The Vertical Search Gap

What this study measures — and what it deliberately does not

The leader in the generic search is absent when it counts

Across 42 market-and-vertical combinations in 14 competitive U.S. metro markets, the integrator that dominates the generic “commercial security systems” search was absent from the vertical-specific search 90.5% of the time — and where any vendor did occupy that vertical search, it was almost always a competitor who got there with the bare minimum of effort.

A ranking is not a customer

Nobody has ever made a payroll on “we come up first on Google.” A high ranking on a page that doesn't convert the buyer who clicks it is a vanity metric — and arguably worse than absence, because it represents time and money spent to produce nothing.

So we want to be clear about what a search ranking is: it's the entry point, not the end-game. It's the door, not the room. We measured the door in this study — specifically, we measured how astonishingly poorly defended the door is — because almost no one in this industry has bothered to walk through it properly. But winning the door is only worth something if you can convert the buyer who walks through it, and that is a separate discipline this study does not measure. We'll return to it at the end. Hold the distinction as you read: the fact that these searches are winnable is the beginning of the opportunity, not the whole of it.

How buyers actually search — and why it exposes a gap

When a facilities director, a school superintendent, or a property manager goes looking for security help, they rarely search “commercial security systems.” They search their problem: a system for a school district, a hospital, a multifamily property. Those vertical-specific searches are the ones that immediately precede a real buying decision — the searcher isn't browsing, they're solving something.

The question this study asks is simple: when the buyer searches that way, does the market's most visible integrator — the one who wins the generic term — show up? And if not, who does?

The sample was fixed before we looked

We wanted a result that couldn't be dismissed as cherry-picking, so we fixed the sample before we looked at anything.

Markets

We took the largest U.S. metro areas by population, split them into two competitive tiers (large and mid-sized — we deliberately excluded small markets, where thin competition would confound the finding and which aren't representative of the markets integrators actually fight over), and selected markets by a fixed, mechanical rule applied before any searching began. Fourteen markets, seven per tier. The list was frozen before the audit started; nothing was added or dropped based on what we found.

Verticals

Three, applied uniformly to every market: K-12 schools, hospitals/healthcare, and multifamily property. Every vertical was run in every market — no selecting which verticals to report per market, which would be a backdoor to cherry-picking.

The instrument

In each market we first identified the “incumbent” — the local or regional integrator that leads the generic head-term search. Then we ran each of the three vertical searches and recorded who appeared, sorting each result into one of four outcomes:

1. The incumbent is present on the vertical search.
2. A different vendor serves it — a non-incumbent local player, or a national firm with a location page.
3. The buyer's own institution occupies the result — the school district's or hospital's own page ranks, with no vendor present.
4. No relevant vendor appears at all — the result is filled by guard-staffing job listings, generic national boilerplate, or a manufacturer's blog.

A limit we state plainly

The automated first pass could not perfectly replicate a localized, personalized-off browser, and it returns “first-returned” results rather than a confirmed ordinal rank. So we treated the automated pass as a first sweep to find the pattern, and then manually verified every apparent vacancy and every featured example in a real, localized browser before trusting it. That verification changed our numbers — see below — and the change is itself part of the finding.

The dominant reality, then, is this: the incumbent has vacated the vertical search, and the space is usually occupied not by no one, but by someone who did very little to claim it. In the large majority of cases where a vendor ranked for the vertical term, that vendor had done the bare minimum of on-page work — the vertical named in the page's URL, its headline, and its copy; a basic landing page; and little else. Several ranking “vertical pages” were, on inspection, a single paragraph about the vertical bolted onto an otherwise generic commercial page. That was enough. The competitive floor to occupy these high-intent searches is on the ground — and the incumbents, overwhelmingly, didn't clear even that.

A meaningful share of these occupying vendors weren't local at all — they were national firms whose templated location pages surfaced ahead of every local integrator in the market. In those cases, a company that isn't even from the market is winning the local buyer's vertical search, because no local player contested it.

The competitive floor is on the ground — and the incumbents didn't clear even that.

The one market where content was excellent — and still lost

In one large market, we found something that at first looked like a contradiction — and turned out to be the clearest confirmation of the deeper point.

Here the incumbent had built a genuinely excellent, buyer-first healthcare security page: zone-by-zone planning for patient flow and restricted clinical areas, a substantive discussion of access-control modernization, privacy-aware video guidance that explicitly warned against lazy “HIPAA-compliant” marketing claims, a buyer's evaluation checklist, proposal red-flags to watch for, and a real FAQ linking to industry and federal standards. This is, more or less, exactly the kind of content that genuinely serves a buyer trying to understand their problem.

And it still lost. The incumbent's page ranked third on the vertical search — behind a competitor whose own “healthcare” page was noticeably thinner (its supporting case studies were for an entirely different vertical) — and the incumbent was omitted entirely from the AI-generated answer summary at the top of the results, which named other providers.

Real, but conditional

Read carefully, the honest takeaway is not “nobody does good work” and it is not “the lane is empty, walk right in.” It's more useful than either:

So the opportunity these findings describe is real but conditional. The searches are winnable, often easily. But winning them is worth doing only if you take the buyer all the way across the finish line — which means doing the ranking work and the buyer-first content work together, not one or the other. That integration is where the actual value sits, and it's the thing this industry, on the evidence, almost universally leaves undone.

Ranking is the entry point. Conversion is the finish line.

We'll say it once more, because it's the frame for everything above: a first-place ranking on a page that doesn't help the buyer is a vanity metric. This study measured the entry point — and found it poorly defended — precisely to raise the more important question it does not answer: of the pages that win these searches, how many actually serve the buyer who clicks? That question is the subject of separate SMR research on how integrators present vertical expertise once a buyer arrives. Ranking gets the buyer to the door. What's behind the door is what pays.

Don't take our word for any of this

The method above is reproducible in about thirty minutes, from your own browser, on your own market and competitors — and what you find may match ours or may not, because every market differs. That's the point: the method reveals what's true for you.

What the method can and can't say

Everything needed to check us

This appendix contains the full method so any reader can reproduce or challenge the study. Company names and specific markets are anonymized in this edition; the method is complete and the searches are exactly reproducible. Run it yourself and compare.

A. The frozen market list

Markets were selected by a fixed rule applied before any searching began, so the sample could not be shaped by what we found. We took the largest U.S. metropolitan statistical areas by population (U.S. Census Bureau population estimates), split the top 66 into two competitive tiers — Large (population rank 1–33) and Mid (rank 34–66) — and within each tier selected every 5th market by rank, starting at the 3rd position. That rule yields 14 markets (7 per tier). Small markets were deliberately excluded: thin competition there would confound the finding, and they aren't representative of the markets integrators compete over.

The rule is deterministic — anyone applying it to the same population ranking reproduces the identical list.

The specific 14 metros are named in the full non-anonymized edition. The selection rule above lets any reader regenerate the exact list from public Census population data. The list was frozen before the audit began. Nothing was added or dropped based on results. Where a selected market lacked an identifiable local/regional head-term incumbent, it was handled by a pre-stated rule, documented, not silently swapped.

B. The exact search phrasings

One head term to identify the incumbent, plus three vertical-intent terms, run identically in every market with [city, state] appended:

All three vertical terms were run in all 14 markets — no selecting which verticals to report per market. Total: 14 markets × 3 verticals = 42 market-vertical pairs.

Search conditions. Searches should be run in a clean, logged-out, personalization-off state with location set to the target market — not the researcher's location. Results are a point-in-time snapshot; positions vary by location and change over time. Date every run.

C. The four-outcome scoring rubric

Each market-vertical pair is sorted into exactly one outcome, based on who appears for the vertical term relative to the head-term incumbent:

Two dimensions were recorded per pair: whether the incumbent serves the vertical, and whether any local integrator does (the “empty lane vs. already-claimed” distinction).

Scope discipline. This rubric measures who ranks — a single variable. It does not score content quality or conversion. A thin page that ranks still counts as outcome 2; that a minimal-effort page can win is itself part of the finding, addressed in the discussion, not smuggled into the score.

D. Verification protocol — how the numbers were corrected

An automated first pass produced the initial dataset. Because automated search cannot perfectly replicate a localized browser and cannot confirm true ordinal rank, we then manually verified, in a real localized browser: every apparent vacancy (outcome 4) and every example featured in the report. Verification reclassified five of seven initial “vacancy” calls to outcome 2 after finding a ranking vendor the first pass had missed — which is why the reported genuine-vacancy figure (~5%) is lower than the first-pass figure (17%). The corrected, lower number is the one reported. The reclassification is documented pair-by-pair in the full edition's record.

E. Per-pair record (anonymized)

The complete 42-pair record — each market-vertical pair, its outcome category, the incumbent-level and category-level findings, the first-returned result, and supporting evidence — is maintained in full. In this anonymized edition it is summarized by pattern.

Outcome distribution across 42 pairs:

• Outcome 1 — incumbent present4 · 9.5%
• Outcome 2 — other vendor serves29 · 69.0%
• Outcome 3 — institution-owned2 · 4.8%
• Outcome 4 — no relevant vendor / genuine vacancy2 verified · ~5%

Outcome-2 count reflects post-verification reclassification (down from 7 first-pass vacancies); the pre-verification distribution and the full named per-pair table appear in the complete edition.

Headline metrics:

• Incumbent absent on vertical terms — consistent across both tiers38/42 · 90.5%
• Genuine vacancy (no vendor at all) — corrected down from 17% first-pass~5%
• Where a vendor serves (33 pairs): incumbent ~12% · non-incumbent local/regional ~73% · national-with-location-page ~15%

Headline figures rest on all 42 pairs and are reported with confidence. The finer four-outcome breakdown rests on smaller counts and is directional — the shape of the pattern, not precise proportions.

F. Reproduce or challenge it

Everything above is the complete instrument. Take the selection rule, regenerate the market list, run the four exact search phrasings in a clean localized browser, score each result against the four-outcome rubric, and compare your distribution to ours. Your specific results will differ — markets change, and search is localized and personalized — but the method is the method. If the pattern doesn't hold when you run it, that's worth knowing too. That's the point of publishing the instrument: so you don't have to take our word for any of it.