SEO Diagrams: The Six Different Things People Mean

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

“SEO diagram” covers about six unrelated drawings that share a name — a crawl-index-rank flow, a site architecture tree, an internal link map, a topic cluster, a URL structure and a redirect map. This page separates them, explains what each is for and when you need it, and sets out what to actually look for in the three that most affect whether a site performs.

What is an SEO diagram?

Any drawing used to explain how search works or how a site is organized. In practice the phrase covers about six different things, and people arrive looking for whichever one their problem needs.

That ambiguity is the reason a search for it returns such a scattered set of images. Someone planning a site wants a site architecture diagram; someone explaining ranking to a client wants a crawl-index-rank flow; someone auditing links wants an internal link map. They are unrelated drawings sharing a name.

Only two of these are diagrams you draw once and keep. The architecture and cluster maps are living documents; the redirect map is a working file that matters intensely for about two weeks and then becomes a record.

What does the crawl-index-rank diagram actually show?

Four stages: discovery, crawling, indexing, and ranking. Its value is that it makes visible the three places where a page can silently drop out before it ever competes.

Most explanations of SEO jump straight to ranking factors, which skips the part that causes the majority of real problems. A page that is not discovered, not crawlable, or not indexed cannot rank at all, and no amount of optimization changes that.

Discovery

The engine learns the URL exists — from a link, a sitemap, or a redirect. A page with no internal links and no sitemap entry may never be found.

Crawling

The engine fetches the page. Blocked by robots.txt, server errors, or being buried too deep to be worth reaching.

Rendering

Increasingly its own stage. Content that only appears after JavaScript runs may or may not be seen, depending on the render budget.

Indexing

The engine decides whether to store it. Thin, duplicated or canonicalised-away pages get crawled and then discarded.

Ranking

Only now do the familiar factors apply — relevance, links, quality signals, intent match.

Serving

What actually appears, which is increasingly a summary or an answer rather than ten blue links.

How should a site architecture diagram be drawn?

As a tree, with the homepage at the top and depth measured in clicks. The single number to read off it is how many clicks it takes to reach the pages that matter.

Three clicks is the conventional target and it is a guideline rather than a law. What matters is that important pages are not buried behind pagination, faceted filters, or a category level that exists only to tidy the navigation.

Orphan pages are the most common finding and the easiest to fix. They usually appear because a page was published outside the normal flow — a landing page, a campaign page, something built for an email — and nobody ever linked to it from the site itself.

What is a topic cluster diagram for?

Planning content so that a group of pages reinforces one another instead of competing. It is the practical answer to accidentally publishing four pages about the same thing.

The structure is a pillar page covering a subject broadly, surrounded by pages covering specific questions in depth, all linked to the pillar and to each other where relevant. Drawn out, it makes cannibalization obvious before it is published rather than after.

The pillar

Broad, covers the subject at survey depth, links out to every cluster page.

The cluster pages

Each answers one specific question thoroughly, and links back to the pillar.

The links between them

Where two cluster pages genuinely relate, link them. Not everything to everything.

One intent per page

The rule the diagram exists to enforce. Two pages chasing the same intent split their own signals.

The gaps

Drawing the cluster shows you which questions you have not covered, which is frequently more useful than the plan itself.

The pruning candidates

Old posts that overlap a cluster page should be merged or redirected, not left to compete.

How does a URL structure diagram differ from an architecture diagram?

Architecture describes how pages relate; URL structure describes what they are called. They are usually drawn together and they are not the same decision, which is why one frequently ends up contradicting the other.

A site can have a sensible three-level hierarchy and URLs that encode a completely different shape — flat slugs under a hierarchical navigation, or folders that no longer match the categories they were named for. Neither is fatal on its own, but the mismatch makes every later decision about canonicals, breadcrumbs and redirects harder than it needed to be.

The trailing-slash row causes more confusion than its size suggests. If both forms return a page rather than one redirecting to the other, every URL on the site exists twice, and the split shows up in reporting long before anyone traces it to its cause.

Do you need a redirect map diagram for a migration?

You need a redirect map. Whether it is drawn or held as a spreadsheet matters far less than whether every old URL that had traffic or links appears in it.

This is the least glamorous item on the list and the one whose absence causes the most damage. Migrations lose rankings overwhelmingly because URLs changed and nothing pointed the old ones at the new ones — not because of anything subtle.

Redirecting everything to the homepage is the classic shortcut and it is treated as a soft 404 — the old page’s signals are not carried across, and the visitor lands somewhere that does not answer why they clicked.

What tools do people use to draw these?

Anything that draws boxes. The diagram is a thinking aid, and the value is in the decisions it forces rather than in the rendering.

The one genuinely useful distinction is between drawings you make by hand — architecture, clusters — and drawings generated from real data, such as an internal link map produced by a crawler. The generated ones tell you what the site actually is; the hand-drawn ones tell you what you intended.

How should an SEO diagram be made accessible?

A diagram that exists only as an image is unavailable to anyone using a screen reader, and unavailable to the search engines you are drawing it for. The fix is to publish the underlying structure as text alongside it, not to write a longer alt attribute.

Alt text has a practical ceiling. It works well for “a bar chart showing crawl requests falling by roughly half after the robots.txt change” and badly for a site architecture with forty nodes, because a single unstructured sentence cannot convey a hierarchy. Trying to force it produces alt text nobody can follow.

For anything structural, publish the same information twice: the diagram for people who read visually, and a nested list, a table, or a short passage of prose for everyone else. A site architecture diagram becomes a nested list of sections and their children. A crawl-index-rank flow becomes an ordered list of stages. The text version is usually easier to keep current than the image, and it is the version a language model can actually read.

This is also why an SVG beats a flattened PNG for these. Text inside an SVG stays selectable and searchable, the diagram stays legible when someone zooms, and it does not blur on a high-density display the way an exported bitmap does.

Who is the diagram actually for, and how does that change it?

Draw a different diagram for an engineer, an executive and a writer. The most common reason a diagram fails is that one version is shown to all three.

An engineer needs precision: real URL patterns, actual status codes, the specific rule that fires. Ambiguity there causes implementation mistakes, and a tidy diagram that omits the awkward exceptions is worse than no diagram, because it will be built exactly as drawn.

An executive needs the shape and the consequence. Forty nodes is noise; three boxes and an arrow showing where the traffic is being lost is a decision aid. If the diagram needs narrating to make sense, it is the wrong diagram for that room.

A writer or subject-matter expert needs to see where their work sits and what it should link to. That is a topic map, not an architecture, and the useful version usually omits the technical layer entirely. Redrawing the same information three ways is not duplicated effort — it is the difference between a diagram that gets acted on and one that gets admired and filed.

What a process diagram is for

Showing which decisions depend on which, so people stop arguing about order. A diagram that lists activities without dependencies is a checklist drawn sideways.

Frequently asked questions