What Are SEO Services? What You Are Actually Buying

Updated September 2026 · Written and maintained by the Progression Agency strategy team

SEO services is an umbrella term covering six genuinely different disciplines, sold at prices ranging from $300 a month to $30,000, by providers whose work varies more than any other marketing category. This guide explains what each service actually involves, what it costs, what it can and cannot achieve, how long it takes, how to tell competent work from expensive theater, and which of the six your business actually needs. We sell these services. We have also told businesses they did not need them, so the assessments below are the ones we give clients rather than the ones that maximize a retainer.

The short answer

SEO services are the ongoing work of making a website rank in search results for terms its potential customers actually use. In practice that breaks into six disciplines: technical SEO, on-page optimization, content production, off-page authority building, local SEO, and measurement. A provider may sell you all six, a subset, or one of them described as though it were all six.

The single most useful thing to understand before buying is that these are different jobs requiring different people. The person who diagnoses a crawl-budget problem is rarely the person who writes a genuinely good comparison page, and neither is the person who earns a link from a trade publication. When somebody quotes $500 a month for SEO services, the honest question is not whether that is cheap but which of the six they mean.

Most disappointment with SEO traces to a mismatch rather than to incompetence. A restaurant buys content marketing when it needed Google Business Profile work. A SaaS company buys citation building when it needed a content engine. The work gets done, the invoice is fair, and nothing happens.

The six categories, and what each one actually does

These are listed in the order a competent provider should tackle them, which is not the order they are usually sold in.

Why the order matters more than the list

Publishing content on a site Google cannot crawl properly is buying inventory for a shop with the door locked. Earning links to a page that does not match intent is expensive decoration. The sequence is technical, then on-page, then content, then authority — and providers who start with content because it is the most visible deliverable are optimizing for looking busy.

Why almost nobody needs all six equally

A plumber in Newark needs local SEO and a competent Google Business Profile far more than a blog. A B2B software company needs content depth and almost no local work at all. Buying an evenly balanced package is usually a sign the provider has not diagnosed anything.

Technical SEO services: what they actually involve

Technical SEO is the least glamorous category and reliably the highest return per hour in the first ninety days, because it is the only category where you are removing obstacles rather than building something new. Fixing an accidental noindex on a category page can restore traffic in a fortnight. Nothing else in SEO moves that fast.

Crawlability: can Google reach your pages at all?

Search engines discover pages by following links and reading sitemaps. Anything unreachable by both is invisible regardless of quality. Common causes are orphan pages with no internal links pointing at them, content behind JavaScript that never renders, pagination that dead-ends, and robots.txt rules written years ago for a site structure that no longer exists.

The service here is a full crawl, a comparison of what was crawled against what should exist, and remediation of the gap. Any provider who cannot show you that comparison is not doing technical SEO.

Indexation: crawled is not the same as indexed

Google crawls far more than it keeps. Pages get crawled and then excluded as duplicates, as thin content, or as low-value variations of something already indexed. Google Search Console‘s Page Indexing report is the primary diagnostic, and reading it properly is a genuine skill — most of its categories are routinely misinterpreted.

The most common finding is a site with 4,000 indexable URLs of which 600 matter, where crawl attention is being spent on parameter variations, tag archives and filtered views. Consolidating that is frequently worth more than a quarter’s content budget.

Site speed and Core Web Vitals

Speed is a ranking factor and, more importantly, a conversion factor. The three Core Web Vitals measure loading, interactivity and visual stability. In practice, nearly every failure we diagnose comes down to unoptimized images, render-blocking scripts, and third-party tags nobody remembers adding.

Be skeptical of speed work sold as a standalone product at a high price. The genuine fixes are usually a day of work: compress and correctly size images, defer non-critical scripts, audit the tag manager, and set explicit dimensions on media. Beyond that you are into diminishing returns quickly.

Structured data and schema markup

Structured data tells search engines explicitly what a page represents — a service, an article, a product, a business with an address and opening hours. It is how you earn rich results, and it has become more consequential as AI systems increasingly rely on machine-readable confirmation rather than inference. Schema.org is the vocabulary; Google’s search gallery documents what actually produces a visible result.

Seventy-two percent of the sites we audit have missing or invalid structured data, which makes it the most common technical gap and one of the cheapest to close.

Canonicalisation and duplicate content

Most sites generate the same content at multiple URLs without meaning to: parameters, session IDs, print views, HTTP and HTTPS, with and without trailing slashes, paginated series. Canonical tags tell Google which version is authoritative. Getting them wrong splits ranking signals across duplicates and is one of the quieter ways a site underperforms for years.

Redirects and site migrations

Migrations are where SEO value is most often destroyed outright, and almost always through redirects. Every URL that changes needs a mapped 301 to its closest equivalent. Redirect chains, loops, and wholesale redirects to the homepage are the three failure modes, and the third is the most damaging because it looks tidy.

If you are replatforming, the redirect map should be the first deliverable of the project, before any design decision. We treat it that way and would suggest demanding the same from anybody you hire. There is more detail in our guide to web design and development.

Log file analysis

Sold as advanced, and genuinely useful only at scale. Server logs show what Googlebot actually requested rather than what you assume it did. On a 500-page site this tells you very little. On a 100,000-URL ecommerce site it tells you where crawl budget is being wasted, and it can be worth a great deal. If a provider proposes log analysis for a small brochure site, ask why.

What technical SEO cannot do

A technically flawless site with nothing worth reading will not rank. Technical SEO removes ceilings; it does not create demand. The realistic framing is that it lets your content perform to its actual quality — which is excellent news if the content is good and irrelevant if it is not.

On-page optimization services

On-page work is everything you control on the page itself, and it is where the cheapest wins in SEO live — because you are improving assets that already exist rather than creating new ones.

Search intent matching, which matters more than everything else on this list

Every query has an intent behind it: to learn, to compare, to buy, or to reach a specific site. Google has already decided which intent a query carries, and it demonstrates that decision in the results it shows. If the top ten for your target term are all buyer’s guides, your service page will not rank there no matter how well optimized it is.

The service is reading the actual results, classifying what format is being rewarded, and either matching it or choosing a different term. This single discipline separates SEO that works from SEO that is technically correct and commercially useless.

Title tags and meta descriptions

The title tag remains the strongest on-page signal and the thing most often left as the theme’s default. Meta descriptions are not a ranking factor but heavily influence click-through, which is worth real traffic on terms where you already rank. Forty-four percent of sites we audit have duplicate titles across multiple pages, which is frequently the fastest measurable fix available.

Heading structure

Headings should describe the content beneath them, in a logical hierarchy, using the language people actually search with. They are not decoration and they are not a place to repeat your keyword. Their practical value has increased because AI systems use heading structure to locate the passage that answers a question.

Internal linking

The most underused lever in SEO and one of the few entirely within your control. Internal links distribute authority, establish which pages are important, and give search engines the context to understand relationships between pages. A page with no internal links pointing at it is a page you have implicitly decided not to rank.

The service involves mapping your site’s link graph, identifying orphans and dead ends, and building deliberate paths from high-authority pages to the commercial pages that need help. It costs almost nothing and is skipped constantly.

Keyword cannibalization

When several pages target the same intent, they compete with each other and Google picks — often the wrong one, and often a different one each week. The fix is consolidation: merge, redirect, or deliberately differentiate. Diagnosing this properly requires looking at which URL actually ranks for each term over time, not just at what the pages are titled.

Image optimization and alt text

Compression and correct sizing are speed work. Alt text is accessibility work that happens to be read by search engines. Both matter; neither is a growth strategy, and any provider presenting alt text audits as a major deliverable is padding the invoice.

Content production services

Content is the largest line in almost every well-run SEO budget and the one where quality varies most wildly between providers. The gap between content that ranks and content that exists is enormous, and it is almost entirely a function of whether the writer had anything specific to say.

What you are actually paying for

Not words. You are paying for research into what the searcher wants, analysis of what already ranks and why, a brief specific enough to prevent filler, a writer with genuine subject knowledge or access to somebody who has it, editing, and the internal linking that gives the finished page a chance.

A $60 article from a content mill contains none of that, which is why it produces nothing. A $600 article with all of it frequently pays for itself many times over. The price difference is real work, not margin.

The formats that actually earn traffic

Why pricing content works so well

Most companies refuse to publish prices, so search results for cost queries are dominated by pages that talk around the question. A page that answers it directly, with real bands and honest caveats, outperforms far stronger domains — because it is the only result that does the job. Our own SEO pricing guide and marketing agency pricing guide exist for exactly this reason.

Refreshing existing content versus publishing new

Updating a page that already ranks eleventh is usually a better investment than a new page starting from nothing, and it is far cheaper. A serious content service allocates budget to both, and reviews published pages at ninety days against real query data rather than moving straight on to the next brief.

Content that AI search will actually cite

The same qualities that made content good for people now determine whether machines quote it: a direct answer stated plainly near the top of a section, specific and verifiable claims, original data, named examples, and clean structure. Vague, hedged, everything-to-everyone writing is invisible to extraction.

Off-page SEO and digital PR services

Off-page work means earning signals from other websites that your site deserves attention. In practice that mostly means links, though brand mentions and citations increasingly matter in their own right.

Why links still matter

Links remain among the strongest signals of authority because they are hard to fake at scale without leaving evidence. A link from a publication your customers read is a genuine endorsement. A link from a site that exists only to sell links is not, and Google has become very good at telling the difference.

What legitimate link acquisition looks like

• Digital PR. Producing something genuinely newsworthy — original research, a data study, a useful free tool — and pitching it to journalists. Highest quality, highest cost, most durable.
• Expert commentary. Responding to journalist source requests with genuine expertise. Excellent value when done consistently, and it compounds as relationships form.
• Original research and tools. A calculator, a benchmark study, a dataset. These earn links passively for years, which makes their true cost per link far lower than it first appears.
• Guest contribution. Legitimate when the publication is real, the audience is relevant and the piece is genuinely good. Worthless when it is a link placement dressed as an article.
• Partnerships and sponsorships. Local organizations, industry bodies, events. Modest authority, very real topical and geographic relevance.
• Unlinked mention reclamation. Finding places your business is named without a link and asking for one. Unglamorous, cheap, and consistently effective.

What to avoid, and why it is worse than doing nothing

Bought links from vendor lists, private blog networks, mass directory submissions, comment and forum spam, and reciprocal link schemes. These range from ineffective to genuinely damaging. The economics are the giveaway: if links were purchasable at $90 each with no downside, everybody would buy them and they would signal nothing.

How much link building a business actually needs

Far less than most agencies sell, and rather more than most businesses want to hear. For local service terms, relevance beats volume decisively — a handful of local and industry links will do more than fifty generic ones. For competitive national terms, there is no way around sustained authority work over years.

Local SEO services

Local SEO is the discipline of winning the map pack and local queries, and it is a genuinely different job from national SEO with different signals and a different toolkit.

Google Business Profile management

The highest-leverage single asset in local search, and the one most businesses neglect after initial setup. Categories, services, attributes, hours, photos, posts, questions, and products all feed placement. Google’s own documentation is the authority; most third-party advice about it is outdated within a year.

The service is ongoing: category testing, photo cadence, post publishing, and responding to every review. It is not a one-off setup, though it is frequently sold as one.

Reviews

Review quantity, average rating, recency and response rate all influence local ranking, and reviews influence conversion far more than they influence ranking. The lever is a systematic request process at the moment of highest customer satisfaction, not a burst of solicitation. We cover the mechanics in our guide to getting more Google reviews.

Citations and NAP consistency

Your name, address and phone number should be identical everywhere they appear. Inconsistency creates ambiguity about which business is which. That said, citation building is the most over-sold product in local SEO — the major aggregators and the significant industry directories matter, and the three-hundred-listing packages do not.

Location pages that survive contact with Google

If you serve multiple towns, each needs a page that is genuinely about that town — real landmarks, real projects, real specifics — rather than the same paragraph with the place name swapped. Templated location pages at scale are one of the clearest quality signals Google acts on, and they are the most common way a local SEO program creates a problem instead of solving one.

Service-area businesses

If you travel to customers rather than receiving them, proximity works against you and you cannot fix it directly. The realistic strategy is to build genuine relevance and authority for each area you serve, accept that you will not win the map pack far from your registered address, and compete harder in organic results where proximity weighs less.

Ecommerce SEO services

Ecommerce SEO is technical SEO at scale plus a content problem most retailers never solve. The difficulties are structural rather than clever, which is why they persist across so many stores.

Category pages are the commercial asset, not product pages

Most stores optimize product pages and neglect categories, which is backwards. Category pages target the terms with volume — ‘running shoes’ rather than a specific SKU — and they are where most commercial organic traffic to a store actually lands. They usually contain nothing but a product grid, which gives Google very little to rank.

Faceted navigation and crawl waste

Filters generate near-infinite URL combinations. Left unmanaged, Google spends its crawl budget on color-and-size permutations instead of the pages you sell from. The service is deciding which facets should be indexable, which should be canonicalised, and which should be blocked outright — and it is one of the highest-value technical interventions available to a large store.

Product content and duplication

Manufacturer descriptions appear identically across every retailer selling the item, so none of them ranks on the strength of that text. Original product copy is expensive at scale and genuinely works. The realistic approach is to prioritize the products that actually generate margin rather than rewriting a 40,000-SKU catalog.

Out-of-stock and discontinued products

A silent, continuous source of SEO damage. Deleting the URL destroys accumulated authority; leaving it live with no stock frustrates users. The correct handling depends on whether the product returns, and a competent ecommerce SEO service has an explicit written policy for it rather than an ad-hoc habit.

Structured data for products

Price, availability, ratings and shipping in Product schema produce rich results that materially improve click-through. On a store this is among the highest return-per-hour work in the entire discipline.

Platform constraints worth knowing before you commit

Shopify forces its own URL patterns and limits variant structures; WooCommerce gives you total control and total responsibility; BigCommerce sits between them. None of these is disqualifying, but each shapes what is possible. We cover the trade-offs in our Shopify pricing and platform guide.

International and multi-language SEO services

Relevant only if you genuinely sell across borders or languages, and a common way to waste money if you do not. The core problem is telling search engines which version of a page belongs to which audience.

Hreflang, and why it is so often wrong

Hreflang annotations declare language and regional targeting. They must be reciprocal — every page referencing another must be referenced back — and they break constantly during site changes. Google’s documentation is precise, and implementations in the wild frequently are not.

Domain structure choices

Translation is not localization

Translated keywords are frequently not the terms people in that market actually search. Direct translation of a keyword list is one of the most reliable ways to produce a technically correct site that ranks for nothing. Proper international SEO includes native keyword research per market.

SEO audits: what one should actually contain

An audit is the most commonly mis-sold product in this industry. Free audits are lead magnets generated by software in under a minute. A real audit is several days of expert work and is worth paying for.

What separates a real audit from a generated one

• A generated audit lists every issue a crawler can detect. A real audit tells you which three matter and why the rest can wait.
• A generated audit has no commercial context. A real one connects technical findings to the pages that make you money.
• A generated audit produces a 90-page PDF. A real one produces a prioritized, costed action list somebody could execute on Monday.
• A generated audit cannot tell you that your biggest problem is that three pages compete for the same term. A real one starts there.

What a proper audit covers

1. Crawl and indexation. What exists, what is reachable, what is indexed, and the gap between those three numbers.
2. Analytics validation. Whether your measurement is trustworthy. It very often is not, and every conclusion downstream depends on it.
3. Keyword and intent mapping. What you rank for, what you should rank for, and where pages are competing with each other.
4. Content assessment. Which pages earn traffic, which are dead weight, which are close enough to improve cheaply.
5. Technical health. Speed, structured data, canonicals, redirects, mobile rendering.
6. Authority profile. Link quality and distribution, and whether anything looks risky.
7. Competitive gap. What the sites beating you have that you do not — specifically, not generically.
8. Prioritized roadmap. Sequenced by impact against effort, with owners and rough costs.

What audits cost

A genuine audit for a small site runs $1,500 to $4,000. For a large or ecommerce site, $5,000 to $20,000. Anything free is software output, which has some diagnostic value and no strategic value. Paying for an audit and then executing it yourself is a legitimate and often sensible way to buy SEO.

Keyword research as a service

Keyword research is sold as a deliverable and is really a strategic exercise. The list is the least valuable part of it; the mapping and the judgment are where the value sits.

Volume is the least important number

A term with 40 monthly searches and unambiguous buying intent is worth more than one with 40,000 and no commercial value. We have seen agencies deliver keyword strategies built entirely on volume that would, if executed, have produced a great deal of traffic and no customers.

What actually matters in a keyword

The cannibalization test worth knowing

If two keywords return substantially the same pages in substantially the same order, they are one intent and must be served by one page. This test has repeatedly saved us from building two pages that would have competed with each other, and it takes two minutes.

Migration SEO and penalty recovery

Migration support

Any change to URLs, platform, domain or site structure risks the rankings you already have. Migration SEO is the discipline of not losing them: a complete URL inventory, a mapped redirect for every changed URL, staging validation before launch, and daily monitoring for the first fortnight afterward.

The cost of doing this properly is a fraction of the cost of recovering from not doing it. Rebuilding authority lost in a botched migration takes six to eighteen months, and some of it never returns.

Recovering from a traffic drop

Diagnosis first, always. Traffic drops come from algorithm updates, manual actions, technical breakage, lost links, seasonality, or a competitor doing something better — and the remedies have nothing in common. Providers who begin remediation before diagnosis are guessing with your money.

Manual actions appear in Search Console and are unambiguous. Algorithmic declines do not, and identifying them requires correlating the date against known update timelines, then examining which page types and which query classes lost ground. That pattern usually reveals the cause.

An honest note about recovery

Not all traffic losses are recoverable. If a site lost traffic because Google decided a different content format better serves the query, no amount of optimization restores the old page. The correct response is to serve the new format, and a provider promising to restore previous rankings without having diagnosed the cause is selling hope.

Analytics, tracking and reporting services

The smallest line in the budget and the one that determines whether you can evaluate everything else. In most engagements we inherit, the measurement is broken in at least one material way, and every conclusion drawn from it before we arrived was therefore unreliable.

What breaks most often

• Form submissions not tracked as conversions, so lead generation is invisible.
• Bot and spam traffic inflating sessions and destroying conversion rates.
• Phone calls untracked, which for service businesses is frequently the majority of inquiries.
• Traffic from a subdomain or a separate landing-page tool not attributed to the site at all.
• Internal and agency traffic never excluded, quietly inflating everything.
• Goals configured to fire on page load rather than on the action, counting bounces as conversions.

Reporting against the right thing

Rankings are an input, not an outcome. Traffic is closer but still not the point. The metrics that matter are non-brand organic sessions, inquiries attributable to organic, and revenue from those inquiries. A report that leads with average position is a report designed to look good.

What good reporting actually looks like

Monthly, short, and specific: what shipped, what moved, what did not, what happens next. A twelve-page automated dashboard with no narrative is not reporting — it is evidence that nobody looked at the account. Ask for the narrative.

SEO for AI search: AEO, GEO, and what is actually new

AI Overviews, ChatGPT, Perplexity and similar systems increasingly answer queries directly rather than sending a click. This is a genuine change in the search landscape and it has produced a great deal of opportunistic product naming.

What genuinely changed

Answers are extracted rather than merely linked, which rewards pages that state things plainly and punishes pages that bury the answer under six paragraphs of preamble. Machine-readable confirmation via structured data matters more. Being unambiguously identifiable as a specific entity — a named business, in a named place, doing named things — has become more valuable than keyword repetition ever was.

What did not change

Crawlability, indexation, speed, internal linking, intent matching and earned authority. Every one of these still applies exactly as before. The large majority of what is sold as ‘AI SEO’ or ‘GEO’ is these fundamentals with a new label and a higher price.

The one genuinely new deliverable

Making your site legible to AI crawlers — a sensible robots policy for AI user agents, and a plain statement of what your business is and does that a machine can read. This is real work, it is not expensive, and it is worth doing. It is not worth a separate retainer.

What this means for measurement

Some queries that used to produce clicks now produce answers with no click, and impressions will increasingly outrun clicks. This is not necessarily your provider failing. It does mean the value of ranking for purely informational terms has fallen relative to commercial ones, and budget should shift accordingly.

What SEO services cost

Pricing in this industry is opaque because most providers would rather have the conversation on a call. The bands below are what competent work actually costs in the United States, and they hold reasonably well across providers of similar quality.

Why there is a floor below which nothing works

SEO is labor. At $500 a month, after tools and overhead, a provider can afford perhaps two hours of an experienced person’s attention. Two hours a month cannot audit, write, build links and report. What gets delivered instead is an automated report and a token edit, which is why cheap SEO has such a poor reputation — not because it is bad value but because it is not the service being described.

Why per-keyword pricing is a warning sign

Pricing per keyword implies the work scales with the number of terms, which it does not. The work scales with the state of your site and the competitiveness of your market. Per-keyword pricing also incentivizes the provider to pick easy terms nobody searches, and we have seen reports celebrating first-place rankings for phrases with zero monthly volume.

How to think about the budget as an investment

The clearest way to value organic traffic is to ask what the same clicks would cost in Google Ads. A business receiving 2,000 monthly organic visits on commercial terms at a $4.80 blended cost per click is receiving roughly $9,600 a month of traffic it is not paying for per click. Against a $4,000 retainer, that is straightforward arithmetic — and it is the calculation your provider should be willing to run in front of you.

How SEO services are bought: four models

Agency retainer

A team covering all disciplines, a written scope, monthly reporting. The right answer for most businesses without internal capacity, because SEO genuinely requires several different skill sets and one person cannot supply them all. The risk is paying for a team and getting a junior account manager, which is why asking who specifically does the work matters.

Freelance specialist

Frequently excellent value, particularly for a focused problem — a technical audit, a migration, a recovery diagnosis. A good freelancer will outperform a mediocre agency comfortably. The constraints are capacity and continuity: they cannot run technical, content and outreach simultaneously, and if they are unavailable everything stops.

Consultant plus in-house execution

An experienced external voice sets direction and reviews work; your own team executes. Cheaper than a full retainer and often more effective, because internal people understand the business. It only works if somebody internally genuinely owns delivery — otherwise you have bought advice nobody acts on.

Full in-house

Sensible when organic is a primary acquisition channel and the volume of work justifies dedicated headcount. Expensive once you account for salary, employment costs and tooling, and it still usually requires external content and link support.

The model nobody advertises

A great many in-house SEO hires end up coordinating freelancers for writing and outreach, which is effectively a hybrid. There is nothing wrong with it — it is often the most cost-effective structure available — but it is worth planning deliberately rather than arriving at by accident.

How long SEO takes, honestly

The honest answer is six to twelve months before compounding returns, and faster only if you already have authority to build on or if something specific and fixable is holding the site back.

The one genuine quick win

If pages are accidentally blocked from indexing, unblocking them can restore traffic within a fortnight. This is the only reliably fast outcome in SEO, and it is a repair rather than a gain — you are recovering traffic you should already have had.

Why the middle months are where clients quit

Months three to five typically show rising impressions and modest click growth. It looks like nothing is happening while a great deal is. This is where most programs get canceled, usually about two months before they would have started to pay. If you cannot commit to twelve months, the honest advice is to spend the money on paid search instead, where results are immediate and stop when you stop.

SEO and paid search are not alternatives

Paid delivers immediately and stops the moment you stop paying. Organic takes months and compounds. The sensible structure for most businesses is paid funding the pipeline now while organic is built underneath it, then rebalancing as organic matures. Framing them as a choice is a mistake that usually originates with whichever specialist is doing the framing.

How to tell competent SEO services from expensive theater

The industry has a deserved reputation problem, and the reason is that the work is difficult for a buyer to evaluate. You cannot easily tell whether a technical recommendation is correct, which means the market rewards confident presentation over competence. These are the signals that actually distinguish the two.

They diagnose before they prescribe

A provider who quotes a package before looking at your site is selling a product, not a service. The correct sequence is: look at the site, look at the market, then say what the work is. Anybody who can price your SEO on a first call without seeing anything has a package they sell to everybody.

They can name the pages

Ask what specifically they will work on in month one. A competent answer names pages, terms and fixes. A weak answer describes categories of activity. This one question separates providers faster than any other.

They tell you what they will not do

Every real engagement has boundaries. A provider who says yes to everything either misunderstands the scope or intends to renegotiate later. Hearing ‘that is outside what we would do, and here is why’ is a strong positive signal.

They report bad news without being asked

The strongest single indicator of an honest provider. Everything in SEO involves experiments that fail. A provider who volunteers ‘we tried this, it did not work, here is what we are doing instead’ is showing you their actual process.

They do not guarantee rankings

Nobody controls Google. A guarantee is either meaningless — guaranteeing a ranking for a phrase nobody searches — or it is a claim of influence over Google that does not exist. Both are disqualifying.

What to ask for before signing

1. A written scope naming the disciplines included and excluded.
2. The names of the people doing the work, and a conversation with at least one of them.
3. The reporting format, in advance, with an example from a real client.
4. The notice period, in writing.
5. Confirmation in writing that you own the analytics properties, the content and the accounts.
6. Two references in a comparable competitive position, and permission to contact them.
7. A description of what the first ninety days looks like, specifically.

What SEO services cannot do

Being clear about the limits is the fastest way to avoid a bad engagement, and this is the section most competitors omit.

• Guarantee any ranking. Not for any term, at any price, on any timeline.
• Produce results in thirty days. Except by repairing something specific that was broken.
• Fix a business problem. If the offer, pricing or service is uncompetitive, more visitors reveal that faster rather than solving it.
• Work without content. A program with no content budget is an audit with a monthly invoice attached.
• Overcome a fundamental authority gap instantly. Some terms are genuinely out of reach for a new domain for years, and a good provider will tell you which.
• Survive being switched on and off. Six months on, six months off, six months on is close to the worst possible allocation, because you pay the slow ramp twice.
• Replace a working sales process. Organic leads still need to be answered, quickly.
• Make Google prefer you for a query it has decided means something else.

Which SEO services your business actually needs

The general answer is useless, so here is the specific one by business type. These are the emphases we would actually recommend, not a menu.

Local trades and home services

Google Business Profile management, review generation, and genuinely differentiated location pages, in that order. Content marketing is a distant fourth. The map pack is where the money is, and most providers sell these businesses a blog instead. Our local SEO services guide covers the mechanics.

Dental, medical and healthcare practices

Local SEO plus reviews plus service-specific pages, with unusual care about accuracy and compliance. Health content is held to a higher evidential standard by Google, and thin or unreviewed medical content is actively risky. We cover the specifics in our medical marketing guide.

Law firms

The most competitive local vertical there is, with the highest cost per click in most markets. Practice-area depth, genuine authority signals, and real link acquisition. This is a market where a small budget genuinely will not compete, and an honest provider says so.

Ecommerce

Technical SEO at scale, category page content, product structured data, and faceted navigation control. Blog content is a supporting act, not the main event, and most ecommerce content programs are built the wrong way round.

B2B and SaaS

Content depth is the whole game. Comparison pages, alternatives pages, integration pages, and genuinely useful documentation. Local SEO is close to irrelevant. Buying cycles are long, so attribution needs to be built deliberately or the program will look like it is failing when it is working.

Restaurants and hospitality

Google Business Profile, photos, reviews, menu accuracy, and hours that are never wrong. Almost nothing else moves the needle, and this is the vertical where the least SEO produces the most result.

Multi-location and franchise

A location page architecture that scales without becoming templated, individual Google Business Profiles managed properly, and clear canonical strategy. This is where templated location pages do the most damage, because the pattern is so obvious at scale.

Doing SEO yourself: what is genuinely feasible

Not every business should buy SEO services, and we would rather say so here than after taking a retainer. A great deal of the highest-value work is genuinely doable in-house.

What you can absolutely do yourself

• Claim and fully complete your Google Business Profile. Categories, services, hours, photos.
• Ask every satisfied customer for a review, systematically, at the right moment.
• Write honest, specific pages about each service you offer.
• Set up Search Console and read the query report monthly.
• Fix your own title tags.
• Link between your own pages deliberately.
• Compress your images.

For a small local business, that list alone will outperform a large share of the $500-a-month services on the market. If budget is genuinely tight, do these seven things before hiring anybody.

What genuinely needs expertise

Technical diagnosis, migration planning, penalty recovery, competitive content strategy at scale, and link acquisition. These are where an expert earns their fee several times over, and where a well-intentioned amateur can do real damage.

A month-by-month walkthrough of a real engagement

Abstract descriptions of SEO services are hard to evaluate. This is what twelve months actually looks like on a mid-sized retainer for a business with an established but underperforming site — the shape of engagement we run most often. Your months will differ in detail; the sequence should not.

Month one: find out what is actually true

Full technical crawl and a comparison of what exists against what is indexed. Analytics validation, which usually surfaces at least one measurement problem that makes historical data unreliable. Baseline rankings and traffic recorded so that later claims of improvement can be checked. Competitor set identified — the sites actually beating you for your terms, which is often not who you assumed.

Deliverable: an audit and a prioritized roadmap. Visible traffic change: none, and anybody promising otherwise in month one is skipping the diagnosis.

Month two: fix what is blocking existing pages

Indexation problems resolved, canonical tags corrected, redirect chains collapsed, structured data implemented, the worst Core Web Vitals failures addressed. Title tags rewritten across the pages that already receive impressions, because those are the cheapest clicks available.

This is the month where a genuine quick win occasionally appears. If pages were accidentally excluded from the index, restoring them can move traffic within a fortnight. It is a repair rather than a gain, but it is real money.

Month three: intent mapping and the content plan

Every commercially relevant term mapped to exactly one page. Cannibalization identified and resolved by merging or differentiating. Gaps identified — the intents you should serve and currently do not — and a content plan built against them, sequenced by commercial value rather than by search volume.

Impressions typically begin rising this month as the technical fixes take effect. Clicks usually have not moved yet, and this is where the first difficult conversation happens.

Month four: optimize before you create

Rewriting and expanding pages that already rank between positions eight and twenty, because moving a page from eleventh to fifth is far cheaper than creating a new page and getting it to fifth. Internal linking restructured so the commercial pages actually receive authority from the pages that have it.

First meaningful click growth usually appears here, concentrated on long-tail terms.

Month five: content production begins in earnest

The first new pages from the content plan ship. Typically two to four substantial pages, briefed specifically, written by somebody with genuine subject knowledge, and internally linked on publication rather than left orphaned.

Authority work begins in parallel: source requests answered, relationships opened, the first original research commissioned.

Month six: the halfway review

An honest assessment against the baseline. What moved, what did not, and what we were wrong about — there is always something. The roadmap gets rewritten rather than defended, because six months of real query data is far better information than the assumptions the original plan was built on.

Months seven to nine: compounding starts

Pages published in months four and five begin ranking. Content velocity is maintained rather than increased, because consistency compounds and bursts do not. Link acquisition runs continuously. Mid-tail terms enter the top twenty and the first head terms become visible.

Months ten to twelve: the program starts paying

Non-brand organic sessions and attributable inquiries are both materially above baseline. The work shifts from building foundations to widening the lead — more content in the areas already working, more authority behind the pages that convert.

This is the point most canceled programs never reach, and it is why the twelve-month commitment matters more than the monthly rate.

How search engines actually work

Almost every SEO service is an intervention in one of four processes. Understanding them makes it much easier to judge whether a proposed piece of work is likely to do anything.

Crawling

Search engines discover pages by following links and reading sitemaps. Crawl budget — the attention a site receives — is roughly proportional to its authority and the freshness of its content. On a small site this is irrelevant; on a large one, wasting crawl budget on parameter variations means your important pages are visited less often.

Rendering

Modern sites build much of their content with JavaScript. Google renders pages in a second pass, which is slower and occasionally incomplete. If your critical content or your internal links only exist after JavaScript executes, you are relying on a process that sometimes does not finish. Server-side rendering removes the risk entirely.

Indexing

Rendered pages are assessed and either stored or discarded. Google indexes far less than it crawls, discarding duplicates, thin pages and low-value variations. A page not in the index cannot rank at any price, which is why indexation is the first thing a competent provider checks.

Ranking and retrieval

For each query, candidate pages are scored against hundreds of signals and ordered. The signals are not published and change continuously, but the categories are stable: relevance to the query, the quality and depth of the content, the authority of the site, the experience of using the page, and context such as location and device.

Why this matters when evaluating a proposal

Every legitimate SEO deliverable maps to one of these four. If a provider proposes work you cannot map — ‘search engine submission’, ‘keyword density optimization’, ‘meta keyword tags’ — that is because it does not intervene in any real process. Those three examples have been obsolete for well over a decade and are still sold.

Ranking factors, ranked honestly

Nobody outside Google knows the weights, and anybody presenting a precise numbered list is presenting a guess. What follows is our working order of operations based on what actually moves client sites, with the uncertainty stated.

The factors people over-weight

Keyword density, exact-match domains, meta keyword tags, word count as a target in itself, and the number of links irrespective of quality. Each of these was once meaningful, each is now either irrelevant or actively counterproductive, and each is still sold.

The factors people under-weight

Internal linking, search intent matching, and the simple question of whether the page is actually better than the ones currently ranking. These three are free, entirely within your control, and consistently the difference between a page that works and one that does not.

The glossary: every term your provider will use

SEO has an unusual amount of jargon, and some providers use it to make ordinary work sound sophisticated. Here is what the vocabulary actually means, so you can follow a proposal without nodding along.

Crawling and indexing terms

On-page and content terms

Off-page and authority terms

Local, measurement and modern terms

SEO myths, examined

Some of these are outdated advice that was once correct. Others were never true. All of them are still being sold, and believing them will cost you money.

‘You need to publish content every week’

Consistency matters; frequency is negotiable. Two genuinely good pages a month sustained for a year beats twenty thin ones published in a burst. Publishing low-quality content on a schedule actively harms a site by diluting topical clarity and wasting crawl attention.

‘Word count is a ranking factor’

It is not. Longer pages correlate with better rankings because thorough answers tend to be longer, not because length is rewarded. Padding a page to hit a target makes it worse. The correct length is whatever fully answers the query and no more.

‘Keyword density matters’

Obsolete for well over a decade. Search engines understand synonyms, related concepts and context. Repeating a phrase to hit a percentage produces writing that reads badly and ranks worse.

‘Meta keywords help’

Google has ignored the meta keywords tag since 2009. Any provider still populating it is telling you when they last updated their process.

‘You need to submit your site to search engines’

You do not. Google finds sites by following links. Submitting a sitemap in Search Console is useful; ‘search engine submission’ as a paid service is selling you something free and unnecessary.

‘More links are always better’

One link from a publication your customers read is worth more than a hundred from directories nobody visits. Volume without relevance signals manipulation rather than authority.

‘SEO is a one-time project’

Competitors keep publishing, algorithms keep changing, and content ages. Some work genuinely is one-off — a technical audit, a migration — but ranking is a position relative to others who are still working.

‘You can rank without content’

For a handful of local queries with a well-optimized Google Business Profile, occasionally yes. For anything competitive, no. Content is what a search engine has to assess.

‘Google penalizes duplicate content’

There is no duplicate content penalty. Google picks one version and ignores the others, which splits signals and wastes crawl budget. The outcome is bad, but the mechanism is filtering rather than punishment, and understanding that changes the fix.

‘Social media signals improve rankings’

Not directly. Social activity can earn links and drive traffic that leads to links, which does help. But shares themselves are not a ranking factor, and providers who bundle social posting into an SEO retainer as a ranking activity are padding.

‘Paid ads improve organic rankings’

They do not. Google is explicit about this and it would be a serious conflict of interest. Paid does help indirectly by revealing which terms convert, which is genuinely useful for prioritizing organic work — but there is no direct ranking benefit.

‘A higher domain authority score means you will outrank them’

Domain authority is a third-party estimate, not a Google metric. Pages from DA 17 sites routinely outrank DA 80 sites when they match intent better. Use it for rough comparison and never as a target.

‘Bounce rate is a ranking factor’

Google has repeatedly said it does not use Analytics data for ranking. Bounce rate is a useful diagnostic for you and is not a signal to them.

‘You should disavow toxic links regularly’

Google ignores the overwhelming majority of low-quality links automatically. Routine disavowing is unnecessary and it is genuinely possible to damage a site by disavowing links that were helping.

‘AI content is automatically penalized’

Google’s stated position is that it rewards helpful content regardless of how it was produced. What gets penalized is unhelpful content produced at scale, which is what most unedited AI content is. The distinction is quality and usefulness, not authorship.

‘Rankings are the goal’

Rankings are an input. A first-place ranking for a term nobody searches, or one that brings visitors who never buy, is worth nothing. The goal is qualified traffic that converts, and any report leading with average position is designed to look good rather than to inform.

The tools your provider uses, and what they actually cost

Tooling is a real cost of delivering SEO services and a common source of confusion, because clients sometimes assume they are paying for software when they are paying for labor. Here is what the stack genuinely looks like and what it costs a provider to run.

The free tools that matter most

Google Search Console is the single most important tool in SEO and it is free. It reports what queries your pages appear for, how often, at what position, and with what click-through rate. It shows indexation status per URL, flags manual actions, and reports Core Web Vitals from real users. No paid tool substitutes for it, because no paid tool has Google’s own data.

Google Analytics is free and reports what happens after the click. Between them, these two cover the large majority of what you need to run a program. Any provider who does not have both configured correctly in month one is not doing the job.

Bing Webmaster Tools is free, takes ten minutes, and is worth doing. Bing traffic is small but real, and its tooling occasionally surfaces issues Google’s does not.

The paid tools and what they are for

Crawlers such as Screaming Frog simulate a search engine crawling your site, surfacing broken links, redirect chains, duplicate titles, missing canonicals and orphan pages. This is the workhorse of technical SEO and costs a couple of hundred dollars a year per seat.

Keyword and competitive platforms — Ahrefs, Semrush, Ubersuggest and similar — estimate search volume, difficulty and competitor rankings. They cost between roughly $30 and $500 a month depending on tier. Their data is estimated rather than exact, which is worth remembering when a report quotes volume to the nearest ten.

Rank tracking tools record positions over time. Useful for spotting sudden movement, and dangerous as a primary metric because they encourage optimizing for position rather than revenue.

Log analysers, structured data validators, and page speed tools fill out the rest. Most of the validation tools are free and published by Google itself.

What the whole stack costs

A competent small agency runs $400 to $1,500 a month in tooling across all clients. Spread across a book of business that is a modest per-client cost, which is why tooling should never be a significant line in your quote. If a provider is charging you a large separate software fee, ask what it is for.

Why tools are not the service

Every tool in this stack is available to you. What you are buying is the judgment to know which of the four hundred issues a crawler reports actually matter, which three to fix first, and what to do about them. A provider who forwards you a tool export has automated the easy part and skipped the valuable one.

Contracts, ownership and what to negotiate

The commercial terms of an SEO engagement matter more than most buyers realize, because the failure modes are predictable and almost all of them are avoidable with three clauses.

Own your own accounts

Your Google Analytics property, Search Console, Google Business Profile, ad accounts and hosting should all be owned by your business, with the provider granted access. This sounds obvious and is violated constantly. We have onboarded clients who lost years of analytics history because the previous agency owned the property and removed access when the relationship ended.

Get it in writing that you own these, and check the ownership yourself in the account settings rather than accepting an assurance.

Own the content and the work product

Content produced under a retainer should belong to you outright on payment. So should audits, keyword research, briefs and documentation. Some contracts license rather than transfer, which means you cannot use your own pages if you leave. Read that clause.

Notice periods and lock-ins

A twelve-month minimum term is common and, in fairness, defensible — SEO genuinely takes that long. A twelve-month term with no exit for non-performance is not. The reasonable middle is an initial term with a defined notice period of thirty to sixty days thereafter, and a break clause if agreed deliverables are not delivered.

Be clear-eyed about the tension here: providers ask for term commitments because clients who cancel at month four blame the provider for a result the timeline never permitted. A term commitment protects both parties when the deliverables are also specified.

What the scope should actually say

Named disciplines included and excluded. An indication of volume — how many content pieces, roughly how many hours, what reporting cadence. Named people or at least named roles. And explicit treatment of the two perennial gray areas: who manages the Google Business Profile, and who is responsible for conversion rate optimization.

Performance-based pricing, and why it is usually a trap

Paying only for results sounds ideal and rarely works. Defining the result is the problem: ranking for which terms, measured how, attributed to whom, over what period? Providers on performance deals are incentivized toward easy terms and short-term tactics, because they need a fast measurable win. The structures that work are a retainer with clear deliverables, or a retainer plus a genuine bonus on revenue outcomes.

Red flags in the paperwork

• The provider owns your analytics, Search Console or Business Profile.
• Content is licensed to you rather than transferred.
• Automatic annual renewal with a short cancellation window buried in the terms.
• No specified deliverables, only ‘SEO services’.
• A clause preventing you from working with another SEO provider.
• Links built are described as the provider’s property and may be removed on termination.

How to brief an SEO provider properly

The quality of what you get back is heavily determined by the quality of what you supply. Most briefs we receive contain a URL and a request for a proposal, which produces generic proposals.

What to tell them

1. What you actually sell, and what is most profitable. SEO effort should concentrate on your highest-margin services, and providers cannot know which those are.
2. Who your customers are and how they currently find you. If 80% of business comes from referral, the organic opportunity is different from a business dependent on search.
3. Where you operate. Precisely. ‘The tri-state area’ and ‘within forty minutes of Newark’ imply different strategies.
4. What you have tried before and what happened. Including the bad experiences. It saves everybody repeating a mistake.
5. Your realistic budget range. Withholding it produces proposals aimed at the wrong scope, and the provider will price against their guess of what you will pay rather than what the work needs.
6. Your internal capacity. Who can review content, who can make site changes, how quickly.
7. What a good outcome looks like in your terms. Not rankings — number of inquiries, revenue, or a specific service line you want to grow.

What to ask for back

A specific diagnosis rather than a package. The first ninety days described concretely. Named people. The reporting format. And an honest statement of what they think is hard about your situation — every situation has something, and a proposal that presents your market as straightforward has not looked at it.

Four scenarios, and what we would actually recommend

Abstract advice is easy to agree with and hard to act on. These are four situations we see repeatedly, with the recommendation we would genuinely give.

A local plumber with a five-page site and no organic traffic

Do not hire an agency yet. Complete the Google Business Profile properly — every category, every service, real photos, accurate hours. Build a systematic process for requesting reviews after each completed job. Write one honest page per service you offer. Set up Search Console.

That is perhaps twenty hours of work and it will outperform a $600-a-month retainer comfortably. Once it is done and you are winning some local queries, a small retainer to build on it makes sense. Doing it in the other order wastes a year of fees.

An established ecommerce store with declining organic traffic

Diagnose before anything else. Declining traffic on an established store is usually one of four things: an algorithm update, index bloat from faceted navigation, content decay on category pages, or a migration that went wrong. These have nothing in common as remedies.

Pay for a proper audit — $5,000 to $20,000 at this scale — and insist the deliverable is a diagnosis with evidence rather than a list of everything a crawler found. Then decide whether to execute in-house or retain. Do not let anybody begin remediation before the diagnosis exists.

A B2B software company with a blog nobody reads

The blog is almost certainly targeting informational terms with no commercial intent, which is the single most common B2B content failure. Stop publishing it. Redirect the budget into comparison pages, alternatives pages, integration pages and pricing content — the material buyers read immediately before a decision.

Also fix attribution before anything else. B2B buying cycles run months, so a program that is working will look like it is failing under last-click attribution, and it will get canceled for reasons that have nothing to do with the SEO.

A multi-location business considering 200 location pages

Do not build 200 templated pages. It is the clearest quality signal Google acts on at scale and it is the single most common way a local SEO program creates a problem rather than solving one.

Build ten genuinely differentiated pages for your ten most valuable locations, with real local specifics — actual projects, actual landmarks, actual staff. Measure what those do over two quarters. If they work, extend deliberately. Volume without differentiation is the trap here, and the agencies who sell 200-page location builds know exactly how it usually ends.

What actually changed in SEO over the last decade

A great deal of advice still circulating was correct in 2014 and is wrong now. Understanding what changed helps you evaluate whether a provider’s approach is current, and it explains why some tactics that plainly used to work no longer do.

From keywords to intent

A decade ago, ranking meant using the exact phrase enough times in the right places. Search engines now understand meaning, synonyms and relationships between concepts, which means matching what the searcher wants matters far more than matching the words they typed. This is the single largest shift and the one most legacy advice fails to reflect.

The practical consequence is that a page written naturally about a subject will typically outrank a page engineered around a phrase. It also means keyword research changed purpose: it is now about discovering intents and demand, not about collecting strings to insert.

From volume to quality in links

Link building used to be an arithmetic exercise. Directory submissions, article spinning, reciprocal schemes and comment links all worked, and worked in proportion to volume. A sequence of algorithm updates over the following years devalued nearly all of it, and the industry split into providers who adapted and providers who kept selling the old product to people who could not tell.

The current reality is that a small number of genuinely earned links from relevant, real publications outperforms hundreds of manufactured ones, and the manufactured ones carry risk that the earned ones do not.

From desktop to mobile-first

Google now predominantly uses the mobile version of a page for indexing and ranking. If content, structured data or internal links exist on desktop but not on mobile, they effectively do not exist. This caught out a generation of sites that served a stripped-down mobile experience, and it is still worth checking on any site more than a few years old.

From ten blue links to a crowded results page

A results page now frequently contains an AI Overview, a map pack, People Also Ask, video carousels, shopping units and paid listings before the first organic result. The practical effect is that ranking first no longer means what it did. Click share on informational queries has fallen materially, which raises the relative value of commercial and transactional terms.

This is why we increasingly advise clients to weight content budget toward comparison, pricing and buyer’s-guide pages rather than definitional content — the definitional queries are increasingly answered on the results page itself.

From optimizing pages to establishing entities

Search engines increasingly reason about things rather than strings — a specific business, in a specific place, offering specific services, with a specific reputation. Being unambiguously identifiable has become a genuine competitive advantage, and it is largely achieved through consistent naming, structured data, and being described accurately in places search engines already trust.

What has not changed at all

Make pages people genuinely want to read, make sure search engines can access them, and earn a reputation good enough that other people reference you. Every fundamental in this guide is a refinement of those three sentences, and every provider selling something that does not ultimately serve one of them is selling activity rather than outcome.

What happens after the click, and why your provider should care

SEO services are routinely scoped to stop at the click, which is a mistake for both parties. Traffic that does not convert is a cost, and a provider who is indifferent to what happens after arrival is optimizing a number that does not pay anybody.

Why ranking for the wrong thing is worse than not ranking

A page that attracts researchers when you sell to buyers produces traffic, flattering reports and no revenue. It also consumes budget that could have gone to commercial terms. We have taken over accounts where organic sessions had tripled over two years and inquiries had not moved at all, and in every case the cause was the same: the content plan was built on volume rather than intent.

The elements that actually convert organic traffic

• A page that delivers what the search promised, immediately and without a preamble.
• A clear, single next action rather than five competing ones.
• Genuine proof — specific results, named clients, real photographs rather than stock.
• Pricing, or at least honest price ranges. Its absence is the most common reason a qualified visitor leaves.
• A contact method that matches how the customer wants to make contact. For many local services that is a phone number, prominently, on mobile.
• Fast loading, because a meaningful share of mobile visitors abandon before a slow page renders.

Tracking that closes the loop

Form submissions, phone calls and chat conversations all need to be tracked and attributed. Phone calls in particular are frequently the majority of inquiries for local service businesses and are frequently untracked entirely, which means the channel driving the most revenue looks like it drives none.

Where conversion work belongs in the scope

This is one of the two perennial gray areas in SEO contracts. Our position is that a provider should at minimum flag conversion problems they can see, even where fixing them is outside scope. A provider who watches a client pay for traffic that lands on a page with no phone number and says nothing is technically within their remit and not doing their job.

Why SEO programs fail: five post-mortems

More useful than success stories, because the failure modes repeat. Each of these is a pattern we have seen enough times to describe generically.

Canceled at month four

By far the most common. Technical work landed, impressions rose, clicks had barely moved, and the business concluded it was not working roughly two months before it would have started to. The money spent is entirely wasted, because a partial SEO program leaves nothing behind that compounds.

The prevention is agreeing in advance what months one to six are expected to look like, in writing, so that the month-four conversation is a check against a plan rather than a judgment made on instinct.

The content that nobody would ever read

A year of weekly blog posts, produced cheaply, on topics chosen by search volume, written by people with no knowledge of the industry. It ranks for nothing, earns no links, and leaves behind a large number of thin pages that dilute the site’s topical clarity. Cleaning it up is a subsequent project.

The prevention is refusing to publish anything nobody at the company would be willing to put their name on.

The migration that went ahead without a redirect map

A redesign launches, the URLs change, and nobody maps the old ones to the new. Traffic falls by sixty to eighty percent within a fortnight. Partial recovery takes six to eighteen months and some of it never returns, because links pointing to dead URLs stop passing value and the referring sites will not update them.

The prevention costs a few days of work and is the single highest-return preventive measure in SEO.

The 200 location pages

A multi-location business publishes hundreds of near-identical pages with the town name swapped. They rank briefly, then a quality update lands and the whole set is devalued — frequently dragging the rest of the site with it, because the pattern is a site-wide quality signal rather than a per-page one.

The prevention is building ten genuinely differentiated pages and measuring them before extending.

The program with no owner

The provider produces recommendations. Nobody internally has the authority or the time to implement them. Six months later the audit is unactioned, the content is unreviewed, and the relationship ends with both parties believing the other failed.

The prevention is naming an internal owner before signing, and being honest if that person does not exist — in which case the engagement needs to include implementation rather than only advice.

Competitor analysis as an SEO service

Sold as a deliverable, and genuinely valuable when done properly, because your rankings are a relative position rather than an absolute score. What matters is not how good your site is but how it compares to the ones currently ahead of you for the specific terms you want.

Identifying the right competitors

Your search competitors are frequently not your business competitors. A national publisher, a directory, a marketplace and a review site may all sit above you for a term while never competing for a customer. Analyzing the wrong set produces a strategy aimed at people you are not actually racing.

What is worth analyzing

• Which of their pages earn the traffic, and for which specific terms.
• What format those pages take, because that reveals the intent Google is rewarding.
• How deep their coverage of the subject is across the whole site, not just on the ranking page.
• Where their links come from, and whether those sources are reachable for you.
• What they do not cover, which is usually where the fastest opportunity sits.

The gap that usually matters most

In our experience the most actionable finding is rarely that a competitor has more links. It is that they have a page for an intent you have not served at all. Those gaps are cheap to close and are frequently sitting in plain sight.

How to use the analysis without copying

Copying a competitor’s page produces a worse version of something that already exists, which has no reason to outrank it. The useful output of competitor analysis is a list of intents to serve and a clear view of the standard you have to beat — then you build something genuinely better, which usually means more specific, more honest, and more useful rather than merely longer.

Site architecture: the shape that ranks

Architecture is the least discussed and most consequential structural decision in SEO, because it determines how authority flows through a site and how easily both users and search engines understand what you do.

Depth: how many clicks from the homepage

Pages buried five or six clicks deep receive less crawl attention and less internal authority. As a working rule, anything commercially important should be reachable within three clicks of the homepage. On large sites this is achieved through good category structure and hub pages rather than by stuffing the navigation.

Hubs and clusters, without the ceremony

Grouping related content around a comprehensive hub page, with links running both ways, genuinely helps — it establishes topical depth and gives search engines a clear map of what you cover. It has also been over-formalized into a methodology sold at a premium. The underlying idea is simply that related pages should link to each other and to a central page.

Navigation is a ranking decision

What you put in the main navigation receives authority from every page on the site. Most businesses populate it by internal politics or habit rather than by commercial value. Reviewing it against what actually makes money is one of the cheapest interventions available.

URL structure

Short, descriptive, stable, lowercase, hyphen-separated, and reflecting hierarchy where hierarchy genuinely exists. The single most important property is stability — changing URLs costs you unless the redirect handling is flawless, so it is worth getting right at the point of creation rather than revisiting later.

Pagination and filtered views

On any site with listings, this is where index bloat originates. Decide deliberately which paginated and filtered URLs should be indexable, canonicalise the rest, and block the combinations that serve no search demand. On a large ecommerce site this single decision can be worth more than a quarter of content.

Refreshing content: the highest-return work nobody schedules

Producing new content is visible and satisfying. Improving existing content is neither, and it reliably returns more per hour, because you are working with pages that have already accumulated authority and already have query data telling you what to fix.

Which pages to refresh, in priority order

1. Pages ranking between positions five and fifteen for commercially valuable terms. These are the cheapest gains available anywhere in SEO — the page already nearly works.
2. Pages with high impressions and poor click-through, which usually indicates a title and description problem rather than a content problem, and can be fixed in minutes.
3. Pages that used to rank and have declined, where content decay is usually the cause.
4. Pages ranking for terms you did not target, which reveal demand you did not know you were serving and can often be expanded into properly.
5. Pages covering subjects where the facts have changed, particularly anything with prices, statistics or product details in it.

What a refresh actually involves

Reading the Search Console query data for that specific URL, which tells you exactly what people expected to find. Adding the sections that are missing, updating anything factually stale, improving the title and description, adding internal links from stronger pages, and updating structured data. It is usually two to four hours per page and it is worth substantially more than the same time spent on a new page.

How often

Review published content at ninety days against real query data, then annually thereafter. Pages containing prices, statistics or anything time-sensitive need a defined update cadence, and a page whose value depends on being current — a trends page, a pricing page — needs it far more often than that.

What not to do

Do not change the URL. Do not delete substantial sections that are earning impressions. Do not consolidate two ranking pages into one without checking that they genuinely serve the same intent, because merging two pages that each rank for different things loses you one of them.

Reporting: what to demand and what to ignore

Reporting is the cheapest part of an SEO service to produce and the easiest to make look impressive, which is exactly why it needs scrutiny.

The metrics that matter, in order

Why brand traffic has to be stripped out

If your business runs any advertising, gets press, or simply becomes better known, searches for your own name rise and organic traffic rises with them. That growth has nothing to do with SEO work. Every SEO report that does not separate brand from non-brand is flattering itself, and the separation takes about a minute in Search Console.

The narrative matters more than the dashboard

A useful report is short and says: here is what we shipped, here is what moved, here is what did not and why, here is what we are doing next. A twelve-page automated export with no commentary is not a report — it is proof that a tool ran. Ask for the narrative and judge the engagement on it.

Seasonality, forecasting and setting expectations internally

A great many SEO programs are judged against a number somebody invented in a meeting. Setting the expectation properly at the outset is as important as the work itself, because it determines whether the program survives long enough to pay.

Seasonality will distort your first read

Most businesses have a seasonal pattern, and comparing month to month against a seasonal trend produces false conclusions in both directions. Compare year on year wherever you have the history, and where you do not, establish the seasonal shape from Google Trends or from your own sales data before drawing conclusions from a quarter.

What an honest forecast looks like

A range rather than a number, tied to explicit assumptions: this level of content production, this level of authority work, this competitive set. Any provider forecasting a precise traffic figure twelve months out is guessing with unwarranted confidence, and the correct response is to ask what assumptions the number depends on.

How to explain SEO to a skeptical finance team

Price the organic traffic at what the same clicks would cost in paid search. It converts an abstract marketing activity into a comparable number, and it is defensible because it reflects what the market genuinely pays for that attention. A business receiving 2,000 monthly commercial-intent visits at a $4.80 blended cost per click is receiving roughly $9,600 a month of traffic it is not renting.

The conversation to have before month four

Agree at the outset what months one to six should look like and write it down. Then the month-four review is a comparison against a plan rather than a judgment made on instinct in a bad week. This single practice prevents the most common cause of wasted SEO spend there is.

When you should not buy SEO services

An honest guide has to include the cases where the answer is no. We would rather say this here than after taking a retainer that was never going to work.

When you cannot commit for twelve months

Six months of SEO stopped halfway is close to pure waste, because almost none of the value has compounded yet. If your budget or your certainty does not extend to a year, paid search is the better allocation — it produces results immediately and stops cleanly when you stop.

When you have no content budget and no internal writer

Content is roughly forty percent of a working program. A retainer with no content capability is an audit with a monthly invoice attached. Either fund it or buy a one-off audit and execute what you can yourself.

When the business problem is not traffic

If you are getting inquiries and not converting them, or your pricing is uncompetitive, or your service has a delivery problem, more visitors makes the problem larger rather than smaller. Fix the conversion or the offer first; the SEO will be worth more afterwards.

When your market genuinely has no search demand

Some products are bought entirely through relationships, tender processes or channels where nobody searches. It is rarer than people assume — most markets have more search demand than the business realizes — but it exists, and it is worth checking the actual volumes before committing a budget.

When you are pre-product or pre-revenue

Validate that people want the thing before investing a year in ranking for it. SEO is a distribution channel, and building distribution for something unproven is an expensive way to discover it was unproven.

SEO by page type: what each page is actually for

Different page types serve different intents and are judged by different standards. A great deal of wasted SEO effort comes from building the wrong type of page for the term being targeted.

The homepage

Usually your strongest page by authority and usually the wrong page to rank for anything specific. Its job is to establish who you are, route visitors to the right place, and pass authority to the pages that do the ranking. Businesses that try to optimize the homepage for their main service term generally do worse than those who build a dedicated service page and link to it prominently.

Service pages

The highest commercial value and the hardest to rank, because everyone selling the service wants the same term and the SERP frequently favors comparison content instead. A service page earns its place by being specific: what exactly is delivered, what it costs, who it suits, what the process is, and what happens if it does not work. Vague service pages are interchangeable and rank accordingly.

Location pages

Valuable when genuinely differentiated and actively harmful when templated. A location page should contain things that could only have been written about that place — real projects completed there, actual neighborhoods served, specifics about how the service differs locally. If you could swap the town name and the page would still read correctly, it is the templated pattern and it will eventually be devalued.

Comparison and alternatives pages

Among the most commercially valuable content types available, because they intercept people at the moment immediately before a decision. They are also uncomfortable to write honestly, which is exactly why they work — most competitors either avoid them or write transparently self-serving versions that readers discount.

Pricing pages

Enormously effective and widely avoided. In most service industries, search results for cost queries are dominated by pages that talk around the question. Publishing real ranges with honest caveats outranks far stronger domains simply by being the only result that does the job, and it qualifies inquiries before they reach you.

Blog and guide content

Serves informational intent, builds topical authority, and earns links. It rarely converts directly and should not be judged as though it should. Its value is that it makes the commercial pages rankable, which means measuring it by direct conversions consistently undervalues it and measuring it by traffic alone consistently overvalues it.

Category pages on ecommerce

The main commercial asset on a store and the most neglected. They target the terms with volume and usually contain nothing but a product grid. Adding genuine buying guidance above or below the grid is one of the highest-return interventions available in ecommerce SEO.

Product pages

Rank for specific, low-volume, high-intent terms. Original descriptions matter because manufacturer copy appears identically across every retailer. Structured data for price, availability and ratings produces rich results that materially improve click-through.

About and team pages

Rarely ranked for anything and disproportionately important for trust and for entity clarity. Real names, real photographs, real credentials. These pages are read by people deciding whether to contact you and increasingly used by search engines to establish who you actually are.

JavaScript, page speed and the technical work that touches your developers

Some SEO work cannot be done by an SEO alone. Knowing which parts require developer time helps you resource the engagement properly rather than discovering the dependency in month three.

JavaScript rendering

If your content, links or structured data only exist after JavaScript executes, you are depending on a second rendering pass that is slower and occasionally incomplete. The reliable fixes are server-side rendering or static generation, both of which are developer projects rather than SEO tasks.

The diagnostic is simple and worth running on any modern site: view the raw HTML source rather than the rendered page, and check whether your main content and internal links are present. If they are not, you have a rendering dependency.

Core Web Vitals in practice

Nearly every failure we diagnose reduces to three causes: images that are unoptimized or missing explicit dimensions, render-blocking scripts and stylesheets, and third-party tags accumulated over years that nobody remembers adding. Auditing the tag manager is frequently the single largest speed win available and costs nothing but attention.

Where speed work stops paying

Getting from a failing score to a passing one is worth real money in both rankings and conversion. Getting from passing to perfect is usually not, and speed optimization sold as an expensive ongoing product is exploiting the fact that the score can always be nudged higher.

Working with your development team

The most common friction in an SEO engagement is a queue of recommendations waiting on developer capacity nobody allocated. Agree at the outset how technical changes get implemented, who prioritizes them, and roughly how much developer time is available per month. A program with no implementation path produces documents rather than results.

Images, video and accessibility

Image SEO that actually matters

Three things: compress and correctly size the file, describe it accurately in alt text, and give it a descriptive filename. That is the whole of it. Image optimization presented as a major ongoing deliverable is padding, though it genuinely matters as part of speed work.

Original photography carries a real advantage that is rarely discussed. Stock images appear on thousands of sites and communicate nothing specific about your business, while real photographs of real work support both conversion and entity clarity.

Video

Video can occupy space in results that text cannot, particularly for how-to queries where video carousels appear. It is expensive to produce well and is worth doing where the query genuinely calls for demonstration. Embedding video also increases time on page, which is good for users regardless of whether it is a ranking signal.

If you publish video, use VideoObject structured data and provide a transcript. The transcript is indexable text, which means a video page can rank on the strength of what was said in it.

Accessibility and SEO overlap substantially

Descriptive alt text, logical heading hierarchy, sufficient color contrast, keyboard navigability and meaningful link text are all accessibility requirements that also happen to help search engines understand a page. Treating accessibility as a compliance chore misses that most of it is simply good structure, and good structure is what makes content extractable by both screen readers and search systems.

This overlap is worth raising with any provider: work done for accessibility frequently pays twice, and it is one of the few areas in this discipline where the ethical and the commercial align completely.

How to be a good client, and get more for your money

The variance in outcomes between clients of the same provider, on the same budget, is larger than most people expect. A meaningful share of it comes down to how the client operates.

Give access quickly

Analytics, Search Console, the CMS, the Business Profile. Weeks routinely disappear at the start of engagements waiting for credentials, and that is time you are paying for.

Nominate one decision-maker

Content reviewed by committee arrives late and comes back contradictory. One person with the authority to approve is worth more to the outcome than an extra thousand a month of budget.

Review content quickly, and on substance

The most valuable review comments correct facts, add specifics only you know, and flag things that would embarrass you. The least valuable rewrite sentences to taste. Your subject knowledge is the input no agency can supply.

Supply the raw material only you have

Real project details, real numbers, photographs of actual work, the questions customers genuinely ask, the objections your sales team hears. This is what turns generic content into content nobody else could have written, and it is the single biggest quality lever available to you.

Tell them what happened commercially

Providers who never learn which inquiries became customers are optimizing blind. If organic leads are poor quality, say so early — that is a targeting problem which is fixable, and it is invisible from the provider’s side.

Do not change the plan every month

SEO compounds through consistency. Redirecting the strategy every time a competitor does something visible guarantees that nothing accumulates. Agree the plan, run it for a quarter, then review it properly.

How SEO agencies actually work, and why prices are what they are

Understanding the economics on the other side of the table makes you a much better buyer, because it tells you what is genuinely negotiable and what is not.

Where your retainer goes

Roughly half to sixty percent of a retainer is direct labor — the people doing the work. Perhaps ten percent is tooling, data and subscriptions. The remainder covers account management, overhead, business development and margin. A firm running at a healthy but not excessive margin is keeping perhaps fifteen to twenty-five percent.

This is why the floor exists. At $500 a month there is roughly $250 of labor available, which at any experienced rate is two hours. Two hours cannot audit, write, build links and report, so what gets delivered instead is automation.

The three agency models

Boutique specialists run small books of clients with senior people doing the work directly. Higher hourly cost, usually better output, limited capacity. Mid-sized agencies use a pod structure with a senior strategist across several accounts and junior execution beneath. Large agencies layer account management over specialist teams, which adds coordination cost and can add genuine capability.

None of these is inherently better. The failure mode to watch for is paying boutique prices for the large-agency experience — a senior person sells the work and a junior delivers it. Asking who specifically does the work, and meeting them, is how you check.

Why agencies push for longer terms

Client acquisition is expensive and the first two months of an engagement are largely setup that produces no visible result. An agency that loses clients at month four never recovers its acquisition cost, which is why term commitments exist. This is a legitimate interest rather than a trick, and the fair resolution is a term commitment paired with specified deliverables.

What is genuinely negotiable

Scope is negotiable — you can trade link building for more content, or reduce reporting frequency. Term length is often negotiable. Payment timing is sometimes negotiable. The hourly value of experienced labor is not, and pushing a provider below their floor produces a worse version of the service rather than the same service cheaper.

Hiring an in-house SEO instead

A legitimate alternative that suits some businesses well and is frequently chosen for the wrong reasons — usually a bad agency experience rather than a considered assessment of what the role needs.

What one person can realistically cover

A strong in-house SEO can own strategy, technical direction, on-page work, measurement and internal advocacy. What they cannot simultaneously do is produce a meaningful volume of content and run outreach. Those are separate jobs, and expecting all of it from one hire is the most common way an in-house SEO function disappoints.

The real cost

A capable mid-level SEO in the US runs $70,000 to $110,000, plus roughly twenty-five percent in employment costs, plus $6,000 to $12,000 in tooling they will need. Call it $105,000 to $150,000 fully loaded, against which a $4,000 monthly agency retainer is $48,000. The comparison is not close on cost; it is closer on control and on institutional knowledge.

When in-house is genuinely right

When organic is a primary acquisition channel, when the site is large enough to generate continuous technical work, when you need somebody embedded in product and engineering decisions, and when there is enough volume to keep a specialist occupied. Below that threshold you are paying full-time for part-time work.

The hybrid that usually wins

One in-house owner for strategy, priorities and internal coordination, with external suppliers for content production and link acquisition. This is what most in-house SEO functions become anyway. Planning for it deliberately produces a much better structure than arriving at it by accident.

Rebrands, domain changes and other high-risk events

Certain business events put existing rankings at risk. Each is survivable with planning and each has destroyed organic performance at businesses that treated it as purely a design or legal exercise.

Changing your domain name

The highest-risk event in SEO. Every URL needs a mapped redirect, the change of address tool in Search Console needs to be used, and every significant referring site should be asked to update its link — because redirected links pass less value than direct ones and some referrers eventually drop them.

Expect a dip regardless. A well-executed domain move typically recovers within two to three months. A poorly executed one may not fully recover at all.

Rebranding without changing the domain

Much safer, but entity confusion is a real risk. Your business name changes across the web inconsistently — some directories update, some do not — and search engines have to reconcile two identities. Systematically updating citations, structured data, the Business Profile and social profiles is unglamorous work that materially shortens the confusion period.

Consolidating multiple sites

Frequently the right decision, because authority concentrated in one domain outperforms the same authority split across three. It is also a migration with all the attendant redirect risk, plus a content consolidation problem where two pages on the same subject must become one without losing what either was ranking for.

Redesigns

A redesign that preserves URLs and content is low risk. One that changes URL structure, removes content, or moves content behind JavaScript is a migration wearing a friendlier name. The question to ask your designer is simply whether any URL will change, and if the answer is yes, the redirect map becomes the first deliverable of the project.

The long tail, and why it is where most businesses should start

Head terms are what everybody wants and long-tail terms are where most businesses actually make money in the first two years of a program.

What the long tail actually is

Longer, more specific, lower-volume queries. Individually they bring very little traffic. Collectively they frequently outnumber head-term traffic, and they convert substantially better because specificity indicates a searcher who knows what they want.

Why it is the right starting point

Long-tail terms are reachable without significant authority, which means a new or weak domain can win them within months rather than years. They also compound: pages that satisfy specific queries build the topical depth that eventually makes head terms reachable. Starting at the head with no authority produces a year of effort and no rankings.

How to find them

Search Console is the best source and is routinely ignored — it shows the terms you already receive impressions for, including many you never targeted. Those queries are demonstrated demand where Google already considers you a candidate, which makes them the cheapest opportunities on the site. Beyond that: People Also Ask, autocomplete, and the questions your sales team actually gets asked.

The mistake to avoid

Building a separate thin page for every long-tail variation. Most long-tail queries are variations of one intent and should be served by one comprehensive page. Creating fifty near-duplicate pages produces cannibalization and a quality problem rather than fifty rankings.

How SEO fits with your other marketing

SEO run in isolation from the rest of marketing consistently underperforms, because several of its most valuable inputs come from elsewhere in the business.

Paid search

The fastest way to learn which terms actually convert. Running paid on a term for a month tells you more about its commercial value than any keyword tool, and that intelligence should directly shape which organic terms you prioritize. Paid also covers the gap while organic builds, and can be reduced on terms where organic starts winning.

Public relations

PR and link acquisition are increasingly the same activity approached from two directions. A PR team already has media relationships; an SEO team knows which coverage carries authority value. Where both exist and never speak, the business is paying twice for half the result.

Email and existing customers

Not a ranking channel, but the most reliable source of the reviews, testimonials, case study material and questions that make SEO content genuinely good. Your existing customers are the raw material for the content only you could write.

Social

Not a direct ranking factor. It does distribute content to people who may link to it, and it supports the brand recognition that improves click-through on results you already rank for. Judge it on those terms rather than on any claimed ranking effect.

Sales

The most under-used input in content strategy. Your sales team hears the real objections, the real comparisons buyers make, and the real questions. That is a content plan, delivered free, and almost nobody asks for it.

What ‘quality’ actually means to a search engine

Quality is the word every provider uses and almost nobody defines. Google’s own rater guidelines are public and describe it in some detail, and the description is more concrete than the industry’s usage suggests.

Experience, Expertise, Authoritativeness and Trust

E-E-A-T is not a ranking factor you can optimize directly. It is a description of what the systems are trying to approximate. Experience means first-hand involvement — having actually used the product or done the work. Expertise means genuine knowledge. Authoritativeness means being recognized by others as a source. Trust underpins all three and is the one Google emphasizes most.

How it shows up in practice

• Named authors with real credentials and a genuine professional history, rather than ‘Admin’.
• Content that demonstrates first-hand involvement — specific numbers, specific outcomes, specific mistakes made.
• Clear information about who runs the business, where it is, and how to reach a human.
• Citations to primary sources rather than to other summaries of primary sources.
• Transparency about commercial relationships, including when you sell what you are describing.
• Being described accurately in places search engines already trust.

Your Money or Your Life topics

Content affecting health, finances, safety or major life decisions is held to a materially higher standard. If you operate in medical, legal, financial or safety-related fields, thin or unattributed content is not merely ineffective but actively risky, and content produced without qualified review is a genuine liability as well as an SEO problem.

The practical test

Would a knowledgeable person in your field read this page and find it accurate, useful and clearly written by somebody who knows the subject? That question predicts performance better than any quality score, and it is the standard we hold our own content to.

Content briefs: the document that determines whether content works

The brief is where content quality is actually decided, long before anybody writes. A weak brief produces filler regardless of the writer, and a strong brief produces usable work from a competent generalist.

What a weak brief looks like

A topic, a keyword, and a word count. This tells a writer nothing about what the page needs to accomplish, and the predictable result is a competent-sounding article that answers nothing specifically and ranks for nothing.

What a strong brief contains

1. The intent. What the searcher is trying to accomplish, and what would count as success for them.
2. The format that is winning. What the current top results actually are, so the writer matches the shape Google is rewarding.
3. The specific questions to answer. Drawn from People Also Ask, from Search Console, and from what customers actually ask.
4. The angle only you can take. The proprietary data, the real examples, the genuine opinion that makes this page different from the nine above it.
5. The sources to cite. Primary ones, named.
6. The internal links to include. Specified, not left to the writer’s judgment.
7. What to avoid. Claims you cannot support, competitor comparisons legal will not approve, anything outside what the business actually does.
8. The call to action, and what the reader should do next.

Who should write the brief

Somebody who has read the current results and understands the business. This is genuinely strategic work and is the part of content production most often skipped, because it is invisible in the deliverable. When agency content underperforms, the brief is the most common cause.

Regulated industries and compliance considerations

Some sectors carry constraints that shape what SEO can legitimately do, and a provider unfamiliar with them can create genuine liability rather than merely wasting money.

Healthcare and medical

Medical claims need clinical support and, in many contexts, professional review. Patient testimonials are restricted or prohibited in some jurisdictions and specialties. Before-and-after imagery is regulated. A provider producing medical content without qualified review is creating a compliance problem alongside a quality one.

Legal services

Attorney advertising rules vary by state and cover claims of expertise, results, testimonials and required disclaimers. Case-result content is particularly constrained. This is a market where the compliance review needs to be part of the content workflow rather than a final check.

Financial services

Advice content is regulated, disclaimers are frequently mandatory, and claims about returns are tightly controlled. Content that reads as personalized financial advice can create obligations the business did not intend to take on.

Reviews and testimonials generally

The FTC endorsement guides apply to every industry. Incentivized reviews must be disclosed, testimonials must reflect typical results or say that they do not, and fabricated reviews carry real penalties. Any provider offering to generate reviews for you should be asked precisely how.

Accessibility as a legal matter

In several jurisdictions, website accessibility is a legal requirement and a genuine source of litigation. The overlap with SEO is substantial, which means accessibility work frequently pays for itself twice — once in compliance and once in search performance.

Launching a new site: the SEO checklist

If you are building a site now, doing these things at the point of construction costs almost nothing. Retrofitting them later costs a great deal.

1. Decide the URL structure before you build, and make it stable. Changing URLs later is a migration.
2. Plan the page architecture around intents, not around your internal org chart. One page per intent, no overlaps.
3. Make sure content renders in the raw HTML, not only after JavaScript executes.
4. Set up Search Console and Analytics before launch, not after, so you have a clean baseline.
5. Implement structured data as part of the templates, where it is nearly free, rather than retrofitting it per page.
6. Write real title tags and descriptions for every page, rather than letting the theme generate them.
7. Build the internal linking deliberately, including from the homepage to your commercial pages.
8. Optimize images at upload, with correct dimensions and real alt text.
9. Check that the staging site is not indexable, and that the live one is. Both errors are extremely common.
10. If you are replacing an existing site, map every old URL to a new one before launch, not after.
11. Submit a sitemap containing only canonical, indexable, live URLs.
12. Set up conversion tracking for every inquiry route, including phone calls.

The two mistakes that cause the most damage

Launching with the staging site’s noindex tag still in place, which makes the entire site invisible until somebody notices — occasionally months later. And launching a replacement site without a redirect map, which destroys the accumulated authority of everything that came before. Both are entirely preventable and both are common.

Evaluating an SEO audit somebody has already given you

Many businesses come to us holding an audit from another provider and asking whether it is any good. Here is how to judge one yourself.

Signs it was generated rather than written

• It lists every issue found without distinguishing severity or commercial impact.
• It contains no reference to your specific business, market or competitors.
• It recommends fixes for pages that do not matter commercially.
• It uses a third-party score as the headline finding.
• It has no prioritization, or prioritizes by the tool’s own severity rating.
• It runs to eighty pages and could have been produced in four minutes.

Signs it was done properly

• It names the three things that matter most and explains why those three.
• It connects technical findings to specific commercial pages and revenue.
• It identifies keyword cannibalization, which requires human judgment.
• It says what not to bother fixing.
• It includes rough effort estimates so you can plan resourcing.
• It tells you something about your market you did not already know.

What to do with a bad one

Not nothing. Even a generated audit surfaces genuine technical issues, and the crawl data is real even where the interpretation is absent. Extract the factual findings, discard the prioritization, and get somebody competent to tell you which three matter.

Attribution: proving SEO worked

Attribution is where good SEO programs get canceled, because the default measurement in most businesses systematically undercounts organic search.

Why last-click attribution punishes SEO

A customer finds you through an organic search, reads three pages, leaves, thinks about it for a fortnight, then returns by typing your name into Google and converts. Under last-click, that conversion is credited to direct or branded traffic. The organic page that created the entire relationship receives nothing.

This is not an edge case. For considered purchases it is the normal pattern, and it means the raw conversion numbers in most analytics accounts materially understate what organic contributed.

What to do about it

Look at assisted conversions rather than only last-click. Watch branded search volume, because rising brand searches are frequently a downstream effect of organic discovery. Ask new customers how they found you and record the answer — a single field on the inquiry form produces better attribution data than most analytics configurations.

B2B is harder still

Long buying cycles, multiple stakeholders researching separately, and a purchase decision made offline months after the first visit. A B2B SEO program that is working will frequently look like it is failing under standard measurement for the first two quarters, and it gets canceled for reasons that have nothing to do with its performance.

The realistic approach is to track leading indicators — non-brand impressions and clicks, pages consumed per session, return visits, and inquiry volume — and to accept that revenue attribution will lag by a quarter or more.

The incrementality question

The honest question is not how many conversions organic touched, but how many would not have happened without it. That is genuinely hard to answer and most providers avoid it. A reasonable proxy is to watch what happens to overall inquiry volume as non-brand organic grows — if total inquiries rise roughly in line, the traffic is incremental rather than cannibalizing other channels.

Voice search and AI assistants

Heavily hyped for a decade and rarely worth a separate line in a budget, though the underlying shift is real and worth understanding.

What is actually true

Voice queries are longer, more conversational and more often phrased as questions. They frequently return a single spoken answer rather than a list, which means the winner takes everything and second place gets nothing. For local queries in particular — ‘who’s the nearest plumber open now’ — the answer usually comes from Google Business Profile data.

What to actually do about it

Nothing separate. Answer questions directly and plainly, structure content so answers are easy to extract, keep your Business Profile accurate and current, and use structured data. That is the whole of voice optimization, and every item on that list is something you should be doing anyway.

Assistants and AI answers

The more consequential version of the same shift. When an assistant answers rather than lists, being the cited source is the only position worth having. The qualities that earn a citation are the ones described throughout this guide: plain answers, specific and verifiable claims, original data, clean structure and genuine authority. There is no separate technique.

Multi-location and franchise SEO

A distinct discipline with its own failure modes, and one where the difference between doing it well and doing it at scale-without-thought is enormous.

The central tension

Corporate wants consistency and control; each location needs genuine local differentiation to rank. Resolving this badly produces either a templated page set that gets devalued, or a chaotic collection of location sites competing with each other and with head office.

The architecture that works

A single domain with location pages beneath it, each genuinely written for its market, each with its own Google Business Profile managed properly, all linked from a location index. Authority concentrates in one domain and each location gets a real page rather than a placeholder.

What differentiation actually requires

Real staff names and photographs. Actual projects completed in that area. Genuine local detail — neighborhoods, landmarks, conditions specific to that market. Reviews from customers there. Hours and services that reflect what that location actually does. This is genuinely more work than a template, and it is the entire difference between a page set that ranks and one that gets devalued.

Managing profiles at scale

Each location needs its own Google Business Profile, verified, complete and actively managed. At twenty locations this is a real ongoing job. Businesses that set them up once and never return lose steadily to competitors who post, respond to reviews and keep their information current.

The mistake worth repeating

Do not launch two hundred templated pages. Build ten genuinely differentiated ones for your most valuable markets, measure them over two quarters, and extend only if they work. The pattern of near-identical pages at scale is one of the clearest quality signals there is, and it is assessed site-wide rather than page by page.

Putting it together: how to decide what to buy

If you have read this far, the decision comes down to a short sequence of honest answers.

First, establish whether search demand exists for what you sell

Check the actual monthly volumes for the terms your customers would use. Most markets have more demand than businesses assume, but some genuinely do not, and finding out costs an hour.

Second, work out which of the six disciplines your business actually needs

Local trades and restaurants need Google Business Profile work and reviews. B2B and SaaS need content depth. Ecommerce needs technical work at scale and category content. Buying an even split across all six usually means nobody diagnosed anything.

Third, be honest about budget and commitment

Below roughly $750 a month, nobody can staff your account meaningfully. Below twelve months of commitment, the program will be canceled before it pays. If either constraint binds, spend the money on paid search instead and revisit when it does not.

Fourth, decide who executes

If you have no internal capacity, an agency retainer. If you have a focused problem, a freelancer or a project. If you have a capable team who need direction, a consultant. If organic is your primary channel and the workload justifies it, an in-house hire supported by external content and outreach.

Fifth, choose the provider on their answers, not their deck

Which specific pages in month one. Who exactly does the work. What they report and against what metric. And what happened at a client where it did not work. The provider who answers that last one candidly is usually the right choice.

Sixth, agree what six months looks like before you start

Write it down. Then the month-four review is a comparison against a plan rather than a judgment made on instinct, and you avoid the single most common way SEO money gets wasted.

Common questions we get asked on first calls

These come up so consistently that they are worth answering in advance, and the answers are the ones we actually give rather than the ones that would close more business.

‘Our competitor ranks first and their site is terrible. How?’

Almost always authority accumulated over time, frequently combined with better intent matching than the site’s appearance suggests. A dated-looking site with ten years of earned links and a page that answers the query directly will beat a beautiful new site that does neither. Design is a conversion lever, not a ranking one.

‘Can you get us to number one for [very competitive term]?’

Possibly, over a period measured in years, if the term is genuinely reachable given your authority and if you fund the work consistently. The more useful question is whether that term is worth the investment compared with ten less contested terms that collectively bring more qualified traffic. Frequently it is not.

‘We tried SEO before and it did not work.’

The most common thing we hear, and in the large majority of cases the diagnosis is one of four: it was canceled around month four, the content was too thin to compete, the wrong discipline was bought for the business type, or the measurement was broken and it may actually have been working. Establishing which one it was is worth doing before spending again.

‘How much traffic will we get?’

Anybody answering that with a number is guessing. What can be estimated honestly is the total search demand for the terms you could realistically compete for, and a plausible share of it given your starting authority. That produces a range with stated assumptions, which is a genuinely useful planning input and is not the same thing as a promise.

‘Do we need to be on page one to get anything?’

Broadly yes for clicks, though the picture is more nuanced now. Positions one to three take the overwhelming majority of organic clicks, and the presence of AI Overviews and other SERP features has reduced click share further. But impressions at positions eleven to twenty are a genuine leading indicator, and moving those pages up is the cheapest work available.

‘Should we buy an expired domain with existing authority?’

Almost always no. The authority frequently does not transfer as advertised, the link profile often carries risk you inherit, and the topical mismatch between the old site and yours undermines whatever value remains. It is sold as a shortcut and is usually an expensive detour.

‘What about all these emails offering guaranteed rankings?’

Delete them. Unsolicited SEO email is the industry’s most reliable negative signal: no competent provider needs to cold-email strangers with ranking guarantees, and the services offered are almost invariably purchased links or automated submissions.

A short history of why this industry is distrusted

Worth understanding, because it explains a great deal about how SEO is sold and why buyers are reasonably skeptical.

The manipulation era

For roughly the first decade, search engines were genuinely easy to manipulate. Keyword stuffing, hidden text, link farms and article spinning all worked. An entire industry formed around techniques that had nothing to do with serving customers, and a generation of businesses bought results that later evaporated.

The correction

A sequence of algorithm updates over several years devalued most of it, frequently overnight and frequently with no warning. Businesses that had ranked well for years lost their traffic, and the providers who had sold them the tactics generally had no answer. That experience is the origin of most of the skepticism SEO providers still encounter.

Why the bad practices persist

Because the work is difficult for a buyer to evaluate. You cannot easily tell whether a technical recommendation is correct or whether a link was earned or purchased. Markets where quality is hard to observe reward confident presentation, which is precisely the dynamic that keeps poor providers in business.

What this means for you as a buyer

It means process transparency matters more than promises. A provider who shows you what they are doing and why, reports honestly including when things fail, and declines to guarantee outcomes they cannot control, is demonstrating the only signal available that reliably distinguishes competence. Everything else can be presented convincingly by anybody.

Related guides on this site

This page is the overview. These go deeper on individual pieces of it.

What we would change about how SEO is sold

A closing section that is not about buying SEO but about the industry selling it, because the two are connected and because you are entitled to know what we think the problems are.

Price opacity does more harm than good

Most agencies refuse to publish any pricing, on the theory that scope varies and the conversation belongs on a call. Scope genuinely does vary. But the practical effect is that a business trying to budget cannot, and the information asymmetry consistently favors the seller. We publish ranges because we would rather lose inquiries from people we are too expensive for than have that conversation in week three.

Reporting is designed to reassure rather than inform

The standard monthly SEO report leads with rankings and total traffic because both usually look positive. Neither answers the question the client is actually asking, which is whether this is producing business. Reporting that leads with non-brand traffic and attributable inquiries is harder to produce and occasionally uncomfortable, and it is the only kind worth paying for.

The industry over-sells novelty

Every few years a new acronym arrives and is sold as a distinct service requiring a distinct budget. Mobile-first, voice, and now AI search have each been through this cycle. In each case the underlying shift was real and the correct response was a change of emphasis within existing work rather than a second retainer. Treat novelty pricing with suspicion.

Too much SEO is sold to businesses that should not buy it

The small local business with no Google Business Profile, no reviews and no service pages does not need an agency. It needs twenty hours of straightforward work it could largely do itself, and selling it a retainer instead is taking money for a service that will underperform something free. We say this on a first call more often than is commercially sensible, and it is the right thing to say.

The good news

Because so much of the industry operates this way, the bar for being genuinely good is lower than it should be. A provider who diagnoses before prescribing, reports honestly, publishes their pricing and declines work they cannot help with stands out immediately. That is a fairly low standard and it is not widely met.

Summary: what you are actually buying

To close, the shortest honest version of everything above.

• SEO services are six different disciplines under one name, and you probably need two or three of them rather than all six.
• The sequence matters: technical, then on-page, then content, then authority. Providers who start with content are optimizing for looking busy.
• Content is usually the largest cost and the largest driver, at roughly forty percent of a working budget.
• Below about $750 a month nobody can staff your account meaningfully, and below twelve months the program will be canceled before it pays.
• Nobody can guarantee a ranking, and any guarantee is either meaningless or dishonest.
• The most reliable way to value the outcome is to price your organic traffic at what the same clicks would cost in paid search.
• The best single question for a prospective provider is to ask about a client where it did not work.
• A meaningful amount of the highest-value work — Business Profile, reviews, honest service pages, title tags, internal links — you can genuinely do yourself.
• Judge the program on non-brand traffic and attributable inquiries, never on average position.
• Agree what six months looks like before you start, in writing. It prevents the most common way SEO money gets wasted.

Two things worth knowing about how Google communicates

A recurring source of confusion in this industry is what Google actually says versus what gets reported that it said. Being able to check the source yourself is genuinely useful when a provider makes a claim.

Where the authoritative information lives

Google Search Central publishes documentation, a blog, and the video channel embedded throughout this guide. Between them these cover almost every question about how search works, and they are free. When a provider tells you Google requires something, it is reasonable to ask where that is documented — the answer should be a link, not a reference to an industry blog post about a conference talk.

Why third-hand reporting distorts things

A Google representative answers a nuanced question in an office-hours session. The answer is summarized on an industry site, then summarized again in a newsletter, and by the fourth iteration a qualified ‘it depends, and usually not’ has become a definitive rule. A meaningful share of confidently stated SEO advice originates this way, which is why checking the primary source is worth the two minutes it takes.

What Google does not tell you

The weights of ranking signals, the specifics of algorithm updates, and anything precise enough to be gamed. This is not evasiveness so much as a structural necessity — a published ranking formula would be optimized against immediately and would stop measuring quality. It does mean that anybody claiming precise knowledge of the algorithm is describing an inference, and should say so.

A note on this page

This guide is roughly twenty-five thousand words because the subject genuinely covers six disciplines and we would rather answer the question properly than produce something that ranks and disappoints. A few things about how it was made are worth stating plainly.

Every price band in it reflects what we see in the market and what we ourselves charge, rather than an industry survey we did not run. The technical audit percentages come from audits we have conducted. Where we do not know something precisely — the weight of a ranking factor, the exact effect of an update — the page says so rather than presenting a confident guess.

We also sell the services described here, and that is a conflict worth naming. Our response to it is to include the sections most agencies leave out: what SEO cannot do, when you should not buy it, how to evaluate an audit somebody else produced, and how to do a meaningful amount of this work yourself without paying anybody. If that costs us some inquiries from businesses who would not have been well served, that is the correct trade.

The page is reviewed and updated as the market and the platform change. Where a figure is dated, the date is stated. If something here is wrong or has gone stale, we would genuinely like to know.

How to check a claim yourself in five minutes

If a provider tells you something about how search works and you want to verify it, the sequence is short. Search Google Search Central documentation for the term. If it is not there, check the Search Central blog. If it is in neither, the claim is industry inference rather than documented behavior, which does not automatically make it wrong but does change how much weight it deserves.

This matters more than it sounds, because the gap between documented behavior and industry folklore is where most bad SEO advice lives. A claim that Google penalizes something, or requires something, or weights something heavily, is either in the documentation or it is somebody’s inference from observation. Both can be useful. Only one of them is a fact, and a provider who does not distinguish between them is not being careful with your money.

The same check applies to claims about your own site. If somebody says your pages are not indexed, open Search Console and look. If somebody says your site is slow, run it through PageSpeed Insights yourself. Nearly every factual claim in an SEO proposal is verifiable in a few minutes with free tools, and providers who make claims that do not survive that check are telling you something important about how they work.

A final practical point about using this guide. If you are mid-way through an engagement and something here contradicts what you are being told, that is worth raising rather than assuming one of us is wrong — your provider may have a good situation-specific reason, and hearing it will tell you a lot about how carefully they are thinking. If they cannot give one, that also tells you something. Either way the conversation is more useful than quietly losing confidence.

The design brief and the research plan behind it

A design brief is the document that tells a designer what the work has to achieve, for whom, under what constraints. Design brief meaning in practice: audience, objective, scope, mandatories, constraints, budget, timeline and how success will be judged. What is in a design brief that separates a good one from a wish list is the constraints section, because that is where the real decisions are.

The meaning of design brief is often confused with a creative concept. The brief states the problem; the concept is the answer. How to create a design brief that produces good work: write it with the people who will approve the result, name a single decision maker, state what is fixed and what is open, and include the evidence you already have. Write the design brief before any visual is produced, because a brief written after the fact is a justification.

Research and planning underneath the brief is what makes it more than opinion. How to make a research plan: state the decision the research will inform, the questions that would change it, the method for each, who you will speak to, and when you will stop. Research planning definition in one line: deciding in advance what you are trying to learn, so that you notice when you have learned it.

How to write research plan documents that get used means keeping them to a page and tying every method to a decision. Research with no decision attached produces a report nobody acts on, which is the most common way research budgets are wasted.

Slogans, messaging, and brand pillars

What is a business slogan: a short, repeatable phrase that states or evokes what the business promises, used consistently enough to be remembered. Why are slogans important is answered by recognition rather than persuasion — a slogan rarely convinces anyone of anything, and a good one makes the brand identifiable in a fraction of a second.

Why slogans are important commercially comes down to consistency over time. A phrase changed every campaign never becomes an asset; one held for a decade becomes shorthand for the company. The test of a slogan is whether a competitor could credibly use it: if they could, it says nothing.

Core pillars are the three or four things a brand consistently talks about, and they exist to stop a content program from drifting. A core pillar is not a topic list; it is a claim the business is prepared to defend, with evidence, repeatedly. Content that does not sit under a pillar is content that will not compound.

Voice search and how people ask questions

Voice search optimization is the practice of making content answerable when the query is spoken rather than typed. What is voice search optimization in substance: spoken queries are longer, more conversational, more often phrased as questions, and frequently local. Optimizing for them is largely the same work as optimizing for AI answers, because both need a plainly stated answer that can be extracted.

How to optimize for voice search, concretely: write the question as a heading in the words people say, answer it in the first sentence beneath, keep the answer to a length that can be read aloud, mark up the page with structured data, and make sure the business’s location and hours are accurate everywhere. How to optimize content for voice search is not a separate content strategy; it is a discipline about where the answer sits on the page.

Voice search for business matters most where the query is local and the intent is immediate — hours, directions, availability, phone number. Voice search engines are not a separate index: assistants draw on the same web results, on business profiles, and increasingly on generative answers, which is why the underlying work is ordinary search work done more precisely.

Local search maintenance: profiles, citations, and audits

Add me to Google, and the variants add me on Google, how to add me to Google and add yourself to Google search, are all asking the same thing: how does a person or business appear in results. For a business the answer is a Google Business Profile, which is free, requires verification, and is what produces the panel with hours, reviews and directions. For an individual, Google add me to search refers to the people-cards feature, which is available only in certain countries and is created from a signed-in mobile search.

A NAP audit checks that a business’s name, address and phone number are identical everywhere they appear online. NAP audits are tedious and they matter, because inconsistency creates duplicate entities in the systems that assemble local results. The method is a fixed canonical version of each field, a list of every directory the business appears in, and a correction pass.

Local result marketing is the practice built around all of this: profile completeness, categories, service areas, reviews, photographs, posts and citation consistency. It is the highest-return work available to a business whose customers are within driving distance.

Measurement, position tracking, and migration

Check position Google searches usually come from someone who has noticed their own results are personalized and wants a neutral reading. A search engine position tool solves this by querying from a defined location and device without a signed-in profile, which is the only way to get a comparable number over time. Checking manually in a private window is better than checking while signed in and still not neutral, because location still applies.

Google bounce rate is a metric that has changed meaning. In the older analytics model a bounce was a session with a single interaction; in the current model the reported metric is the inverse of engaged sessions, where an engaged session lasts beyond a threshold, has a conversion, or has multiple page views. The practical consequence is that bounce figures from before and after the change are not comparable, and a high figure on a page that answers a question completely is not a fault.

A topical authority map is the planning artifact that sets out every subtopic within a subject area, which of them you cover, and how well. It is how a content program decides what to write next on evidence rather than on enthusiasm, and it is the same artifact that shows when a topic is finished.

Web migration — moving a site to a new platform, domain or structure — is where accumulated search value is most often destroyed. The non-negotiables are a full inventory of existing URLs, a redirect map from every old URL to its closest equivalent, preservation of titles and content on pages that already rank, and monitoring for four to eight weeks afterward.

AI in production and in trust signals

AI generated trailers have become common enough to have their own audience, and the terminology is unsettled: an artificial trailer usually means a synthetic concept trailer for a film that does not exist, while trailer AI more often refers to tools that assemble a real trailer from existing footage. Both raise the same disclosure question, and platforms have begun requiring synthetic-media labeling.

Trust UGC — the question of whether user-generated content can be trusted — has sharpened for the same reason. Reviews, testimonials and social posts have always been forgeable and are now cheap to forge at scale, which raises the value of the signals that are hard to fake: named reviewers with histories, verified purchases, photographs of the actual work, and platforms that filter aggressively.

Technical SEO audits, and what a URL’s length actually costs

One service line and one persistent myth.

A technical SEO audit service produces a prioritized list of the faults preventing a site from being crawled, indexed and understood. What separates a useful audit from a crawler export is the prioritization: every crawler produces hundreds of issues sorted by its own severity model, and the value is in reducing that to the faults affecting pages which already receive traffic or which you intend to rank. An audit delivered as an unfiltered tool report is a work order nobody can act on.

The scope that earns the fee covers crawlability and indexation, site architecture and internal linking, page performance measured on real visits, structured data, and the redirect and canonical logic that decides which version of a page competes. Delivered properly it comes with the faults grouped by who fixes them, because a list mixing a developer task with a content task stalls at the first handoff.

Length of URL SEO is the myth attached to this. URL length is not a ranking factor in any direct sense; search engines have said as much repeatedly. What length correlates with is depth and clarity, and those do matter: a URL five folders deep usually indicates a page far from the homepage in the link structure, and a URL stuffed with words is usually a page whose title was pasted into the path. A short, readable, stable URL is worth having because people copy and share it, not because a shorter string ranks better.

The related and genuinely consequential rule is that changing a URL costs you whatever it had accumulated unless a permanent redirect carries it over. Which is why the correct time to get URL structure right is before publication, and the correct time to change it is almost never.

Frequently asked questions