Content cannibalization is the default state of a dealer group site.
Keyword cannibalization on a dealer group site is structural, not accidental.
The platform generates a model landing page. Someone adds a location page for the same model. The inventory filter produces an indexable URL for the same thing again. Then a decade of blog posts sits underneath all of it, written by three different vendors, answering the same buyer question in three different voices.
Google picks one of them. It is frequently not the one that converts, and often it is the homepage, because the homepage is the only URL on the domain with enough authority to be unambiguous. That is why a store can hold a respectable number of rankings and still find most of them resolve to a single page that was never built to close anyone.
The cost is not that the extra pages rank badly. It is that the authority which should have compounded into one strong answer is split across four weak ones, and every new page published without checking makes it worse.
This is also the layer where AI search visibility and classic car dealership SEO diverge least. An engine that cannot tell which of your four pages is the canonical answer resolves the ambiguity the same way a buyer does, by picking someone else. How citation works →
- Model landing page
- Platform-generated, thin, templated across every rooftop in the group.
- Location model page
- The same copy with a city name swapped in. Reads as a duplicate because it is one.
- Inventory filter URL
- Indexable, parameterized, and changing every time stock turns.
- Legacy blog post
- Written in 2019 by a vendor nobody at the group remembers hiring.
All four are answering the same buyer. Google publishes one and the other three spend authority to no effect.
The content cannibalization map is the first deliverable.
Every URL on the domain, inventoried against every query that has clicks or impressions in Google Search Console, with overlap scored and consolidation candidates flagged. Nothing else in the content workstream starts until this exists.
- Which pages overlap, and by how much. Two pages sharing a handful of long-tail impressions are not a problem. Two pages splitting the same commercial query are. The map separates them so the work goes where the money is.
- Which page should survive. Decided on existing performance and link history, not on which one somebody likes. The surviving URL is usually the older one with the inbound links, even when the newer one reads better.
- Which queries the domain already earns. These are cheaper to win than anything net-new, because the site is already in the index for them and merely sits past the click threshold.
- Which queries it does not earn at all. The genuine content gaps, scored against what the group actually has on the lot rather than against national search volume.
For a group, the same map settles which rooftop owns which query, which is the allocation dealer group SEO has to make before any of it gets written. The map is a document you keep, not a slide in a monthly deck. What the deliverables look like →
The cannibalization map is the first thing the audit produces. Which of your pages are competing with each other for the same buyer, and which one should own each term.
Get the free auditThree things happen after the map.
Consolidation
Merge or redirect the overlapping pages so one canonical answer compounds authority instead of three half-versions fragmenting it. Unglamorous, produces no new page to show anyone, and usually the highest-return action available in the first quarter.
Inventory-aware content
New pages mapped to what buyers are researching and scoped against what the group actually sells. A page targeting demand the store cannot fill converts nobody, however well it ranks. The keyword library is the segmentation this runs on. See the categories →
Plan discipline
Each month's actions come from Search Console data and from what competitors are ranking for, not from a preset editorial calendar. Calendars are how dealer sites accumulate the cannibalization problem in the first place.
The number of content actions per month tracks your tier and what the market demands. The qualifying test is the same at every tier: every new page either earns a query you do not already win, or it should not exist. Content is one workstream among the automotive SEO services in each tier, and it is sequenced behind whatever the technical and local reads say is blocking it.
The content types that work for a dealership, and the one nobody counts.
Most dealership content plans produce the same four things: model overviews, seasonal service tips, community posts and manufacturer press releases rewritten. Three of those rank for nothing. The types that earn queries share one property, which is that they state a fact a buyer cannot get from the manufacturer.
Model against rival, trim against trim, with the differences stated plainly. This is the highest-volume research intent in automotive and the place a dealer site can genuinely out-specify a national publisher, because the store has the vehicles.
What a job costs, what a trim adds, what ownership runs locally. The most asked and least answered question class in the category. A published range beats a contact form every time it is compared.
Registration, inspection, emissions and titling for the state, plus seasonal advice specific to the market. Nobody national can write it, and it earns civic references.
Reviews, questions, delivery photos and owner commentary. The only content on the site written by someone with no incentive to sell, which is exactly why engines weight it.
That fourth category is the one nearly every plan omits, because no vendor invoices for it. User-generated content on a dealer site does three jobs at once: it adds text an engine can read on pages that would otherwise be templated, it supplies the independent corroboration that entity resolution runs on, and it answers the specific questions buyers ask in the words they ask them. A review that names the vehicle, the salesperson and the job is a more citable sentence than anything the marketing department will write.
Making it usable is mostly mechanical. Reviews should render as text in the page rather than inside a third-party widget an engine cannot read. Owner questions and answers belong on the model or service page they concern instead of in a support inbox. Delivery photos need real alt text naming the vehicle. And review requests that ask about a specific job produce specific sentences, where a generic request produces a star rating and nothing to quote. What is not acceptable is gating, filtering or incentivising, which is an advertising compliance question before it is an SEO one. How reviews compound → Where compliance sits →
Measuring whether the content did anything.
Dealership content reporting almost always counts output. Pages published, words written, posts scheduled. None of those is a result, and a plan judged on them will keep producing the pages that caused the cannibalization problem.
- Query count per page, not sessions per page. The question a content plan has to answer is whether a page earned queries the domain did not already win. A page drawing traffic on terms three other pages already ranked for has added nothing and has probably subtracted. Search Console answers this at page level, and it is the only report that tests the qualifying rule directly.
- Consolidation counted as a win. When a merge moves a set of queries onto one URL and the total improves, that is the highest-return content action available and it produces no new page to point at. A reporting model that only counts publications will score the best month of the year as the least productive one.
- Citation rate on the questions the content answers. Research content increasingly gets read by an answer engine rather than a person, and the return arrives as a mention rather than a session. Tracking whether the store gets named on its own category questions captures the part analytics cannot see. How citation is measured →
- Assisted contribution, on a window that fits the purchase. Vehicle research happens weeks before a lead. Any model crediting the last touch will report research content as worthless, which is how good content plans get cancelled a quarter before they pay. The full return calculation →
Ninety days is too short a window for this layer, and every honest version of this report says so up front. Impressions and query count move first, position on commercial classes follows, and the lead effect lands last.
Questions about automotive content strategy.
What is a dealership content strategy?
A dealership content strategy is the standing decision about which pages the site publishes, merges, or retires, made against the queries the domain already earns rather than against a topic list. For a dealer group it starts with the content cannibalization map, because most groups already own more pages than they can rank and consolidation is the faster gain. New pages are then scoped to inventory the group actually sells, so the plan does not chase demand the stores cannot fill.
What is keyword cannibalization on a dealership website?
It is what happens when several pages on the same domain compete for the same query. On a dealer group site it is usually structural: a model landing page, a location page for the same model, an indexable inventory filter, and a legacy blog post all answering the same buyer question. Google picks one, frequently not the one that converts, and the authority that should have compounded into a single answer splits across near-duplicates instead.
What is a content cannibalization map?
Every URL on the domain inventoried against every query with clicks or impressions in Google Search Console, with overlap scored and consolidation candidates flagged. It is the first content deliverable in an engagement, because every later decision depends on it. You cannot decide what to publish until you know what you already have and which of it is fighting itself.
Should a dealer group delete old blog posts?
Usually merge or redirect rather than delete. A legacy post that ranks for nothing and duplicates a stronger page should be redirected into that page so its links and history survive. Deleting throws away signal for no gain. The exception is content that is factually wrong, or that describes vehicles and offers the group no longer carries, where leaving it published is a liability in its own right.
How many content pieces per month does a dealer group need?
Volume tracks the tier and what the market demands, not a fixed quota. The test that does not change is the qualifying one: every new page either earns a query you do not already win, or it should not exist. Publishing to hit a number is how the cannibalization problem gets built in the first place.
Where do the monthly content decisions come from?
From Search Console data on your own domain and from what competitors in your market are actually ranking for. Queries where the site already has impressions but sits past the click threshold get worked first, because they are cheaper to win than net-new topics. Nothing on the plan comes from a template shared across clients. How the competitive read feeds it →
Does content strategy cost extra?
No. It is included in the published tier price: $1,165 per rooftop per month at Core, $1,615 at Compete, $2,425 at Command, with a banded volume discount for groups (10% off at 2–3 rooftops, 20% at 4–7, 28% at 8–14, 32% at 15+) and a flat one-time onboarding fee charged once per group. Month-to-month, no annual contract. Full pricing →
Find out what your domain is already fighting itself for.
The baseline audit covers AI citation first, but the Search Console read that comes with it is usually where the uncomfortable part shows up. Most groups discover they have been paying to publish pages that compete with pages they already own.