Keyword Cannibalisation: When Your Own Pages Compete and Both Lose

Some of the most frustrating SEO problems are not caused by competitors. They are caused by your own pages. Keyword cannibalisation is when two or more URLs on your site target the same query, splitting the ranking signal that should have concentrated on one, so both underperform. On large sites it is rampant and rarely diagnosed, and I have watched it quietly hold down pages that were doing everything else right.

What Cannibalisation Actually Does

Search engines want to show one best result from your site for a given query. When you give them two pages that both look like the answer, they hedge: they split link signals, relevance, and click data across the two, rank whichever looks marginally better at any moment, and swap them in and out. Neither page accumulates the authority it needs to break through, so both sit stranded, usually together on page two, trading places while a competitor with a single strong page owns the top.

On a large marketplace I worked on across four markets, this pattern was everywhere: a blog post and a category page both chasing the same term, or two near-identical category URLs generated by filters, each suppressing the other. The fix was rarely new content. It was deciding which URL should own the term and making every signal point at it.

How to Diagnose It

How to Fix It

Once you have found it, the fix is a decision followed by consolidation. Pick the URL that should own the term, usually the one with the most authority, the best intent match, or the most commercial value, and then make everything point at it.

Done across a site, this is some of the highest-return technical work available, because it takes pages you already have and lets them finally rank. No new content, no new links, just the removal of the self-inflicted competition that was capping them.

Why Large Sites Cannibalise Themselves

Small sites rarely have a serious cannibalisation problem, because they do not have enough pages to compete with themselves. Large sites cannibalise constantly, and usually not on purpose. Faceted navigation spins up multiple near-identical category URLs for the same set of products. A blog post written for traffic ends up targeting the exact term a category page needs. Two teams, content and merchandising, each create a page for the same topic without knowing the other exists. At scale, self-competition is the default state, not the exception, which is why diagnosing it has to be a routine audit rather than a one-off.

This is also why cannibalisation and index bloat travel together. The same filter and parameter URLs that bloat the index are frequently the ones competing with your clean category pages for the same terms. Fixing one often fixes the other: canonicalising or pruning the duplicate URLs both trims the index and ends the cannibalisation, concentrating the signal on the single page you actually want ranked. On a large site, treating them as one problem rather than two is usually the faster path to both fixes.

Shopify Collection Pages: Where E-commerce SEO Is Actually Won

Ask most Shopify owners which pages matter for SEO and they will say product pages or the blog. The real answer, for commercial search, is usually the collection page. It is the page that targets the high-volume category term, the page a buyer lands on ready to choose, and the page most stores treat as a bare product grid with no thought given to search at all. Get collection pages right and you win the terms that actually drive revenue.

Why Collection Pages Carry the Commercial Load

Category-level searches, the cookware, the wall decor, the water bottles, carry far more volume than any single product term, and the searcher is close to buying. The collection page is what should rank for them. On the store I studied, collection pages were the largest source of commercial, non-blog clicks, yet they were also declining, precisely because they were being left as thin grids while competitors built theirs out as proper landing pages.

The competitors who were winning those category terms were not doing anything exotic. They simply treated the collection page as a page, not a shelf: a clear H1 and title targeting the category term, a short block of genuinely useful copy, a crawlable and paginated product grid, and internal links and footer content that passed authority around the catalogue. The store that was losing had grids. The stores that were winning had landing pages.

What Turns a Grid Into a Landing Page

Finding the Collections Worth Building

The fastest way to find collection-page opportunities is a competitor gap: the high-volume category terms your competitors rank for because they have a dedicated, built-out collection and you do not. Some are pages you already have and have under-optimised; others are collections you have not created yet. Both are among the highest-return work available to a Shopify store, because they target commercial demand directly.

None of this is glamorous, and that is the point. Collection page SEO is unglamorous, high-return work that most stores skip because the blog is more fun and product pages feel more urgent. The stores that quietly own their category terms are the ones that treated the collection page as the commercial landing page it actually is.

Shopify’s Built-in SEO Traps, and How to Fix the Duplicate Content It Creates for You

Shopify is a genuinely good SEO platform in most ways, which is exactly why its built-in traps catch people off guard. The defaults that make it easy to launch a store also quietly generate duplicate URLs, bloat your index, and hide crawl problems behind a clean-looking storefront. Auditing Shopify stores, I see the same handful of issues again and again. None are hard to fix once you know they exist. Most stores never look.

How One Product Becomes a Dozen URLs

The most common Shopify SEO problem is duplication the store creates for itself. A single product ends up reachable through its clean URL, through variant parameters, through filter and sort parameters, and through the /collections/all/ path, all serving near-identical content. Multiply that across a catalogue and you have thousands of thin, competing URLs where you meant to have a few hundred strong ones.

On one store, the root product URL was even redirecting to the URL with a variant ID appended, so the version being linked and the version being indexed were not the same page. Each of these paths splits the ranking signal that should have concentrated on one clean URL. The buyer never notices. Google does, and it hedges by ranking none of them well.

The Traps Worth Auditing First

The Fix List

Individually these look minor. Together they decide whether your crawl budget is spent on the pages you want ranked or wasted on duplicates and dead ends. Fixing them rarely produces an overnight jump, but it removes the ceiling that keeps a technically messy store from ever climbing, no matter how good the content on top of it is.

How to Audit These Traps Yourself

You do not need enterprise tooling to find most of these. The audit takes an afternoon and a crawler. Start with a site colon search for your domain and compare Google’s page count to the number of real products and collections you actually have. A large gap points straight at parameter and variant bloat. Then open Search Console Page Indexing and read the excluded reasons: a pile of pages marked duplicate, alternate with canonical, or crawled and not indexed is the fingerprint of the traps above.

Next, crawl the site and bucket the URLs by type, separating clean product and collection URLs from variant, filter, and catch-all paths, so you can see how much of the crawl is being wasted on duplicates. Check robots.txt for whether the parameter patterns are blocked, view source on a few blog and template pages to confirm each has exactly one real H1 describing the page, and load a collection on a throttled connection to see whether the full product grid is genuinely crawlable or quietly lazy-loaded into invisibility. Every problem on the fix list reveals itself in one of those checks.

Vanity Metrics vs Revenue: How to Read an E-commerce SEO Report Without Being Fooled

Every e-commerce SEO report has a number that looks great and a number that matters, and they are rarely the same number. Sessions are up, impressions are up, keyword counts are climbing, and revenue is flat or falling. I have read a lot of these reports for Shopify and D2C brands, and the single most useful skill is knowing which figures to trust and which are flattering noise. Here is how to read one honestly.

The Traffic That Does Not Sell

On a content-led store I analysed, the blog drove close to eight in ten organic sessions. Impressive on a slide. But almost every actual order came from a collection or product page. The blog was top-of-funnel traffic that informed and left; the money was made further down, on a fraction of the sessions. A report that celebrates total sessions, without splitting where revenue is made, tells you almost nothing about whether the store is healthier.

This is the core trap of e-commerce SEO reporting: the biggest number is almost always the least commercial one. Blog sessions, impressions, total ranking keywords, these scale easily and feel like progress. Revenue, conversion rate on commercial pages, and non-branded commercial rankings scale slowly and are what actually pay wages. A good report leads with the second set and treats the first as context.

The Metrics That Flatter, and What to Read Instead

How to Structure an Honest Report

None of this means blog traffic, impressions, or keyword counts are worthless. They are leading indicators, and they matter. But they are the top of the funnel, and a report that leads with them, without ever connecting them to revenue, is measuring effort instead of outcome. The fix is not fewer metrics, it is the right order: revenue first, then the commercial pages behind it, then the traffic that feeds them.

How to Read a Down Month Without Panicking

Every store has a month where the graph dips, and the difference between a good SEO relationship and a bad one is how that month gets read. The wrong response is to tear up the strategy. The right one is to ask, in order: did non-branded commercial revenue actually fall, or just total sessions? Did attribution shift, with direct or unassigned absorbing revenue that organic really earned? Is the whole category simply down this month, seasonal for everyone? Only once those are ruled out is it a performance problem worth reacting to.

On the store I analysed, a month where total sessions fell looked alarming on the summary slide, but the commercial pages were stable and the drop was concentrated in blog traffic and a seasonal dip across the whole category. Reading it by page type and by branded split turned a panic into a footnote. That is the entire value of an honest report: it tells you when to act and, just as importantly, when not to.

The Shopify Blog Engine Most Stores Ignore (And Why It Quietly Wins the Rankings)

Most Shopify owners treat the blog as decoration. A few recipe posts, a gift guide at Christmas, nothing that feels like real SEO. Then they wonder why an established store with hundreds of products cannot break page one for anything but its own brand name. Working on a Shopify home and lifestyle brand, I saw the pattern that most stores miss: the blog was not decoration, it was the single biggest source of organic visibility on the whole site.

Why the Blog Outranks the Store

When you actually measure where a content-led store ranks, the split is startling. On the brand I worked with, blog URLs held roughly two in three of the top-three positions and made up close to half of every ranking keyword on the domain. The collection and product pages, the pages the team obsessed over, were the minority at the top of search. The blog was doing the heavy lifting, and almost nobody had planned it that way.

The reason is intent. Product and collection pages answer commercial queries, the ones with the highest competition and the deepest pockets bidding against you. Blog content answers the informational queries that come first: how to style a small living room, what size photo frame for an A4 print, which cookware is safe for induction. Those searchers are not ready to buy yet, but they are your buyers, three touches earlier. Win them at the question stage and you are the brand they remember at the purchase stage.

What the Blog Actually Captures

How to Build a Blog Engine That Sells, Not Just Attracts

A blog that pulls traffic but never revenue is only half-built, and I will come back to that trap. Done properly, the blog is engineered around the buyer, not around whatever was fun to write. The stores that win treat it as a system.

The Catch Worth Knowing Up Front

Here is the honest caveat, and it is the subject of the next piece in this series: blog traffic is not blog revenue. On the store I described, the blog drove the overwhelming majority of sessions while almost every actual sale happened on a collection or product page. That does not make the blog worthless, it makes it the top of a funnel that has to be connected to the bottom. Build the engine, then make sure it is wired to the till.

Technical SEO First, Content Second: The Order Most Enterprise Teams Get Backwards

The most expensive mistake in enterprise SEO is not doing the wrong things. It is doing the right things in the wrong order. I have seen teams pour budget into content while the technical foundation underneath could not rank any of it, then conclude SEO does not work. The work was fine. The sequence was fatal.

Why Order Decides Outcome

SEO initiatives are not independent. They depend on each other, and some are prerequisites for others actually paying off. Build content on a site with broken indexing and the content joins the pile of pages Google never surfaces. Chase rankings while cannibalisation splits your signal and you fight yourself. Scale pages before fixing speed and every new page inherits the same ceiling. The right work in the wrong order produces motion without progress, and burns the budget and the patience you needed for the work that matters.

The Sequence That Works

This is why a serious enterprise SEO strategy is a prioritised backlog, not a list of ideas. Dozens of technical initiatives, each tagged by theme and quarter, shipped in an order that respects the dependencies. The order is not administrative detail. It is the strategy itself, and it is the difference between a program that compounds and one that spins.

If your SEO has been busy but flat, before you commission more content, ask a harder question: can the pages you already have actually rank? If the technical foundation cannot lift them, more content is the most expensive way to stay in the same place.

What Happens When You Skip the Order

The cost of wrong sequencing is not abstract. Ship content onto a site with broken indexing and the pages join the crawl-budget graveyard, unseen, and you conclude your content is not good enough when the truth is Google never rendered it. Chase rankings while cannibalisation splits your signal and you spend months watching two of your own URLs trade places on page two, neither ever reaching page one. Scale the catalogue before fixing site speed and every new page inherits the same slow ceiling, so the more you publish the more mediocrity you own. Each of these is a real failure mode I have been called in to reverse, and in every case the underlying work was competent. Only the order was wrong.

The Backlog Is the Strategy

A serious enterprise SEO plan does not look like a content calendar. It looks like a prioritised technical backlog: dozens of specific initiatives, each tagged by theme, owner, and quarter, sequenced so dependencies are respected. Structured data before content that relies on it. Cannibalisation fixes before ranking pushes. Speed and crawl control before scale. The document that matters is not a list of ideas, it is an ordered queue, and maintaining that order under pressure, when everyone wants the visible content win now, is a large part of what a consultant is actually for.

The test for whether your SEO is sequenced or scattered is simple. Ask why the current priority is the current priority. If the answer is that it unblocks the next thing, you have a sequence. If the answer is that it seemed worth doing, you have a to-do list, and on a large site a to-do list produces motion without progress.

How I Build the Sequence for a New Site

A sequencing strategy is not a philosophy, it is a document, and building it follows the same order every time. The point is to end with a queue where every item exists because it unblocks the next one, not because it seemed worth doing.

The Signals That Your SEO Is Scattered, Not Sequenced

You do not need an audit to sense when sequencing has gone wrong. The symptoms are consistent, and once you have seen them a few times they are hard to miss.

If more than one of those is true, the problem is almost never effort and almost never talent. It is order. Fixing the sequence, not adding to the pile, is what turns the same work into results.

None of this means content does not matter. It means content is the reward for a foundation that can carry it, not a substitute for building one. The teams that compound are the ones patient enough to fix the machine first and disciplined enough to hold the order when the pressure to publish arrives. That patience, more than any single tactic, is what separates SEO that grows from SEO that merely stays busy.

Index Bloat: Why Your Biggest SEO Problem Might Be Having Too Many Pages

Most site owners believe more indexed pages is better. On large sites, the opposite is usually true. Index bloat, thousands of low-value URLs crawled and indexed, is one of the most common and least understood problems in enterprise SEO, and I have watched it quietly cap sites that were doing everything else right.

How Bloat Happens, and What It Costs

Large sites generate URLs automatically: filtered category views, faceted navigation, search result pages, tag archives, pagination, session parameters, and near-duplicate product variants. Left uncontrolled, Google crawls and indexes tens of thousands of them. On one marketplace I worked on, a huge share of indexed pages should never have been indexed at all. The cost is real and threefold.

The Controls That Fix It

The instinct when traffic is flat is to add more pages. If you have index bloat, that instinct makes the problem worse. The fix is subtraction first: prune the index to the pages that deserve to be there, concentrate crawl budget and authority on them, and only then consider what genuinely new pages are worth adding.

A Real Pattern at Scale

To make this concrete without naming the client: on one global marketplace I worked on, the indexed page count ran into the hundreds of thousands, while the pages that earned meaningful traffic numbered in the low thousands. The gap was faceted navigation and filter combinations, every colour, size, and sort order generating its own indexable URL, plus tag archives and internal search pages Google had happily crawled and stored. The store had not published that content deliberately. The platform generated it, and nobody had told Google to ignore it.

The fix was not more content, it was subtraction. Deciding which filtered views deserved to be landing pages and blocking the rest, canonicalising duplicate paths, and applying noindex to the archives and search pages. Reclaiming crawl budget and concentrating authority on the pages that could actually rank did more for the site than any content campaign of that period. On large sites, the highest-leverage SEO work is often deletion.

How to Run the Diagnosis Yourself

You do not need enterprise tooling to find out whether you have this problem. The check takes an afternoon.

Crawl Budget: How Google Actually Spends Its Time on Your Site

Crawl budget is the amount of crawling Google is willing to do on your site, and on large sites it is finite. It comes down to two forces: how much Google can crawl without straining your servers, and how much it wants to crawl based on how fresh and valuable it judges your pages. Every request spent fetching a filtered URL that will never rank is a request it did not spend on a page that could. On a site of a few hundred pages this is invisible. On a site of hundreds of thousands, it is the ceiling.

You can see the problem in server logs and in Search Console. When a large share of crawl activity lands on parameter URLs, internal search pages, and endless filter combinations, while your genuinely important pages are crawled rarely, crawl budget is being spent on nothing. The symptom owners notice first is different, though: new content that takes weeks to appear in the index, because Google is busy re-crawling junk it was never told to ignore.

The Prune and Consolidate Playbook

Fixing bloat is not one action, it is a sequence, and the order protects you from removing something that was quietly earning. This is the playbook I run, biggest buckets first.

What Recovery Looks Like After a Prune

The first thing that improves is usually not rankings, it is crawl efficiency: important pages get crawled more often and new content gets indexed faster, because the budget is no longer wasted. Rankings follow over the next one to three months as the remaining pages absorb the internal authority once scattered across thousands of dead URLs. On the marketplace I described, reclaiming crawl budget and concentrating authority did more for the site than any content campaign of that period. The lesson holds: on a large site, the highest-leverage SEO work is often what you remove, not what you add.

International SEO Across Four Markets: What Multi-Domain, Multi-Language Growth Actually Takes

Most SEO advice assumes one site, one language, one market. The moment you cross a border, the rules change, and a strategy that worked domestically starts quietly breaking. I learned this running SEO for a marketplace across Mexico, Turkey, India, and Indonesia, multiple domains and languages under one roof, and I want to share what international SEO actually demands, anonymised and without the client’s numbers.

Where International SEO Breaks First

The failures are predictable, and they are almost never about content. They are about architecture, the signals that tell Google which version of your site belongs to which searcher. Get those wrong and your own pages compete against each other across borders, splitting the authority that should have lifted one clean winner.

The Decisions That Shape Everything

The through-line across all of it: international SEO is an architecture discipline first and a content discipline second. The stores that win globally are not the ones with the most translated pages. They are the ones whose structure tells Google, cleanly and consistently, which version serves which searcher.

A Practical Hreflang Checklist

Because hreflang fails silently, it needs verifying rather than trusting. This is the check I run on every international site, and it catches the errors that quietly cost markets their rankings.

Migration: The Highest-Risk Moment in International SEO

If there is one moment where international SEO equity vanishes, it is a migration: moving to a new domain, changing structure, or merging markets. I have watched a migration cause two countries’ traffic to blend so badly that the analytics were untrustworthy for months, with rankings transferring to the wrong market’s version. Migrations are not a technical formality. They are the single highest-risk event in an international program, and they need a redirect map, per-market parity checks, and a rollback plan before anything ships.

The rule I hold to: never migrate more than one variable at a time if you can help it. Change the domain or the structure or merge the markets, but not all three in one weekend, because when traffic moves you need to know which change caused it. On large international sites, that discipline is the difference between a controlled move and a quarter spent guessing.

Localise, Do Not Translate

The cheapest and most common international SEO mistake is running content through a translation tool and calling it localisation. Translation converts words; localisation adapts to how a market actually searches and buys. The terms differ, the currencies differ, the seasonal patterns differ, and the buying norms differ. A page that is linguistically correct but culturally and semantically wrong ranks for nothing, because it matches no real query. Real localisation usually means less content done properly per market, not everything duplicated in every language.

Domain Structure: The One Decision You Cannot Cheaply Undo

Before hreflang, before content, before a single keyword, international SEO forces one architectural choice: how to structure the domains. There are three options, and the gap between them is the gap between authority that concentrates on one strong property and authority that scatters across several weaker ones.

Country-code domains like example.mx and example.tr send the strongest local signal, but each is a fresh start with its own authority to earn, so a four-market rollout means four times the link building and four times the patience. Subdirectories like example.com/mx/ keep every market under one domain, so a link earned anywhere lifts them all, which is why they suit most companies expanding into new regions. Subdomains sit in the middle, one domain with markets held in separate sections. The right answer depends on how genuinely separate your markets are and how much authority you can afford to build in parallel, not on which option sounds most correct.

This decision matters more than any other because it is the one you cannot reverse cheaply. Moving between structures later is exactly the migration event that costs markets their rankings for a quarter or more. Choose it deliberately at the start, for the resources you actually have, not the ones you wish you had.

Measuring International SEO Without Fooling Yourself

The quickest way to mismanage a multi-market program is to average it. One blended dashboard hides the market quietly failing behind the market carrying the numbers, and by the time the blended line dips, you have already lost a quarter you did not need to lose. Every market needs its own baseline, its own targets, and its own report, read side by side rather than summed into a single comforting figure.

Read this way, the weak market surfaces early, while there is still time to intervene. Read as an average, it stays hidden until it is large enough to drag the whole program down, which is precisely when it is hardest and most expensive to fix.

The Position 11-100 Trap: Why Enterprise Sites Rank for Everything and Get Traffic From Nothing

A few years ago I led the SEO strategy for a global used-vehicle marketplace operating across four markets: Mexico, Turkey, India, and Indonesia. Multiple domains, multiple languages, a progressive web app, and millions of monthly impressions. On paper it looked like an SEO success story. It ranked for tens of thousands of keywords in every market. The dashboards were enormous.

Then you looked one level deeper and found the same quiet catastrophe in all four countries. I want to walk through it here, anonymised and with illustrative numbers rather than the client’s, because it is the single most common and most misunderstood problem on large sites, and almost nobody writes about it honestly.

Ranking for Everything, Getting Traffic From Nothing

Here was the pattern, repeated in every single market. The site ranked for somewhere between twenty-five and forty thousand keywords. Of those, a few hundred sat in positions one to three. A thousand or two sat in positions four to ten. And the overwhelming mass, twenty-thousand-plus keywords per market, sat in positions eleven to one hundred. Page two and beyond, where click-through rates round to zero.

The one-line summary I wrote against every market in that strategy deck was the same sentence: we are not capturing the traffic opportunity that already exists, because our rankings are not high enough to earn clicks. That is the trap. A site can look like an SEO giant in a rank tracker and behave like a ghost in the analytics, and the gap between those two views is thousands of keywords stuck one page too low to matter.

Why Enterprise Sites Fall Into This

Large sites, and marketplaces especially, generate pages at industrial scale. Every listing, every category, every filter combination, every location becomes a URL. Google discovers and ranks a huge number of them, but ranks most of them weakly, because the site never gave any individual page the signals it needed to climb. The footprint grows automatically; the authority to rank it does not. So you end up broad and shallow: present everywhere, strong nowhere.

The Fix: Technical First, Content Second, Never the Other Way

The instinct on a site like this is to write more content. That instinct is wrong, and expensively so. When twenty-eight thousand keywords are stranded on page two, the problem is almost never a shortage of pages. It is that the pages you have cannot climb, because the technical foundation underneath them is broken. So the strategy sequenced hard: fix the machine first, then feed it.

The Backlog Nobody Sees, and the Order That Matters

A real enterprise SEO strategy is not a document of ideas. It is a prioritised backlog of dozens of specific technical initiatives, each tagged by theme, owner, and quarter, shipped in sequence. Structured data, facet optimisation, pagination, breadcrumb restructuring, URL rebuilds, hreflang, the internal linking scheme, and a long tail of page-speed work from image decoding attributes to Brotli compression. Unglamorous, every one of them, and collectively the thing that lifts a stranded footprint off page two.

The order is the strategy. Fix indexing before you build content, or the new content joins the twenty-eight thousand pages nobody sees. Fix cannibalisation before you chase rankings, or you keep splitting your own signal. Fix speed before you scale, or every new page inherits the same ceiling. On a large site, sequencing is where consultants earn their fee, because doing the right things in the wrong order produces motion without progress.

What This Means for Your Site, Even If You Are Not a Marketplace

You do not need four markets and a progressive web app to be in the position 11-100 trap. Any site that has grown past a few hundred pages can have it: an e-commerce catalogue, a large blog, a programmatic content play, a directory. The diagnostic is simple and you can run it this week. Pull your ranking keywords, bucket them by position tier, and look at the shape. If the eleven-to-hundred bucket dwarfs everything above it, you are not under-ranked because you lack content. You are under-ranked because your existing pages cannot climb, and more content will make it worse, not better.

The fix is the same at every scale: technical foundation first so pages can rank, prioritisation second so you push the keywords closest to page one and worth the most, and only then content, aimed at the gaps that actually convert. That sequence is the difference between a rank tracker that looks impressive and an analytics dashboard that pays the bills.