TL;DR — On 31 August 2026 Search Engine Land published GMBapi’s Q2 year-over-year Google Business Profile cut: US calls −11.9%, website clicks −12.5%, direction requests +21.1%. Desktop Maps impressions flipped from −17.9% in Q1 to +3.2% in Q2. Rank trackers that only watch the Map Pack will call this a healthy listing. It is not. Customers are finishing more of the job inside Google Maps. Do not mix this with the 24 Aug reinstatement-reviews gap or the 25 Aug category study.
What changed on 31 August 2026
Michel van Luijtelaar (co-founder of GMBapi; disclosed in the Search Engine Land byline) published the Q2 numbers on Search Engine Land on 31 August 2026, 16:00 UTC. The same figures are on GMBapi’s own post (originally February 2026, updated through Q2).
The claim is not “local SEO is dead.” That was the April BrightonSEO read, based on an early Q1 cut where US actions and impressions looked roughly halved. GMBapi later revised Q1. Corrected US Q1 year over year: calls and website clicks each −15.8%, direction requests +31.3%. Q2 is the same shape, slower.
This is not a Google Search ranking incident. Google’s ranking status history still tops out at the August 2026 spam update (started 18 Aug, ended 21 Aug). Search Central’s newest documentation item on 31 Aug is the European Search Dataset Licensing Program page — DMA data sharing, not Business Profile actions.
Do not treat this post as a patch for:
- 3 Jul 2026 public reviews vanishing
- 9 Jul 2026 dashboard “You have no reviews yet”
- 24 Aug 2026 reviews missing after a suspension is reinstated
- 25 Aug 2026 specific-vs-generic category / completeness (1.8 million profiles)
- 28 Aug 2026 EEA site-reputation manual actions
Those are separate failure modes. This one is: the listing is live, the Map Pack looks fine, and calls are the wrong KPI.
The Q2 US numbers (year over year)
GMBapi, US Google Business Profiles, Q1 vs Q2 2026 against the same quarter in 2025:
| Action / surface | Q1 2026 vs Q1 2025 | Q2 2026 vs Q2 2025 |
|---|---|---|
| Calls | −15.8% | −11.9% |
| Website clicks | −15.8% | −12.5% |
| Direction requests | +31.3% | +21.1% |
| Desktop Maps impressions | −17.9% | +3.2% |
| Desktop Search impressions | +12.3% | +13.9% |
| Mobile Search impressions | −20.6% | −20.1% |
| Mobile Maps impressions | (noisy — ads) | +30.4% |
GMBapi’s Q1 US comparison used 5,000–10,000 US profiles. They flag mobile Maps impressions as a weak standalone metric: a few large advertisers switching Google Ads on or off move it. They also disclose the portfolio is roughly doubling locations year on year, and new trials import eighteen months of history — so the mix is not a frozen panel.
The read they (and Search Engine Land) put on Q2: US customers are going straight to Google Maps instead of starting on a Search results page, and the desktop Maps reversal says that is no longer mobile-only.
Sterling Sky / Jepto (179 Google Business Profiles) is the earlier context, not this quarter’s panel: AI-powered local packs often show two businesses instead of three, often without a click-to-call button, and surface 32% as many unique businesses as the traditional Map Pack. That is why a rank tracker can stay green while the phone goes quiet. Do not paste it onto the Q2 % table above.
The EU and the UK are not a delayed US
GMBapi split Q2 by region. It is not “the US first, Europe next year.”
EU — 11,073 locations (little paid-search experimentation vs the US):
| Metric | Q2 note |
|---|---|
| Calls | −5.2% |
| Website clicks | −5.1% |
| Direction requests | +21.7% in Q1 → +13.1% in Q2 |
| Desktop Maps impressions | −16.5% in Q1 → −34.7% in Q2 |
| Mobile Maps impressions | +13.5% |
| Mobile Search impressions | +22.4% |
EU calls/clicks are a milder drop. EU desktop Maps is the opposite of the US reversal — accelerating down. GMBapi leaves open whether that share is going to Apple Maps / ChatGPT or whether DMA rules give other directories more room. They do not pick an explanation.
UK — 833 locations (smallest cut; do not build a strategy on the Maps line):
- Mobile Maps impressions −70.8% year over year in Q2, vs +22.4% in Q1
- Direction requests still +35.8% (Q1 was +35.6%)
- Website clicks +15.3% in Q2 vs +2.7% in Q1
GMBapi’s own caveat: a handful of accounts toggling ads can swing 833 locations. What they will not explain away is the action side — directions still up, website clicks faster. If your UK report leads with Maps impressions, you will panic the wrong number.
Two systems: Map Pack rank vs whether AI will name you
Traditional Maps ranking is still proximity, relevance, engagement, prominence. Search Engine Land’s write-up of the same data: AI Mode and Gemini layer web context, entity matching, brand authority, and review sentiment on top of the Business Profile.
One system decides if you appear on the map. The other decides if Google’s AI will answer the customer without a click.
Reviews are the raw material for that second system. In Q2 Google prompted reviewers with structured tags (atmosphere, price, cleanliness) instead of free text only, and pushed customers to review places they had already visited. GMBapi adds: Google is using a customer’s own past reviews to prompt them back to the business inside Maps, and is pushing longer reviews. That is a plausible reason direction requests rise while calls and website clicks fall — the answer is on the listing.
SOCi’s 2026 Local Visibility Index (cited in the same Search Engine Land piece; not a GMBapi panel): AI platforms recommended 1.2% of locations on ChatGPT, 7.4% on Perplexity, 35.9% in Google’s 3-pack. Profile accuracy averaged 68% on ChatGPT and Perplexity vs 100% on Gemini (Gemini pulls Maps data). Locations ChatGPT recommended averaged 4.3 stars; Perplexity 4.2. That is a gate, not a nice-to-have star count.
This is not a reason to write Reddit threads or a thin “best of” listicle. It is a reason to keep review velocity and replies on a drip, and to keep the listing accurate — the same work as steady review velocity vs historical count and specific category + completeness.
What to do this week (and what not to do)
- Export calls, website clicks, and direction requests for Q2 vs last year. If the phone line is down and directions are up, the listing is being used as a map, not abandoned.
- Put direction requests on the weekly report. A calls-only dashboard is now a lie by omission.
- Do not buy LSAs from the April “local SEO is dead” slide. That slide used the uncorrected Q1 half-drop. This dataset does not justify a paid-ads pivot. Map Pack vs local PPC is a different argument (review velocity vs $3k/mo ads).
- Do not burst-request Google reviews to “fix Q2.” Google’s spam systems flag spikes after silence. After each completed job, send one WhatsApp review request, not a blast of old customers.
- Reply to what lands. What to automate vs approve still holds: draft fast on 4–5 stars, hold 1–3 for a human.
- Do not bulk-edit name, address, or category because calls dipped. Category work is the 25 Aug study. Missing reviews after reinstatement is the 24 Aug gap. This post is reporting.
At VijayaTech Labs the productized version of that drip — capture, reply, report — starts at $180/month.
Limitations
- GMBapi says the panel is biased toward home services, grows by roughly 2× locations year on year, and imports 18 months of history on new trials.
- Q1 was revised. Do not quote the BrightonSEO “roughly halved” figure as current.
- Mobile Maps impressions are noisy because of ads. GMBapi told readers not to lean on that line in Q1.
- UK = 833 locations. The −70.8% mobile Maps print is flagged by the authors as not yet a UK pattern.
- EU desktop Maps vs US desktop Maps is an open question (other apps vs DMA), not a finding.
- SOCi’s 1.2% / 7.4% / 35.9% and 4.3 / 4.2 star gates are a different study from the GMBapi Q2 % table. Do not merge the samples.
- Correlation of “Maps is the destination” is GMBapi’s read, not a Google ranking-document change.
- Alternatives that fail: ignore directions and declare the listing dead; buy LSAs from the uncorrected Q1 slide; blast past customers for replacement reviews; mix this with a reinstatement ticket.
FAQ
My Google Map Pack rank is unchanged. Why did calls drop?
Rank is not the journey. In GMBapi’s Q2 2026 US cut, calls were −11.9% and website clicks −12.5% year over year while direction requests were +21.1%. More of the visit is completing inside Maps.
Should I buy Local Services Ads because GBP calls fell?
Not from this dataset. The April “buy ads / local SEO is dead” message used a Q1 cut that GMBapi later corrected (US calls and clicks −15.8%, not roughly −50%). Q2 decelerated. LSAs are an ads product, not the review engine.
What should I report besides phone calls from the Google Business Profile?
Direction requests, on the same weekly line as calls and website clicks. Those three US Q2 figures were −11.9%, −12.5%, and +21.1%.
Is this the same problem as reviews missing after Google reinstated my profile?
No. Reinstatement with empty reviews is a separate recovery gap (24 Aug 2026). This post assumes the listing is live and the reviews are still there.
Sources
- Calls and clicks keep falling as Google Maps becomes the destination — Search Engine Land, Michel van Luijtelaar, 31 Aug 2026 16:00 UTC
- Local SEO in 2026: Why Your Rankings are Steady but Your Calls are Vanishing — GMBapi (Q2 figures; page updated August 2026)
- SOCi 2026 Local Visibility Index — cited in the Search Engine Land article for ChatGPT / Perplexity / 3-pack rates
- Latest Google Search documentation updates
- Google Search ranking incidents
- The Google Business Profile Algorithm in 2026
- Specific GBP Category Beats Generic
- Google Review & GMB Automation — VijayaTech Labs, pricing checked 1 Sep 2026