Category used to be cosmetic. It isn't anymore.
For most of Google Business Profile's history, primary category was a modest local-pack ranking input — worth getting right, rarely worth agonizing over, and easy to treat as a brand-consistency question: pick the category that best represents the company, apply it everywhere, move on.
That changed when Google confirmed AI Overviews now pull directly from Business Profile data to decide which businesses are eligible to appear in an AI-generated local answer. The mechanism isn't a ranking adjustment — it's an inclusion gate. A location's category is matched against the query's intent as a controlled vocabulary, not evaluated the way page content or brand messaging is.
The practical consequence: a multi-location business that enforces one primary category across every branch, in the name of consistency, is now silently excluding whichever branches don't actually match that category from an entire class of AI-driven queries — not because the branch does the work poorly, but because the category field says it doesn't do the work at all.
The mismatch nobody notices until it's diagnosed
Franchise and multi-location brands standardize on one category deliberately — it's simpler to manage, avoids inconsistent branding across the profile network, and was, until recently, a reasonable tradeoff.
A concrete example of the failure mode
A plumbing franchise standardizes every branch on "Plumber" as primary category. One specific branch, due to its local market, actually does 60% of its revenue in drain and sewer work — a distinct, separately searched service. Categorized as "Plumber" only, that branch is structurally excluded from AI Overview answers to drainage-specific queries, even though it's the branch best equipped to handle them.
Why this is invisible without auditing
The branch's website, reviews, and general search rankings look fine — nothing about its organic performance signals the problem. The exclusion is specific to AI-answer eligibility for the mismatched service category, which doesn't show up in standard rank-tracking the way a lost keyword ranking would.
Why brand consistency isn't the right frame here
Consistency is a reasonable goal for logo usage, tone, and NAP data. Category is a factual claim about what a specific location actually does, and factual claims should vary by location precisely when the underlying business does.
Auditing category against revenue mix, not brand guidelines
The fix isn't picking a better single category — it's treating category as a per-location question with a factual answer, audited against what each branch's revenue mix actually looks like, not against what the brand would prefer every branch to represent.
Pull revenue mix by service line, per location
Not company-wide averages — the specific branch's actual service breakdown. A branch materially over-indexed on a secondary service relative to the brand average is a candidate for a different primary category, or for using GBP's secondary-category slots deliberately rather than as an afterthought.
Use secondary categories as real inclusion levers, not filler
GBP allows multiple categories. Secondary categories extend eligibility for AI-answer inclusion beyond the primary match — treating them as a genuine part of the eligibility strategy, rather than something added once and forgotten, closes some of this gap without abandoning a sensible primary category.
Re-audit when a location's business genuinely shifts
A branch that adds a new service line or sees its revenue mix shift needs its category reassessed — this isn't a set-once decision, it's a periodic check tied to how the actual business at that location is changing.
If a specific location's real revenue mix diverges meaningfully from the brand's default category, that divergence should show up in its GBP category — treating it as inconsistency to correct is exactly backwards now that category is a matching mechanism, not branding.
What this means for how multi-location profiles get managed
This doesn't mean abandoning standardization everywhere — NAP consistency, review response cadence, and photo quality genuinely benefit from a uniform playbook across a location network. Category is the one field where the correct answer is allowed, and often needs, to be different per location.
For a franchise or multi-location operator managing dozens or hundreds of profiles, the practical shift is treating category audit as an ongoing operational task tied to real business data, not a one-time setup decision made at brand launch and left alone. The businesses catching this early are the ones whose AI-answer visibility won't quietly erode as this mechanism becomes a larger share of how local search actually surfaces answers.
We do this for a living
If you'd rather not build this yourself, these are the services where it lives.