Everything Rankbank looks for

Rankbank checks your site against 58 rules. Each one either fires or it doesn't, based on what the last read found and what your search data says, and each one that fires becomes a finding card. This page is the whole list.

It's a reference rather than a read-through. Skim the family headings, find the one you saw on a card, and read that row. If you want the shape of a card rather than the catalog behind it, start with findings: how to read a card.

How to read these tables

Every family gets one table with three columns: what the finding is called, what makes it fire, and why it's worth your time. Two tags appear in the "fires when" column:

[needs Search Console] means the rule can't run without your Google Search Console data. There are 27 of these. Everything else works from the read of your site alone. See connecting Google Search Console.

[needs market data] means the rule needs third-party market data, which isn't available for every domain. Exactly one finding depends on it. When it's absent, the finding simply doesn't appear, which is the intended behaviour rather than a degraded one.

Two rules also decide how findings behave in bulk. A family that affects more than three pages rolls up into one finding naming up to twelve examples, so "43 pages have no main heading" is one card rather than 43. And findings are de-duplicated by key, so the same fact never appears twice.

Snippets: titles and descriptions

Twelve findings about how your pages present themselves in search results. This family is the cheapest work on the whole list, because none of it needs new content, new links, or anybody else's cooperation.

Finding	Fires when	Why it matters
Ranks well, wins few clicks	A page has at least 200 appearances, sits at position 20 or better, and earns under 60% of the clicks typical for its position. [needs Search Console]	Google already shows this page to thousands of people, so the hard part is done. They read your title and choose somebody else.
No title	An indexable page has no title tag at all.	The title is the strongest on-page ranking signal there is, and the headline of your listing. Without one Google invents a label from whatever text it can find.
Pages sharing one title	Two or more pages carry the exact same title.	Google can't tell which of them to rank, so they compete with each other and often all sink. Searchers see identical listings with no way to choose.
Titles cut off	A title is longer than 60 characters.	Google truncates around sixty characters, and it sometimes rewrites over-long titles entirely, which takes control of your listing away from you.
Titles too short	A title is shorter than 15 characters.	One or two words can't match the phrases people actually type, and they win no clicks against fuller listings.
Titles that only say your company name	A page other than the homepage is titled with nothing but the brand name.	You already rank first for your own name. Every other page titled that way is invisible for everything else it could rank for.
No description	An indexable page has no meta description.	Google clips whatever sentence on the page contains the search words instead, often mid-thought, sometimes a cookie notice.
Descriptions cut off	A description is longer than 165 characters.	Anything past the cut is invisible, and the reason to click is usually the part that gets lost, because writers put it at the end.
Descriptions too short	A description is shorter than 70 characters.	Most of the line Google gives you is left blank, and Google often fills the space with a scraped sentence instead.
Pages sharing one description	Two or more pages carry the same meta description.	Identical descriptions make different pages look like copies of each other, to searchers scanning results and to Google deciding which deserves to rank.
Homepage listing undersells the site	The homepage's title is under 25 characters, or it has no description.	The homepage listing is the most-seen snippet you own and often the first impression a stranger gets. It should say what you do and who for.
Title and headline disagree	A page's title and its H1 share less than 30% of their words.	Searchers click a promise and land on a different one. Many go straight back, and Google reads that bounce as an answer about your page.

Rankings within reach

Eight findings about pages that are close to earning more than they do. All of them need Search Console, because "close" is only knowable from position data.

Finding	Fires when	Why it matters
One push from the clicks	A page sits between positions 4 and 10 with at least 300 appearances, and has at least the site's median number of incoming internal links. [needs Search Console]	The top three take roughly half of all clicks and position eight takes about three percent. The same effort on a page ranking fortieth would achieve nothing.
On page one, and unsupported	Same as above, except the page has fewer incoming internal links than the site's median. [needs Search Console]	Internal links are the one lever you fully control, and Google reads them as your own opinion about which pages matter. Right now your site is voting against this one.
Stuck on page two	A page sits between positions 11 and 20 with at least 200 appearances. [needs Search Console]	Position eleven earns roughly one percent. But ranking there means Google already considers the page a serious candidate, which makes these the highest-leverage upgrades on the site.
A search you nearly rank for	One specific search has at least 100 appearances and sits between positions 4 and 10. [needs Search Console]	Page-level averages hide query-level wins. One well-aimed paragraph, using the words people actually typed, is often the difference between position seven and position two.
Losing ground	A page earned at least 10 clicks in the previous window and has since slipped 3 or more positions. [needs Search Console]	Rankings decay quietly. Recovering position eight from eleven is ordinary work; recovering it from thirty is a project. This is the moment that costs least to act on.
Climbing	A page improved by 3 or more positions, or earned at least 10 clicks and grew them by 50% or more. [needs Search Console]	Momentum is a signal Google responds to. The crawler is already revisiting the page, so this is the moment to add links and fill gaps, not after it stalls.
Demand proven, page not good enough	A page has at least 500 appearances but ranks worse than position 20. [needs Search Console]	Google keeps auditioning the page and keeps benching it. The demand is proven. This is not a title tweak, it's a rewrite, and it's worth doing because the audience is already there.
Top three, clicks taken	A page ranks in the top three with at least 300 appearances but earns under half the clicks typical for that position. [needs Search Console]	Something sits above you: a featured snippet, an AI answer, or a wall of ads. You can't remove them, but the featured slot is usually winnable.

Internal links and architecture

Ten findings about how your pages support each other. Internal links need no outreach, no budget and nobody's permission, which is why this family is worth working before anything harder.

Finding	Fires when	Why it matters
Earning traffic, linked from nowhere	A page nothing links to is earning appearances in Google. [needs Search Console]	The sharpest finding this tool makes: real search traffic is landing on a page your own site pretends does not exist. It is already earning, so linking it is pure upside.
Nothing links to it	A page has no incoming internal links and no search appearances.	Visitors can't stumble on it and Google may not be sure it still exists. Each one is a decision: link it, or retire it.
Best pages, least support	A page in the top 20% of the site by clicks has fewer incoming internal links than the site's median. [needs Search Console]	Your internal linking should mirror what actually earns. Redirecting a few links to something already working beats starting something new.
Buried, and still earning	A page 4 or more clicks from the homepage has at least 100 appearances. [needs Search Console]	Depth is one of the ways Google reads importance. A page this deep that Google still shows is fighting its own site's architecture.
Reachable only from the menu	A page linked only from primary navigation, with nothing in the body of any page pointing at it, is earning appearances. [needs Search Console]	Google discounts anything repeated on every page as boilerplate. This page is already being shown, so it's worth more than a menu entry.
Linked only from the footer	A page's only incoming links are footer links.	Footer links are the faintest signal a site sends. If a page is worth having, one real sentence should point at it.
Links that say nothing	A page with at least two incoming links has 30% or more of them worded "click here", "read more", "learn more", "this page" or similar.	Link text is a ranking hint and an accessibility feature at once. You already have the links, they're simply mute.
Two of your pages compete	Two or more of your pages rank for the same search, each with at least 50 appearances at position 30 or better. [needs Search Console]	Google splits the vote between them, so each ranks worse than a single agreed contender would. Consolidating usually works within weeks, which is rare in SEO.
Linking to everything	A page links out to more than 150 of your own pages.	A page divides its strength among the pages it links to, so a page linking to hundreds passes each of them almost nothing.
Links taking a detour	A page with incoming internal links is reached through one or more redirects.	Every hop costs a fraction of a second and a little of the strength the link would pass. Chains also break silently over time.

Both orphan findings deliberately skip pages that are supposed to have no links pointing at them: 404 and error pages, "not found", "thank you" and search-results pages. A thank-you page with no inbound links is how it is meant to be, and listing it would teach you to skim past findings, which costs more than the finding is worth.

Content

Nine findings about what's actually on the page, and what isn't.

Finding	Fires when	Why it matters
Thin, and already being shown	A page under 300 words has at least 100 appearances. [needs Search Console]	The highest-return writing on the site: the audience is already ordered, you simply have not stocked the shelf.
Too little on the page	An indexable page is under 300 words with fewer than 100 appearances.	Pages this light rarely rank, and in quantity they dilute how the whole site is judged. Expand or merge and redirect.
No main heading	An indexable page has no H1.	The main heading tells a reader in the first second that they landed in the right place, and it's one of the signals Google uses to work out what a page covers.
Several main headings	A page has more than one H1.	More than one main heading muddles what a page is about, for readers skimming and machines parsing alike.
A search you have no page for	A search with at least 100 appearances ranks worse than position 30, isn't a brand search, and no page of yours targets it. [needs Search Console]	A pre-validated content idea: the demand is measured rather than guessed, and Google already considers you relevant enough to show.
A topic with no lead page	A group of related pages exists with no page leading it.	Search rewards the hub-and-spoke shape. Without a lead page the authority those pages earned has nowhere to pool, and each competes with its siblings.
Related pages not linking to their lead	A topic has a lead page, and members of the group don't link to it.	Without those links Google sees a handful of unrelated pages rather than a body of work.
A question the page never answers	A search starting with who, what, why, how, when, where, can, does, do, is or are has at least 100 appearances, and the question's real words never appear close together anywhere in the page's text. [needs Search Console]	Google matched the page by vocabulary rather than substance. A tight paragraph answering the literal question wins the click, and frequently the featured box above position one.
Wrong kind of page for the search	A commercial search lands on a blog page, or an informational search lands on a product page, at position 5 to 20. [needs Search Console]	The mismatch is structural. No amount of on-page tuning fixes it. Change the shape of the page, or aim it at a different search.

Technical health

Nine findings about the machinery. Four of them are roll-ups of the structural problems on the Problems to fix page, restated in the catalog's own voice so they can be priced and prioritised alongside everything else.

Finding	Fires when	Why it matters
Earning traffic while telling Google to drop it	A page carrying a noindex tag is earning clicks. [needs Search Console]	The only urgent finding in the catalog. A standing eviction notice: the traffic survives only until Google next re-processes the page. Usually an accident, and always losing money on a timer.
Pages Google knows and your site forgot	Search Console reports at least 50 appearances for URLs a crawl of your own links never reached. [needs Search Console]	Old campaign pages, orphaned sections, a whole area the navigation forgot. Free traffic landing in a place your site does not acknowledge.
Links to dead pages	Any broken internal link or broken crawled page exists.	Anyone who clicks hits an error and Google follows the same dead end, wasting the crawl budget it allocates you. Among the cheapest fixes on any list.
Pages crediting the wrong page	Any missing canonical, canonical pointing elsewhere, or canonical pointing off-site exists.	A page hands its ranking credit away and quietly disappears from results, while looking completely fine to a visitor.
Linking to pages you hide	Your pages link to a page marked noindex.	You're telling Google to ignore a page while pointing your own links at it. Whatever strength those links carry goes nowhere.
Sitemap problem	The sitemap fails to load or parse.	Search engines use the sitemap as your table of contents. A broken one slows down how fast new pages get discovered, which matters most when you're publishing.
Images with no description	Any page has images without alt text.	Alt text is read aloud to screen-reader users and read by Google Images, a search engine in its own right. An image with no description is invisible to both.
Bare links when shared	An indexable page is missing its og:title or og:description.	Shared in Slack, WhatsApp or LinkedIn, the page appears as a naked URL rather than a card. A bare URL gets a fraction of the clicks a card does.
No structured data	A site with at least 5 indexable pages has structured data on 10% of them or fewer.	Star ratings, FAQ drop-downs, breadcrumbs and recipe cards are only offered to pages that describe themselves in structured data.

Off-site authority

Three findings about what the rest of the web thinks of your domain. These are the ceiling on everything above, and the only part of the list you cannot fix alone.

Finding	Fires when	Why it matters
Nobody links to you	The site has at least 5 pages and zero referring domains.	With no links at all you're competing on content alone against sites that are not. The first ten referring domains matter far more than the next hundred.
Little weight with search engines	The site has at least 20 pages, at least one referring domain, and authority under 20/100.	That caps how well any page can rank no matter how good the writing is. The first ten or twenty referring domains move a small site visibly.
Beaten by heavier sites	At least three rated competitors are known and their median authority is 15 or more points above yours. [needs market data]	"Your authority is 12/100" is a number without a scale. The sites Google puts above you for the same searches are the only scale that matters. Nothing on the page closes a gap this size.

Both of the first two carry the link-building callout on their cards, with routes into link swaps and guest posts. "Beaten by heavier sites" adds an impact note of its own: "Authority moves over months, not days. It is the ceiling on everything else, which is why it is worth starting now."

See authority for the meter, the five bands, and what actually moves the number.

What your search data says

Four findings about the shape of your traffic overall rather than any one page.

Finding	Fires when	Why it matters
Traffic that already knew your name	At least 50 clicks in the window, and 70% or more of them come from searches for your brand. [needs Search Console]	Branded searches are loyalty, not discovery. The real measure of search work is clicks from people who did not know you existed.
A new search you appear for	A search with at least 100 appearances shows up that wasn't in the previous window. Needs a previous window to be absent from, so it never fires on a first sync. [needs Search Console]	New impressions are an audition. What happens over the next few weeks decides whether Google keeps showing you, and making the page answer it properly now is far easier than winning it back later.
Searches you stopped appearing for	A search that earned at least 10 clicks in the previous window has disappeared entirely. [needs Search Console]	Disappearing is different from slipping. It usually means the page changed, broke, was deindexed, or was comprehensively beaten. Either way it's measured traffic you had and no longer have.
Pages Google never shows	At least 30% of a site's indexable pages, on a site with at least 10 of them, have zero appearances. [needs Search Console]	A page never shown to anyone is either not indexed, targeting a phrase nobody searches, or too similar to another of yours. In quantity they dilute how the whole site is judged.

A new search you appear for carries the Win badge, so it surfaces in the green Working panel rather than in your to-do list.

Wins

Rankbank could have shipped without this family. It didn't, for a stated reason: a tool that only ever nags is a tool people stop opening. Wins close the loop from "we told you to do this" to "here is what it did".

Finding	Fires when	Why it matters
Traffic is growing	Total clicks are up 15% or more on the previous window. [needs Search Console]	The number this whole product exists to move. Worth noting what changed in the period that produced it, because the same lever usually works twice.
Reached page one	A page moved from worse than position 10 to position 10 or better. [needs Search Console]	The single biggest step change in search: clicks go from roughly nothing to something real. Pages that just moved tend to keep moving with a little support.
Reached the top three	A page crossed onto page one and landed at position 3 or better. [needs Search Console]	The top three take about half of all clicks between them. From here a better title is worth more than any amount of further ranking work.

Two more findings carry the Win badge without living in this family: Climbing (in rankings within reach) and A new search you appear for (in what your search data says). Both are moments to spend effort rather than to celebrate and move on, and their fix steps say so.

Which findings need what

If you've just connected Search Console and are wondering what changed, this is the answer: 27 of the 58 rules only run once search data exists, and they include most of the ones that carry a clicks-per-month price. Before you connect, Rankbank can tell you what's structurally wrong. After you connect, it can tell you what that's costing.

The one finding that needs market data, Beaten by heavier sites, will not appear for most domains. Neither will the market line on other cards, or the coverage tile. None of that is broken. See the Search traffic page for what the coverage tile shows when it does appear.

Everything priced in clicks uses the same click-through curve, and findings: how to read a card explains exactly how that figure is worked out and what it does and doesn't promise.