B2B SaaS Marketing Agency: What Actually Builds Pipeline

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

An eleven-month cycle, a buying committee of six to ten people, and a CAC payback period that decides whether the business can afford to grow. This is what B2B SaaS marketing genuinely involves — the content that produces pipeline, the metrics that mislead, the difference between product-led and sales-led motions, and how to judge an agency.

The short answer

B2B SaaS has more marketing tooling, more published advice and more agencies than any other category, and the widest gap between what gets measured and what produces revenue. Most of that gap comes from measuring a one-month metric against an eleven-month cycle.

Who this is for

B2B software companies between roughly $1m and $50m ARR, product-led, sales-led or hybrid. The motion matters enormously and is addressed directly rather than assumed.

The two ratios everything answers to

Lifetime value against acquisition cost, and how long that acquisition cost takes to come back. A channel with a great LTV to CAC ratio and a twenty-four month payback can still be unaffordable, and a channel with a modest ratio and a four-month payback can be exactly right.

Where SaaS pipeline actually comes from

Organic content is the largest source at efficient companies

Around a quarter of qualified pipeline, and considerably more at product-led companies. It also compounds, which nothing else on the list does. The catch is six to twelve months before it contributes, which is longer than most marketing leaders are given.

Product-led signups where a free tier exists

The most efficient acquisition available, with CAC payback measured in months rather than years. It also depends entirely on decisions made by product rather than by marketing, which is a recurring organizational problem.

Referral closes better than anything

Roughly one in seven deals, closing faster and negotiating less. It cannot be switched on, which is why companies over-invest in the channels that can and why referral stays unmanaged.

Outbound is predictable and getting harder

Response rates have fallen for years and cost per meeting has risen. It still works when targeting is genuinely narrow and you have something specific to say, and it fails as a substitute for differentiation.

Paid search is high intent and extremely expensive

Category terms in established SaaS categories regularly exceed fifty dollars a click. It works, and it works far better when the landing experience answers the question rather than demanding a demo.

Review sites are the shortlisting layer

Where a field of twenty becomes a field of four. Placement is partly paid; review volume and rating are not, and those are what buyers read.

Partnerships and integrations are under-invested

Slow, unglamorous, and among the most durable pipeline sources in software. Every integration is a page, a co-marketing opportunity and a search term nobody is competing for.

CAC payback, which decides what you can afford

Under twelve months is comfortable

You recover acquisition cost within the first year of the relationship, which means growth is largely self-funding. Product-led and referral acquisition usually sit here.

Twelve to eighteen months is workable

Most content and organic acquisition lands in this band. It requires either capital or patience, and it is the normal condition for a healthy B2B SaaS business.

Beyond eighteen months strains cash flow

Outbound, events and paid social frequently sit here. They are not necessarily wrong; they are expensive in a specific way that shows up as a cash problem rather than as a performance problem.

Payback and LTV to CAC are different questions

A channel can have an excellent lifetime ratio and an unaffordable payback period. Companies that optimize only for the ratio frequently run out of money while being technically efficient.

Contract length changes everything

Annual prepaid contracts shorten payback dramatically against monthly billing at the same ACV. That is a pricing decision with a direct marketing consequence and it is rarely discussed as one.

Measure it by channel, not in aggregate

A blended payback figure hides the fact that one channel is funding the business and another is consuming it. This is the single most useful segmentation in SaaS marketing measurement.

The go-to-market motion decides everything else

Product-led marketing lives on organic acquisition

Self-serve signup means the marketing site is the sales team. Search visibility, documentation, onboarding content and free-tier design carry the entire top of funnel, and SEO leverage is higher here than in any other business model.

Sales-led marketing lives on enablement and quality

Fewer, better-qualified opportunities and a sales team that needs materials for six different committee members. Volume metrics are actively misleading in this motion.

Hybrid is the most common and the least well served

The usual failure is optimizing the whole site for the motion producing less revenue, because it is the one the team finds more interesting.

The free tier is a marketing decision

What it includes, where it stops, and how obvious the upgrade moment is. In product-led companies this is the most consequential acquisition decision made, and it is usually made without marketing in the room.

Sales cycle length differs by an order of magnitude

Days for self-serve, quarters or years for enterprise. Reporting them together produces an average describing neither and hides which motion is actually growing.

An agency should ask which one you are, first

Before anything else. An agency that proposes the same plan regardless has one plan.

The content that produces pipeline

Comparison pages are the highest-intent traffic available

Somebody searching ‘your product versus competitor’ has narrowed the field to two. They will read a comparison from somebody, and it is usually your competitor’s. Writing an honest one — including where the other product is genuinely better — outperforms every other page type we have measured.

Alternatives pages

‘[Competitor] alternatives’ is searched by people actively dissatisfied with a product you compete with. High intent, low competition from the incumbent, and consistently under-built.

The pricing page is the most visited and worst executed

Buyers go there first and frequently leave. ‘Contact us for pricing’ removes you from consideration for a meaningful share of evaluations, and ‘starting from’ figures with no context do nearly as much harm.

Publish the shape of your pricing at minimum

What drives the price, what a typical customer of a given size pays, what is included. Buyers building a business case cannot proceed without a number and will source one from an analyst or a review site instead.

Integration pages, one per integration

Specific, low competition, and searched by people already using the other product. Most SaaS companies have one combined integrations page listing logos.

Use-case pages by role and by industry

Buyers self-identify by role and vertical rather than by feature. ‘For finance teams’ and ‘for healthcare’ pages convert better than feature pages because they answer ‘is this for someone like me’ first.

Public documentation is an SEO asset

Indexable documentation ranks unusually well on long-tail technical queries, reaches evaluators directly, and builds real trust. Companies that gate documentation forfeit all of it.

Security and compliance content

SOC 2, data handling, subprocessors, uptime, access controls. Publishing this openly shortens enterprise deals measurably and almost nobody does it, because it is nobody’s favorite thing to write.

Customer stories need numbers

A specific figure, a named role and an honest account of what was difficult. Anonymous, number-free case studies are the most common and least useful format in B2B software.

What does not work: gated ebooks

They produce MQLs and very little pipeline. The gate costs you the readers who would have shared it and collects details from people who wanted the PDF.

Review sites and where shortlists form

This is the reduction stage

G2, Capterra, TrustRadius and category equivalents are where a field of twenty becomes a field of four. Vendors with thin review profiles do not survive that reduction regardless of product quality.

Placement is partly paid, presence is not

You can pay for position and category sponsorship. You cannot pay for review volume or rating, and those are what buyers actually read.

Review volume is a process

Ask at the right moment — after successful onboarding, after a support interaction that went well, after a renewal. Quarterly pushes produce a spike that looks exactly like what it is.

Category placement matters and shifts

Being in the right category with the right competitors affects who you are compared against. Worth reviewing annually because these categories change.

Respond to critical reviews specifically

Without defensiveness, and mention what changed if something did. Prospective buyers read responses more carefully than reviews.

Cloud marketplaces are under-used

AWS, Azure and Google Cloud marketplaces let buyers purchase against existing committed spend, which removes procurement friction entirely. Mid-market SaaS companies frequently overlook this.

Metrics, and the ones that mislead

Stop leading with MQLs

The metric most agencies report and the one least connected to revenue at this cycle length. It is trivially gameable with a gated PDF and it tells a sales team nothing about whether to call anybody.

Pipeline created is better and still early

The right direction, and still a leading indicator eight to eleven months ahead of revenue.

Closed-won by first touch is the honest number

Difficult to measure, lagged by nearly a year, and the only figure that tells you whether the program worked. Set it up now so that next year you can answer the question.

Win rate by source is the underused metric

Some sources convert poorly and close superbly. Referral and organic frequently do. Judging channels on conversion to opportunity alone systematically undervalues them.

Sales cycle length is a content metric

One of the clearest effects good B2B content has is shortening the cycle, because buyers arrive having answered more of their own questions. Almost nobody tracks it as a marketing outcome.

Attribution is directional, not truth

With six to ten people, eleven months and most of the process invisible, no model is accurate. Use it to compare relative movement over time and treat any model claiming precision with suspicion.

Net revenue retention is a marketing metric too

Expansion depends on adoption, which depends on onboarding content, documentation and enablement. Marketing that stops at closed-won leaves the more profitable half of the relationship unattended.

What it costs

The most common misallocation

Companies under $5m ARR buying multi-channel plans that spread a small budget across six channels, none of which reaches useful volume. One channel done properly beats six done adequately below roughly $10m in revenue.

Agency cost for B2B SaaS

$5,000 to $25,000 a month for a mid-market company, excluding media. Below about $4,000 you are buying execution on a narrow scope, which is sometimes exactly right.

When not to hire an agency at all

Before product-market fit. Marketing cannot manufacture demand for something people do not yet want, and the money is better spent on the conversations that establish whether they will.

In-house versus agency

The scarce skill is writing accurately about a technical product. That is harder to outsource than most companies expect, and it is why the best arrangement is frequently one strong internal writer plus outside technical SEO and paid media.

How to judge a B2B SaaS marketing agency

Ask which motion they think you run

If they cannot distinguish product-led from sales-led, or do not ask, they are proposing a plan for a business they have not identified.

Ask about your ACV, cycle length and win rate

An agency quoting without these three numbers is pricing from a template. All three change what the correct plan looks like.

Ask what they would build first

Comparison pages and pricing content in most cases. An ebook is the wrong answer.

Ask how they will report

Pipeline created, win rate by source, and eventually closed-won by first touch. If MQLs lead the report, the report is optimized for the agency.

Ask what they would tell you not to do

Everybody competent has a list. In SaaS it usually includes gated ebooks, undifferentiated thought leadership, and expecting pipeline inside a quarter.

Ask who writes the content

Technical content written by somebody who has not used the product is confidently wrong in ways your evaluators will notice immediately.

Technical SEO for a SaaS site

The documentation subdomain question

The single most common structural decision costing SaaS companies organic traffic. A subfolder consolidates authority; a subdomain splits it. Where the platform makes a subfolder impossible, optimize both deliberately rather than ignoring one.

Rendering trips up modern stacks

Marketing sites built as single-page applications frequently render content Google sees late or not at all. Invisible in a browser, obvious in Search Console.

Programmatic pages, carefully

Generating a page per integration or use case is legitimate when each contains genuinely different substance. Hundreds differing by a variable are assessed as a pattern and can affect the whole site.

Sales enablement, which is also marketing

The materials sales actually uses

A one-page comparison, a security summary, a business case template and three customer stories with numbers. That covers most of what a B2B SaaS sales team asks for, and most marketing departments produce something else.

Test whether it is used

The most useful enablement metric is whether the sales team sends what you made. It is measurable, uncomfortable, and predicts revenue contribution better than content volume.

Objection handling belongs on the website

Every objection your sales team hears repeatedly is a question buyers have before they contact you. Publishing the answers shortens the cycle and improves the conversations that do happen.

The champion is the most under-served person

Somebody inside the company wants to buy and has to persuade five colleagues. A business case template they can forward is the most leveraged content in B2B SaaS and almost nobody builds it.

Losing well

Sales teams learn more from lost deals than won ones and marketing rarely hears any of it. A monthly fifteen minutes on why deals were lost redirects content priorities better than any keyword research.

Competitive intelligence, kept current

Competitor pricing and positioning move quarterly. A comparison page written two years ago is wrong, and being caught being wrong about a competitor costs more credibility than the page earned.

Content mapped to the buying committee

Six to ten people, each deciding something different. Most SaaS content addresses one of them.

The economic buyer

Deciding whether the business case holds. They need pricing, an honest ROI framing and customer stories with real numbers in them. Most SaaS sites give them adjectives.

The technical evaluator

Deciding whether it will work with the existing stack. They want documentation, integration detail, API reference and architecture information, and they want it without talking to anybody.

The end user

Deciding whether this makes their job easier or harder. Use-case pages by role, product tours and onboarding content reach them, and almost nothing on a typical SaaS site is written for them at all.

The security reviewer

Deciding whether it exposes the company. SOC 2 status, subprocessors, data residency, access controls and incident history. Publishing this openly is the single most under-used content opportunity in enterprise SaaS.

Legal and procurement

Deciding whether the terms are acceptable. Standard contract terms, data processing agreement, insurance, SLA. Deals die here more often than in product evaluation.

The champion

The person inside the company who wants to buy it and has to persuade five colleagues. A business case template, an honest comparison and a security summary they can forward is the most leveraged content in B2B SaaS and almost nobody builds it.

Existing users, for expansion

Net revenue retention depends on adoption. Documentation, changelog and advanced use cases are marketing work in most companies and assigned to nobody in many.

Analysts and advisers

Deciding whether to recommend you. Clear category positioning, differentiation that survives a feature grid, and proof points they can cite.

What to do at each stage of company

The right plan at $1m ARR is not a smaller version of the right plan at $20m.

Under $1m ARR

One channel, done properly, plus comparison and pricing content. Founder-led writing usually outperforms anything outsourced at this stage because the founder knows what customers actually say.

$1m to $5m ARR

Content depth, review site presence and the first paid search experiments. This is where compounding begins and where most companies mistakenly spread budget across six channels.

$5m to $20m ARR

A full content program, paid media, sales enablement and partnership work. Also where the first genuine hiring decisions about in-house marketing happen.

$20m ARR and above

Category building, account-based marketing with a real target list, brand investment and partnerships. The risk here is abandoning the content and referral discipline that produced the growth.

Before product-market fit

No agency, no retainer, no content program. Marketing cannot manufacture demand for something people do not yet want, and the money is better spent on conversations that establish whether they will.

At every stage

Comparison pages and pricing transparency. They are the two highest-return content types at $1m and at $50m, and they are blocked internally at both.

Objections we hear, answered

Every one of these has been said to us, and two of them are partly right.

‘Our product is too complex to explain simply’

Then your buyers cannot explain it internally either, which is why deals stall at the champion stage. Simplification is not dumbing down; it is the work.

‘We cannot publish pricing, every deal is different’

Publish the shape: what drives price, what a customer of a given size pays, what is included. Buyers will find a number somewhere and it will be less accurate than yours.

‘Comparison pages will help our competitors’

Your competitors already know your product. The comparison is being written by somebody; the only question is whether the version buyers read is yours.

‘Content takes too long’

It does, and it is the only channel that compounds. The correct response to a long payback is starting earlier, not choosing a channel that never gets cheaper.

‘Our buyers do not search’

Somebody at the buying company does — usually the champion or the technical evaluator, weeks before you are contacted. That is precisely the stage you are absent from.

‘We tried SEO and it did not work’

Ask what was published. Almost every unsuccessful SaaS content program we review produced thought leadership and no comparison pages, no pricing content and no integration pages.

‘Our category is too new’

Then problem-led content reaches people who do not know the category exists, which is a larger audience than the category term has. Category creation is expensive; problem content is not.

‘We need pipeline this quarter’

Then paid search and outbound, and be honest that both cost more per booked customer and neither improves. Content is the answer to next year, not to this quarter.

Writing comparison pages properly

The highest-return content in B2B SaaS, and the easiest to do badly.

Write the comparison you would want to read

Honest about where the competitor is better, specific about where you are, and free of the tells that make comparison pages read as marketing. Buyers can spot a rigged grid instantly and it costs more credibility than the concession costs deals.

Use a real feature grid, and lose some rows

A grid where you win every row is not believed. A grid where you lose three and win seven, with an explanation of who should choose which, is believed and converts better.

Say who should choose them

‘If you need X, they are the better choice’ is the most persuasive sentence on a comparison page. It also filters out the customers who would have churned.

Keep pricing current on both sides

Being wrong about a competitor’s pricing is the fastest way to lose the reader. Set a quarterly review, because their pricing moves.

Address migration explicitly

Somebody comparing is somebody considering a switch. What migration involves, how long it takes and what you do to help is the practical question underneath the comparison.

Do not disparage

Comparative advertising has legal limits and reputational ones. Accuracy is both safer and more persuasive than criticism.

Build alternatives pages separately

‘[Competitor] alternatives’ is a different search from ‘[you] vs [competitor]’ and deserves its own page addressing somebody actively dissatisfied rather than somebody choosing.

Update after every competitor release

A comparison page that is six months stale is a liability. This is ongoing work rather than a project.

Security and compliance content

Where enterprise deals actually stall, and all of it is publishable.

What to publish about security

SOC 2 status and report availability, subprocessor list, data residency options, encryption at rest and in transit, access controls, SSO and SCIM support, uptime history and incident communication policy.

Why publishing it shortens deals

Every item above appears in a security questionnaire. Publishing it means the reviewer can self-serve most of the questionnaire, which removes weeks from the enterprise cycle.

It also filters

Companies whose requirements you cannot meet find out early rather than in month three. That is a saving for both sides and it protects your win rate figures from deals that were never winnable.

A trust center beats a PDF

A maintained page, not an emailed document. It ranks, it is linkable by your champion, and it does not go stale in somebody’s inbox.

Compliance roadmap, honestly stated

If you are pursuing SOC 2 or ISO 27001 but not there yet, say so with a timeline. Buyers respect a stated roadmap and distrust silence.

Do not overstate

Claiming certifications you do not hold is a straightforward path to a failed deal and a damaged reputation with an auditor-adjacent buyer.

Integrations and partnerships

Slow, unglamorous, and among the most durable pipeline sources in software.

Integrations are search terms nobody contests

‘[Your product] Salesforce integration’ has genuine search volume and almost no competition, because the only two companies who could write it usually do not.

One page per integration, with substance

What it does, what it syncs, what it does not, how long setup takes, and what breaks. A logo grid is not an integration page.

Co-marketing is the easy ask

Partners generally want integration content too. A joint page, a shared announcement or a listing in their marketplace costs nothing and reaches an audience already using the other product.

Marketplace listings are underrated

Salesforce AppExchange, HubSpot, Shopify, Atlassian and cloud marketplaces all drive genuine discovery, and listings are frequently thin or absent for mid-market SaaS.

Integrations shape retention too

A customer with three integrations configured is substantially harder to displace than one with none. Driving integration adoption is a retention activity as much as an acquisition one.

Prioritize by what customers already use

Not by partner size. The integration your existing customers keep asking for is worth more than the one with the biggest logo.

Positioning, and why most SaaS marketing fails before it starts

The single most common reason a B2B software company’s marketing underperforms has nothing to do with channels. It is that nobody can say, in one sentence, who the product is for and what it replaces. Marketing then spends its budget describing a product to an audience that has not been defined, and every channel underperforms simultaneously, which is usually misread as a channel problem.

A workable position answers three questions without hedging. Who specifically has this problem badly enough to pay. What they are doing today instead, whether that is a competitor, a spreadsheet, an internal build or nothing at all. And what changes for them once the problem is solved. If the honest answer to the second question is ‘a spreadsheet’ or ‘nothing’, the entire content strategy is different from one where the answer is a named competitor, because there is no existing search demand to capture.

The spreadsheet competitor

When the real alternative is a manual process, nobody is searching for your category. Comparison pages have nothing to compare against and category terms have no volume. What works instead is problem-first content: the symptoms of the manual process, its hidden costs, and what the process looks like once automated. This is slower and less measurable than capturing existing demand, and it is the correct approach when existing demand does not exist.

The named competitor

When buyers already know the category, demand capture beats demand creation by a wide margin. Comparison pages, alternatives pages and integration pages reach people already in the market. The mistake here is inverted: companies with abundant existing demand spend on thought leadership that creates demand they already have, while leaving the high-intent pages unbuilt.

Positioning is testable

Ask five recent customers why they bought. If the answers cluster, the position is real. If they scatter across five unrelated reasons, marketing is going to keep producing inconsistent results no matter how it is executed, and no agency can fix that from the outside.

Free trial, freemium and demo request: choosing the conversion event

The conversion event a SaaS site optimizes for determines the entire funnel behind it, and it is chosen far more casually than it deserves. Three models dominate, and they are not interchangeable.

Time to value is the deciding variable

If a new user can reach a genuinely useful outcome inside a session, a free trial without a credit card will outperform a demo gate substantially. If reaching value requires importing data, connecting systems or configuring permissions, a trial produces a stream of users who sign up, fail to reach value, and churn silently, which looks like a marketing volume success and a revenue failure.

The card-required question

Requiring a card at trial start reduces signup volume sharply and raises trial-to-paid conversion sharply, and the two effects do not reliably cancel. The only honest way to decide is to run it, which requires enough volume to reach significance. Below roughly a few hundred trials a month, this test cannot be resolved and the decision should be made on operational grounds instead.

Category creation, and when it is an expensive mistake

Creating a category is the most celebrated and least frequently appropriate strategy in B2B software. It requires convincing a market that a problem it has not named is worth budget, and then that your name for it is the right one. The published successes are heavily survivorship-biased; the failures simply ran out of money quietly.

Category creation is defensible when the product genuinely does something no existing category describes, when the company has enough runway to spend two to three years before search demand exists, and when a founder or executive is willing to be the public voice of it. Absent any one of those three, the strategy usually collapses into paying to educate a market that then buys from an incumbent who arrives later with more distribution.

The cheaper alternative

Position inside an adjacent category that already has budget and search demand, and differentiate within it. Buyers find you through terms they already search, budget already exists in their planning, and procurement has a box to put you in. This is less exciting and dramatically more capital-efficient.

The hybrid

Sell into an existing category and describe the new one in parallel — category language on brand and thought-leadership surfaces, existing-category language on the pages that capture demand. This keeps the pipeline running while the narrative builds, and it is what most successful category creators actually did, whatever the retrospective case study says.

The pricing page, treated properly

The pricing page is the most-visited page on nearly every B2B software site and the one most often argued about internally. The argument is usually framed as transparency versus flexibility. In practice the choice is narrower than that, because buyers who cannot find a number do not conclude that pricing is flexible. They conclude it is expensive, and they look for a number elsewhere.

Somebody will publish an estimate of your pricing whether or not you do. Review sites, comparison posts written by competitors, and increasingly AI assistants summarizing all of it will produce a figure. If you have not published one, that figure will be wrong and unflattering, and you will have no way to correct it.

Publishing a shape rather than a price

Companies with genuinely variable enterprise pricing can still publish the structure: what drives cost, which tier a company of a given size typically lands in, and a starting figure. This gives a buyer enough to build an internal business case, which is the actual job of the pricing page, without committing to a number for a deal that has not been scoped.

What belongs on it besides prices

• A comparison of what each tier includes, written in outcomes rather than feature names
• The answer to what happens at renewal and whether prices increase
• Whether implementation, onboarding or support cost extra, stated plainly
• How overages or seat additions are billed mid-term
• An FAQ addressing the objections your sales team hears about price specifically
• A link to the security and compliance material, because procurement arrives here first

The most common own goal

A ‘contact us’ tier with no context at all, sitting beside two priced tiers. It reads as a wall rather than an invitation, and it is the point at which a large share of qualified enterprise evaluators leave to compare somebody who told them something.

Retention, expansion, and marketing’s share of net revenue retention

Net revenue retention decides more about a B2B software company’s valuation than new logo growth does, and it is treated almost universally as a customer success metric. Part of it is genuinely marketing’s, and the part that is gets systematically under-resourced because it does not appear in a pipeline report.

Adoption drives retention. If customers do not use the features they bought, they do not renew, and whether they use them depends heavily on whether they know the features exist and understand what problem each solves. That is a communication problem, and marketing is the function that solves communication problems at scale.

What marketing owns in retention

• Onboarding sequences that drive to a specific first outcome rather than a feature tour
• Feature announcement that explains the problem solved, not the capability shipped
• Documentation and education content, which is read far more by customers than by prospects
• Customer stories that show existing customers uses they had not considered
• Expansion-path content aimed at the buyer who already has one team on the product
• Community, where it exists, which is the highest-retention surface in software

Why expansion content outperforms acquisition content on cost

An existing customer already trusts the product, already has budget approved, and does not need to be convinced the category is real. The cost of producing content that expands an account is a fraction of the cost of acquiring a new one, and almost every company under-produces it because expansion revenue gets attributed to customer success rather than to the content that caused it.

Attribution in a business with an eleven-month cycle

Attribution is the single most contested subject in B2B SaaS marketing and the one where confident claims should be trusted least. When a purchase involves six to ten people across eleven months and dozens of touches, no model assigns credit correctly. The useful question is not which model is right but which model’s specific distortion you are willing to accept.

The ‘how did you hear about us’ field

An open-text self-reported attribution field on the demand form is, for most companies, more useful than the analytics stack. It is the only source that captures podcasts, private communities, word of mouth, conference conversations and AI assistant recommendations, none of which appear in any tracked session. It is imprecise and it is honest about being imprecise, which puts it ahead of a model that is equally imprecise while presenting three decimal places.

What to instrument now for next year

Closed-won revenue by first-touch channel, held for eighteen months. Almost nobody has this, because it requires deciding to store it before you need it. Set it up now and in eighteen months you will be able to answer the question that currently gets answered with an opinion.

Vertical SaaS and horizontal SaaS need different marketing

A product sold to one industry and a product sold across many industries share almost no marketing mechanics beyond the vocabulary. Treating them the same is a frequent and expensive error, usually made by an agency applying a playbook built for the other type.

Vertical: credibility is the whole game

In a vertical market the audience can tell within two paragraphs whether the writer understands their work. Generic marketing content actively damages credibility, which is why vertical SaaS content should usually be written or heavily shaped by someone from the industry. Low search volume is not the obstacle it appears to be, because a few hundred monthly searches from an industry with a five-figure contract value is a substantial commercial opportunity.

Horizontal: differentiation is the whole game

In a horizontal market there is abundant search volume and abundant competition for it. The content that wins is the content that takes a position specific enough to be disagreed with. Comprehensive neutral overviews of a well-covered subject are the most-produced and least-effective asset class in B2B software marketing.

In-house, agency, or fractional: choosing the operating model

The build-versus-buy decision in marketing is usually made on cost, and cost is the least useful input. The better frame is which capabilities you need continuously and which you need in bursts.

What genuinely belongs in-house

Product marketing, positioning and pricing. These require access to customer conversations, the product roadmap and internal politics that no external partner will ever have in sufficient depth. Companies that outsource positioning get positioning that sounds like their competitors’ because it was written from the same public inputs.

What an agency does better

Execution requiring specialized skill at variable volume: technical SEO, paid media management, content production at scale, and analytics implementation. These need depth of practice that only comes from doing the same work across many companies, and they are needed in uneven quantities.

Where fractional leadership fits

A fractional head of marketing makes sense in a narrow window: when the company needs senior judgement about strategy but does not yet have the volume of work to justify a full-time executive salary, and when there are people or agencies available to execute the resulting decisions. A fractional leader with nobody to execute produces strategy documents, which is the most common way this arrangement fails.

The arrangement that usually works at $2m to $20m ARR

One in-house product marketer who owns positioning and messaging, one in-house generalist who owns operations and the day-to-day, and an agency for content production, technical SEO and paid. This keeps the knowledge that must be internal internal, and buys the skills that are uneconomic to hire.

What the first ninety days should actually look like

Any agency engagement that opens with three months of strategy before anything ships is spending your runway on its own onboarding. Equally, one that starts publishing in week one has not understood the business. The reasonable shape is diagnosis and shipping in parallel, with the earliest work being the things that are obviously correct regardless of what the diagnosis concludes.

Weeks one to three

Technical audit and analytics verification, because every later measurement depends on it. Interviews with sales and with three to five recent customers. A full inventory of what already exists and how it performs. In parallel, ship the pages that are correct under any strategy: the pricing page improvements, the first comparison page, the security page.

Weeks four to eight

The content program begins in volume, prioritized by commercial intent rather than by search volume. Fix the technical problems the audit found, in order of how much traffic they suppress. Establish reporting that shows pipeline created by source, and be explicit that the numbers will be unreliable for the first two quarters because the cycle is longer than the reporting period.

Weeks nine to thirteen

First honest review. At this point leading indicators exist — rankings, qualified traffic, trial or demo volume, conversation quality reported by sales — and lagging indicators do not, because eleven months have not passed. An agency that presents revenue attribution at ninety days in a business with an eleven-month cycle is presenting something it constructed rather than measured.

What should be true at ninety days

• Analytics measure what they claim to measure, verified rather than assumed
• The high-intent pages exist: pricing, comparison, security, top integrations
• Content is publishing on a schedule that has held for at least six weeks
• Sales can name specific content they use in deals
• Reporting shows pipeline by source, with the caveats stated plainly
• Nobody is being asked to judge an eleven-month cycle on a three-month result

Getting cited by AI assistants, not just ranked by Google

A growing share of B2B software evaluation now starts inside an AI assistant rather than a search results page. A buyer asks which tools solve a problem, gets a shortlist of three to five named products with a sentence of justification each, and proceeds from there. Being absent from that shortlist is a new and largely unmeasured way to lose deals, and it does not show up in any analytics report because no click ever happened.

This is not a separate discipline from search, and the agencies selling it as one are overcharging for it. Assistants are trained on and retrieve from the open web, so the same substantive, specific, well-structured content that earns rankings is what gets quoted. What differs is the format the content needs to be in for a machine to extract a clean, attributable claim from it.

What makes a page quotable

• A direct answer in the first two sentences under each heading, before the context and caveats
• Specific numbers with the basis stated, because a model will not repeat a figure it cannot frame
• Headings phrased as the question a person would actually ask
• Comparison content in table form, which extracts far more reliably than the same facts in prose
• Named entities used consistently, so the model can connect the page to the company
• Dates and last-updated stamps, which weigh heavily in whether stale content gets surfaced

The structured data that actually matters here

Schema markup does not force a citation, but it removes ambiguity about what a page asserts. The types that carry weight for B2B software are Organization to establish the entity, SoftwareApplication to describe the product, FAQPage built from real Question and Answer pairs, and Product with offers where pricing is public. Google documents the implementation requirements for FAQ structured data and for software application markup, and its guidance on AI features in Search is explicit that there is no separate markup for AI surfaces — the ordinary rules apply.

Crawler access, which is where most companies quietly lose

A page cannot be cited if the crawler was blocked from reading it, and a surprising number of B2B software sites block AI crawlers by default through a security vendor’s rule set that nobody chose deliberately. The relevant agents are documented: OpenAI publishes its bot list and user agents, Google documents Google-Extended among its crawler fleet, and independent directories such as Dark Visitors track the wider set. Access is governed by the ordinary robots.txt standard, and a substantial portion of training data still originates from Common Crawl.

The decision to allow or block is genuinely yours to make and there are defensible arguments both ways. What is not defensible is not knowing which you have chosen. Check the file, check what your CDN’s bot rules do on top of it, and make the choice on purpose.

The llms.txt question

A proposed convention, llms.txt, offers a curated markdown map of a site for language models to read. Adoption by the model providers is not universal and nobody should claim it is. It costs very little to publish, it does no harm, and it is worth doing on the same basis as a sitemap — cheap, structured, and useful if it is read. Treating it as a ranking factor would be overstating it considerably.

How to tell whether any of this is working

Ask the assistants directly and repeatedly. Run the ten questions a real buyer would ask about your category, record which products get named and how they are described, and repeat monthly. It is manual, it is imprecise, and it is currently the only honest measurement available. Any agency presenting a confident percentage for AI visibility is presenting a proxy it invented; the useful version of this report is a list of the actual prompts, the actual answers, and whether your name appeared.

Where the biggest gains usually are

For most B2B software companies the fastest improvement is not markup at all. It is publishing the specific comparative and pricing facts that assistants are asked for constantly and that most vendors refuse to state. If a model cannot find what you cost, who you are better than and on what dimension, or which systems you integrate with, it will summarize a competitor who said all three.

Questions SaaS companies ask

Watch before you hire an agency

Related: tech marketing agency covers technology marketing more broadly, marketing agency pricing covers what agencies charge, and generative engine optimization covers being cited in AI answers.

Frequently asked questions