SEO · CONTENT · MARKETING · 11 JULY 2026 · 7 MIN READ
What your competitors rank for, and why it does not matter
A gap analysis is a list of things somebody else decided to do. Useful as a map of page types, useless as a plan — and your own Search Console already knows more.
Competitor keyword research is useful for one narrow thing: discovering *page types* and query shapes you have not considered. It is close to useless as a plan, because a competitor’s rankings reflect their catalogue, their link profile, their brand and years of accumulated decisions — none of which you inherit by targeting the same terms. The estimated volumes attached to those keywords are modelled, not measured, and your own Search Console holds measured data about queries you already rank for. Start there.
IN SHORT
- A competitor gap report tells you what somebody else chose to do, not what you should do — their rankings rest on assets you do not have.
- Third-party keyword volumes are estimates produced by a vendor’s model; Search Console impressions and clicks are counts of what actually happened on your site.
- Google documents that position in the Performance report is “the topmost position occupied by a link to your property or page in search results, averaged across all queries”, so a single rank-tracker number is a simplification of a distribution.
- Google states that results depend on “your location, past Search history, and Search settings”, which is why two people checking the same ranking see different answers.
- The genuinely useful output of competitor research is structural: which page types exist, how they are organised, and which intents they serve.
- Most stores are not constrained by knowing which pages to build. They are constrained by being able to ship them.
What a gap report actually contains
Run any competitive tool and you get the same artefact: a few thousand keywords your competitors rank for and you do not, sorted by estimated volume. It looks like a plan. It is a record of other people’s decisions, filtered through a model.
Two things are wrong with treating it as a plan. The first is attribution: a competitor ranking for a broad category term is usually not doing so because of the page — it is because of a decade of brand search, coverage, and links that the page happens to sit behind. Building the equivalent page does not transfer any of that. You get the page, not the position.
The second is that the numbers are modelled. Keyword volume from a third-party tool is an estimate derived from clickstream panels, Google Ads data and the vendor’s own extrapolation. That is not a criticism of the vendors — the true number is not publicly available, so estimation is the only option — but it means the figure is a vendor’s opinion. Ranking a plan by opinion-with-decimal-places is how a quarter gets spent on terms that were never going to convert.
Your own data is measured, and you already have it
Search Console reports impressions, clicks and position for queries where you have already appeared. That is a count of real events on your site, not a model of someone else’s. It is also full of the most valuable category in SEO: queries where you rank on page two for something you were not trying to rank for. Those are pages that Google already considers relevant, which is most of the work.
It is worth knowing how Google defines the numbers, because misreading them causes bad decisions. Google documents position as “the topmost position occupied by a link to your property or page in search results, averaged across all queries”, and explains that “position in Google Search is typically calculated from top to bottom on the primary side of the page, then top to bottom on the secondary side of the page”. So an “average position 8.4” is an average across queries, devices and locations — not a rank you can check by searching.
That averaging is also why rank tracking disagrees with itself. Google states plainly that its systems use “your location, past Search history, and Search settings to determine what is most relevant for you at the moment”, and that “the language of your query determines how most results will be displayed”. There is no single ranking to track. Treat a tracker as a trend line for a portfolio and stop arguing about individual positions.
The part of competitor research that is worth doing
Ignore the keyword list. Read the site.
What you are looking for is structural, and it does not need a tool: which page types exist that you do not have, what intent each one serves, and how the catalogue is organised into them. A competitor with buying guides per use case, size and fit pages per category, and comparison pages between their own ranges has made three structural decisions. Those are transferable in a way that a keyword is not, because they are about how a catalogue is presented rather than about who has the authority to rank.
Three questions that produce more than a gap report:
- Which queries do they have a dedicated page for, where we have only a product or a collection? That is a page-type gap, and it is actionable.
- What do they answer on the page that we answer only in a PDF, a support article, or not at all? Answering questions in the place where they are asked is the cheapest SEO work available.
- Where are they weak? A category page with forty words of copy and no internal links is a competitor telling you what they have not got round to. That is a better target than the thing they are strongest at.
Where the real constraint is
Here is the honest version, and the reason we would talk most stores out of a bigger research budget: almost nobody is stuck because they do not know which pages to build. They know. The list has been in a spreadsheet since last year. They are stuck because publishing a new kind of page takes a developer, a ticket and three weeks, so it never gets to the top of anybody’s list.
That is a production problem wearing an SEO costume. A store where a merchandiser can assemble a buying guide from existing blocks on a Tuesday afternoon will out-publish a store with a better keyword list and a development dependency, every single time. A [section library](/services/build/shopify-section-library) — composable, merchandiser-arrangeable blocks — is usually the higher-return investment, and it is not close.
The test is simple. Ask how long it takes to put a new page type live today. If the answer is weeks, buying more research is buying a longer backlog.
When competitor research does earn its place
There are cases where it is genuinely the right tool, and they share a shape: you do not yet know the shape of the market.
- Entering a new market or country, where you have no Search Console data because you have no presence yet.
- Launching a category you have never sold, where the vocabulary buyers use is genuinely unknown to you.
- Diagnosing a specific loss — you ranked, you no longer do, and you want to see what now occupies the results. Looking at what Google currently rewards for that query is real evidence.
- Sanity-checking a plan already built from your own data, to see whether you have missed an intent entirely.
What we would do instead
Spend a day, not a quarter. Export Search Console queries for the last twelve months and find the ones with impressions and no clicks — pages Google is showing that nobody chooses. Group them by intent rather than by keyword. Then look at three competitors for half a day, writing down page types only, and add any type your grouping suggests you are missing.
That produces a shorter list than a gap report and every item on it is supported by evidence about your own site. Then spend the rest of the budget on being able to publish it — which is the part that was actually broken.
Questions this raises
Is competitor keyword research worth doing at all?
For discovering page types and query shapes, yes — a half-day of reading competitor sites is genuinely useful. As the basis for a content plan, no. A competitor’s rankings depend on their brand, links and catalogue, none of which you acquire by publishing a page about the same term.
Are third-party keyword volumes accurate?
They are estimates. True search volume is not published, so vendors model it from clickstream panels, ads data and extrapolation. Different tools give different numbers for the same term, which is the clearest evidence that the figure is a model output rather than a measurement. Use them for relative scale, never as an input to a revenue forecast.
Why does my rank tracker disagree with Search Console?
They measure different things. Google documents Search Console position as the topmost position of your site averaged across all queries, and states that results depend on location, past search history and settings. A tracker checks one query from one location, so it is a single sample of a distribution that Search Console reports the mean of.
What should I look at on a competitor’s site instead of their keywords?
Page types and structure: which intents have a dedicated page, how the catalogue is organised into them, what is answered on the page rather than buried in support content, and where their pages are visibly thin. Those observations transfer between sites in a way that individual rankings do not.
We have no Search Console history — what then?
That is the strongest case for competitor research, and it applies to new markets, new categories and new brands. Use it to learn the vocabulary and the page types, publish, and then switch to your own data as soon as you have three months of it. Competitor data is a substitute for evidence, not a form of it.
Should we target the terms our biggest competitor ranks first for?
Usually not. The terms a market leader dominates are the ones where their accumulated advantage matters most, so they are the most expensive place to compete. The pages where they are thin, or the intents nobody has a page for, cost far less and convert better because the match is closer.
NEXT STEP
Free store audit
A senior Shopify engineer reviews your storefront, theme performance and checkout, then sends a prioritised list of fixes.