Section 1
The evidence that the pack is the firebreak
The sharpest data on this comes from a Whitespark study published in May 2025. Researchers ran 540 local queries across Houston, Phoenix, and Denver in six service categories, including plumbers, dentists, and personal-injury lawyers. Across all queries, AI Overviews appeared 68 percent of the time. Taken alone that number reads like bad news for local firms, and plenty of coverage stopped there. The finding that matters is the split underneath it. The study found an inverse relationship between Overviews and local packs, sorted by intent. On explicit local queries, the "plumbers in Phoenix" shape, an AI Overview appeared only 15 percent of the time while a local pack appeared 93 percent of the time. On informational queries, the "how much do lawyers charge" shape, the numbers flipped: Overviews on 92 percent, local packs on 6 percent. The 68 percent headline is an average that sits on top of two nearly opposite behaviors. Where the searcher signals they want to hire someone nearby, Google still shows businesses, not a summary. Where the searcher wants a fact, Google answers and the visit ends. That inverse relationship is the whole thesis in one data point. The queries closest to a booking are, for now, the least colonized by the answer box. The informational traffic that AI deleted was never the traffic that booked jobs anyway. It was the top-of-funnel that fed the funnel. The bottom of the funnel, the hire-intent local search, still resolves to a pack. And the pack converts. About 44 percent of local searchers click a result in the local 3-pack, against roughly 29 percent for the organic blue links below it, 19 percent for paid, and 8 percent for the "more places" expansion, based on aggregated local-pack behavior data. Within the pack, the top listing draws around 17.6 to 17.8 percent of clicks, the second about 15.4 percent, the third about 15.1 percent. Businesses that sit in the pack pull an estimated 126 percent more traffic than those ranked four through ten. The pack is not a nice-to-have. On a local query it is where the attention is, and it is the surface the Overview has not swallowed.
Section 2
The framework: why it survives, where it breaks, and who you are competing against
Run three lenses on the pack and state what each one cannot see. The pack looks like a gift. Treating it like one is how you lose it. Lens one: why the pack resists the answer box (comparative statics) Move the same variable that broke everything else, "Google now answers in-box," and trace it through the pack specifically. It does not move the pack much, and the reason is structural. Google's own three local ranking factors are relevance, distance, and prominence. Two of those, distance and prominence, are not facts a model can generate. Distance is a physical relationship between the searcher and a business. Prominence is an accumulated reputation, built from reviews, citations, and real-world signals that took years to earn. An Overview can restate a specification. It cannot manufacture proximity, and it cannot fabricate a decade of five-star reviews for a business that does not have them. So on a hire-intent query the cheapest correct answer Google can give is still a ranked list of nearby, well-reviewed firms. The pack is the output of a computation the answer box cannot shortcut. As the surfaces around it decay, the pack also gains relative value. When the ten blue links below lose a third to two thirds of their clicks to the Overview, the pack's share of the remaining attention rises. The firebreak is not just holding. It is becoming a larger fraction of what is left. This lens gets the direction right and stops there. It tells you the pack is resilient today. It does not promise the interface stays put, and the interface is exactly what Google controls. Comparative statics, the resilience lens • Assumes: a stable results interface where the pack keeps its current role. • Fits because: distance and prominence are computations prose cannot replace, so hire-intent queries still resolve to businesses. • Breaks when: Google redesigns the local result, folding the pack into an AI-brokered or fully paid surface. • Counteracts: the panic that "AI killed search," by isolating the one surface it structurally cannot answer. • May reinforce: complacency that the pack is permanent, when it is a product decision away from changing. Lens two: the pack's own tipping points (threshold) The pack is resilient against AI. It is not unconditional. It has three thresholds of its own, and missing any of them takes you out regardless of how good your work is. The first is proximity. Distance is a ranking factor, which means there is a geographic band around the searcher inside which you can rank and outside which you effectively cannot for that search. A shop across town from a dense customer cluster is fighting physics. This is why service-area reality, where your Business Profile says you operate and where you genuinely have address and review density, is not a cosmetic setting. It is the threshold that decides whether you are even eligible for a given searcher's pack. The second is prominence, and it behaves like a threshold rather than a dial. Reviews are the largest lever most operators can actually move, and there is a tipping quality to them. Below a certain review count and rating relative to your local competitors, you are invisible in the pack no matter how relevant you are. Cross it, and you become a default. Google has also confirmed that behavioral engagement with your listing, the calls, the direction requests, the clicks, feeds back into ranking. Prominence compounds. The firms already in the pack get the engagement that keeps them in the pack. The third threshold is the one Google controls, and it is the dangerous one. Google is steadily inserting paid surfaces above and around the pack. Local Services Ads now sit at the very top of many local results, above the organic pack, and their cost per lead has been climbing as displaced demand crowds in, with home-services LSA leads averaging around 53 dollars in early 2026 and competitive metros running 90 dollars and up. Every paid pixel Google adds above the free pack pushes the free result down the page. The pack is free today. The share of the screen it gets for free is shrinking. Threshold, the pack-vulnerability lens • Assumes: the pack has discrete tipping points in proximity, prominence, and Google's own monetization. • Fits because: distance eligibility and review-count effects behave as thresholds, not smooth gradients. • Breaks when: you treat "rank in the pack" as a single achievement rather than three separate gates you must clear and hold. • Counteracts: the belief that good service alone earns the pack, when eligibility and prominence gate you first. • May reinforce: over-indexing on review count while ignoring the slower monetization threat above the pack. Lens three: the operators chasing the citation (behavioral) The third lens points at your competitors and at the advice they are following. The loud consensus right now is answer-engine optimization, getting named inside the AI answer. It is seductive because it feels like fighting the machine on its own ground. It also quietly abandons the surface that still produces the call. Every hour and dollar spent engineering a citation in a summary that ends the session is an hour not spent on the Business Profile, the reviews, and the local signals that still convert. Pew found only 1 percent of AI Overviews produced a click on the cited source. Being named in the box is close to a billboard on an empty road. Meanwhile the map pack, the surface the herd is drifting away from, is where 44 percent of local clicks still go. The opportunity in the herd's mistake is direct. While competitors chase citations, the operator who does the unglamorous local-SEO work faces a thinner field for the surface that actually books jobs. The behavioral edge is not cleverness. It is refusing the fashionable channel and owning the boring one. Behavioral, the consensus lens • Assumes: operators over-weight the novel, socially-endorsed tactic and under-weight the boring one that still works. • Fits because: AEO is the loud advice while its click economics are near zero and the pack's are not. • Breaks when: an operator has real conversion data separating pack-driven calls from AI-driven ones and can judge on evidence. • Counteracts: the pull to spend the recovery budget on citations because the trade group is talking about them. • May reinforce: dismissing AEO entirely, when it has a genuine defensive role on branded and reputational queries. The structure-break flag Here is what all three lenses depend on, and the honest limit on the word "free." The pack is free of a per-lead charge today because Google has chosen to render hire-intent queries as an organic list. That is a product decision, not a law. Google has monetized every high-intent surface it has ever owned. It put ads on top of organic search, then more ads, then the shopping carousel, then Local Services Ads on top of the local pack. The trend line points one way. The "last free channel" is a lease, not a deed, and the landlord has a history of raising the rent. Own the pack as if it is temporary leverage you are using to build something you actually control, not as if it is a permanent home.
Section 3
The playbook: levers, then a dated portfolio, then a history check
The levers, ranked by what moves the pack Pull these in order. The early ones gate the later ones. Eligibility first. Claim and fully complete your Google Business Profile: correct primary category, accurate service areas that match where you truly operate, hours, phone, and a booking or call action. An incomplete or miscategorized profile is not eligible for packs it should win. This is the cheapest, highest-return move on the list and most firms do it halfway. Prominence second, and this is the lever with the most room. Build a post-job review request into your close-out process so review velocity is steady rather than sporadic, because recency and count both feed prominence, and respond to reviews so the profile reads as actively managed. Keep your name, address, and phone consistent across the web so Google trusts the entity. Add real photos, because engagement with the listing is itself a ranking signal. Engagement third. The calls, direction requests, and clicks your listing earns feed back into ranking, so the profile should be built to earn them: clear photos, current hours, a reason to call now. Prominence compounds through use, which means the work of getting into the pack and the work of staying in it are the same work. Everything above the pack, treated as a throttle, not a base. Local Services Ads and paid local can buy you the top slot while you build the organic pack underneath. Use them to bridge, cap the spend, and watch the cost per lead, because the whole reason to own the free pack is to stop renting the paid one. The dated portfolio Do now, zero regret. Complete the Business Profile and switch on a post-job review request today. Both are cheap, reversible, and correct in every future, whether Google monetizes the pack next year or leaves it alone. Every review you bank now is prominence you keep. Hedge, cheap insurance. Run Local Services Ads or paid local at a capped budget as a bridge while your organic prominence builds, so a slow month does not starve the pipeline. Bounded cost, real protection against the gap between "started the local-SEO work" and "ranking from it." Defer, with a trigger. Do not pour money into a full multi-location local-SEO agency retainer or a review-generation platform until your own manual process proves the pack converts for you. Pre-commit the trigger: "when my Business Profile is driving a countable share of booked calls and I have hit the ceiling of what I can do by hand, then I buy the tooling." Scale what works. Do not scale a hypothesis. The history check The reference class for "a free Google surface" is not encouraging, and it is the honest base rate. Organic search was once a free channel that firms could live on. Google added ads above it, then more, until the free result sat below the fold for commercial queries. The local pack itself used to show seven listings, then Google cut it to three and put Local Services Ads on top. The pattern is consistent: Google keeps the free surface useful enough to hold attention, then monetizes the attention. The base rate says the pack's free run has a clock on it. Which is exactly why the pack is a means, not an end. Use its resilience now, while it lasts, to build the two things Google cannot lease back from you: a base of repeat customers and a reputation that shows up wherever they look. The operators who treated free Facebook reach or free organic search as permanent got hurt when the rent came due. The ones who used the cheap window to build owned demand came through.
Section 4
A note for UK and EU operators
The mechanics of the Google pack are the same in the UK and EU, but two things shift the picture, and a companion piece handles them in depth. First, UK service demand runs heavily through directory platforms like Checkatrade, where members pay in the region of 60 to 120 pounds a month plus per-lead fees, which means the pack competes for attention not only with Google's own paid surfaces but with an entrenched paid directory the way US firms compete with Angi. Owning your Google Business Profile is how you stop paying that directory rent. Second, the regulatory environment is different: AI Overviews reached EU users roughly nine months after the US and in a limited set of member states, held back by the Digital Markets Act and related rules, which has so far given EU firms a slower, more staggered version of the collapse. That delay is a window, not a reprieve.
Section 5
What this framework cannot see
Name the blind spots. This read assumes Google keeps rendering hire-intent local queries as an organic pack, which is a product decision that could reverse the day Google decides to broker local bookings itself. It leans on a small, US, three-city Whitespark sample and on aggregate click studies, so your category and geography could behave differently, and the only ground truth is your own profile insights and Search Console. It treats reviews and prominence as earnable through honest work, which is largely true but sits next to a real and ugly problem of review manipulation and extortion that a local-SEO playbook does not solve on its own. And it assumes proximity works in your favor, when for a firm physically outside its best market's search radius the distance factor is a wall no optimization climbs.
Section 6
The fitness test
You should make the map pack your primary investment if you serve a defined local area, your customers hire on a "near me" basis, and your Business Profile is currently incomplete or under-reviewed. For you the pack is the highest-return, least-exposed surface available, and the field is thinning as competitors chase citations. Spend the next quarter clearing the eligibility and prominence thresholds and you win a surface AI cannot take. You should not lean on the pack, and should invest elsewhere first, if your business is not proximity-gated, if you sell regionally or nationally rather than to a neighborhood, or if you are physically outside the geographic radius of the market you want. Distance is a ranking factor you cannot optimize away, and for you the pack is a wall, not a door. Build owned demand and referral systems instead. Either way, stop treating the map pack as a box to check on the way to the fashionable AI work. It is the one high-intent surface the answer box cannot replace, it converts better than anything around it, and it is free for now. Own it while that sentence is still true, and use the window to build the demand Google cannot lease back.