GBP Call Clicks Are Falling. Here's the Data and What's Driving It.
Your profile still sits in the top three for your trade and your town. Views look normal. Yet the Calls figure in your Performance tab has been sliding for months, and you have checked everything twice.
You are probably not imagining it, and probably not the cause. Evidence published since December 2025 points at Google's own results pages: the call button is vanishing from map results, ads that keep theirs have spread, and a newer AI layout shows fewer businesses.
Most of that evidence is American, so each figure below says where it comes from. Every Google help page cited was checked on 19 October 2026.
The short version
- Sterling Sky's December 2025 charts for 179 profiles across 34 US law firms show January to October call clicks about 15% down on 2024 (our sum), while website clicks held. We found no UK equivalent.
- In Smallbiz Edge's March 2026 study of 2,580 mobile searches across 15 cities, 19.8% showed a call button in the map pack. Plumbing searches showed one twice in 90.
- Ads kept their call buttons and spread through 2025, though the one before-and-after study of Local Services Ads, from 2020, found a small effect.
- A fall in Search Console impressions in September 2025, or from late April 2026, was most likely a reporting change, not lost visibility.
- Google's Calls figure counts taps on a call button that has gone from most trade map results. Count the calls themselves.
Are Google Business Profile call clicks really falling?
In the one multi-profile dataset we could find, yes. Sterling Sky, a local SEO agency in Canada and the US, asked the reporting platform Jepto to chart Business Profile clicks for 179 profiles across 34 law firms in the USA, from January 2024 to October 2025. Clicks-to-call declined. Website clicks did not.
Joy Hawkins published the charts in The State of Local SEO in 2026 on 17 December 2025, and the page has since been redated to June 2026. Adding up the monthly totals printed on the charts, January to October call clicks fell from 101,237 in 2024 to 86,368 in 2025, a drop of about 15%. Website clicks over the same months rose about 2%, from 99,709 to 101,300, though they too slipped from June 2025. Those sums are ours.
Hawkins says Sterling Sky had noticed the trend all year, with calls falling even on profiles whose rankings held. Her reading of the split: website buttons sit on desktop listings and call buttons on mobile ones, so the loss looks specific to mobile.
These are American law firms, not trades, and one dataset. We found nothing comparable for the UK, so ask anyone quoting a UK percentage where it came from.
Where has the call button gone?
Off most map results for trades. In a study published in March 2026, the marketing agency Smallbiz Edge searched 172 keywords on a mobile screen in 15 cities: 13 in the US, plus London and Toronto. Of 2,580 results, 510 showed a call button in the map pack, or 19.8%.
Every trade below scored under 7%, against 19.8% across all 41 niches. The rates are pooled across the 15 cities.
| Trade in the study | Results with a call button |
|---|---|
| Roofing | 1 of 90 (1.1%) |
| Plumbing | 2 of 90 (2.2%) |
| Electricians | 2 of 60 (3.3%) |
| HVAC (heating and air conditioning) | 5 of 90 (5.6%) |
| Locksmiths | 4 of 60 (6.7%) |
| House cleaning | 0 of 60 |
The pattern followed a label. Where Google headed the map results "Places", 76.2% had a call button. Under "Businesses", the label most trade searches got, 0.9% did. Service-area businesses kept the button 11.6% of the time, against 26.5% for businesses customers visit. Emergency searches did no better, at 19.0%.
London scored highest of the 15 cities at 24.4%, against 19.2% across the US cities, which the authors think may reflect Google changing the US first. One city is a pointer, not a UK benchmark.
Two other changes point the same way. Sterling Sky's December article showed Google swapping call buttons for images in some industries' local packs, with a handyman search from late 2025. In April 2026, Search Engine Roundtable reported Google testing a profile panel that hides the call button behind a numbered "show more" button. We found no Google announcement of either.
Why that shows up in your Calls figure
Google's Calls metric is a tally of taps, the times someone "clicked on the call button". With no button in the map results, a customer has to open your profile first, or read your number and dial it, and dialled calls were never counted. Each extra step probably loses some people. Nobody has measured how many still ring.
Are ads taking the calls instead?
Some, probably, because ads kept their call buttons. Across the queries Sterling Sky tracks, Local Services Ads appeared on roughly 11% of them at the start of 2025 and 31% by November. Paid ads inside the map pack went from 1% of its mobile reports at the start of 2025 to almost 22% by December.
Sterling Sky works in Canada and the US and does not split these by country, so they tell you nothing about the UK.
When Local Services Ads launched for US lawyers in 2020, Sterling Sky compared 110 lawyer listings before and after. Calls as a share of non-branded searches fell from 1.32% to 1.24%, about five fewer actions a month per business, two of them calls. Real, but small, and the only before-and-after measurement we found.
In the UK, Google's Local Services list includes plumbers, electricians, roofers, heating and air conditioning, cleaning and 23 other home-service categories. Locksmiths and painters and decorators are not on it. You are charged for each valid lead.
What do AI Overviews and AI local packs have to do with it?
Less for your phone than for your blog. AI Overviews mostly answer questions, and a question usually comes a step before hiring anyone. Local Falcon's May 2025 whitepaper, covering 60,000 local queries for 4,423 businesses in 20 countries, saw them on 58.3% of informational searches and 17.2% of commercial ones.
Sterling Sky blames them for falling blog traffic. We found no study of their effect on Business Profile calls. Getting your business named inside Google's AI Overviews is a separate job, and mostly the same Google work you already do.
AI local packs are the bigger worry, for now in the US
These are a version of the map pack generated by Google's AI. When Sterling Sky wrote about them in December 2025, it had seen them only on mobile and only in the US, on about 7% of the keywords it tracks through the rank tracker Places Scout. They show one or two businesses, different ones from the normal pack, and no call buttons.
Places Scout counted 5,943 unique businesses in the AI packs against 18,330 in normal three-packs. That is about a third as many (5,943 ÷ 18,330 = 32%), and 88% of the 322 markets checked had fewer businesses in the AI version. Sterling Sky's own January 2026 follow-up calls this "32% fewer". Its raw counts mean about 68% fewer.
Did your impressions drop too?
If the fall is in Search Console impressions, it may not be real. Two reporting changes cut impression counts without anybody losing a position: Google ending support for a results-page parameter called num=100 in September 2025, and a Search Console logging fix in April 2026. Neither concerns your Business Profile's calls.
What num=100 actually was
num=100 was a URL parameter: added to the address of a Google results page, it showed 100 results instead of 10. It stopped working in mid-September 2025, and a Google spokesperson told Search Engine Land it was "not something that we formally support."
Rank trackers used it to load 100 results at a time. By Google's definition, Search Console counts an impression whenever your result is on the current page of results, scrolled to or not. So, as Brodie Clark explained, a tracker loading a 100-result page logged an impression for a site at position 99. When that stopped, Tyler Gargula of LOCOMOTIVE found 87.7% of 319 Search Console properties lost impressions, and 77.6% lost unique ranking terms (Search Engine Land, September 2025).
The April 2026 logging fix
Google's data anomalies page records a logging error that stopped Search Console "accurately reporting impressions from May 13, 2025 until April 27, 2026". With it fixed, "you may notice a decrease in impressions". Clicks were not affected; CTR and average position were. Search Engine Land reported that past data stays uncorrected.
Across either date, judge your website by clicks, not impressions.
How should you measure calls now?
Stop reading Google's Calls figure as your call count. It counts button taps, and with fewer buttons it can fall while your phone rings as often as ever. Count the calls themselves, then use the profile's other numbers to work out which kind of drop you have.
If Google's figure is all you have, start by tracking clicks-to-call properly, alongside the calls that actually reach you. A forwarding number on the profile, or asking every caller how they found you, gives a count Google's layout cannot move.
Then line up 12 months of views, website clicks and calls. The table is our reading of the evidence above, not a published method.
| What fell | Most likely reason | Check first |
|---|---|---|
| Calls only; views and website clicks held | Fewer call buttons in the results | Your trade and town on your phone, then your tracked calls |
| Calls and website clicks; views held | People see you but do not act | The profile itself |
| Views as well as calls | Lost visibility | Rankings, the category, any suspension |
| Search Console impressions only | A reporting change | Search Console clicks |
The first row matches the US law-firm data: calls down, website clicks steady (views were not published). If views held but website clicks fell too, people are seeing you and not acting, so work through the profile itself. If views fell, start with rankings, the category and whether the listing has been suspended. Google's performance figures also mix "organic search results and Google Ads", so ad spend moves them.
What can a UK trade actually do about it?
You cannot put the button back. You can make sure a customer who taps through to your profile, instead of ringing from the map, finds a reason to call. You can test paid placements against booked jobs, and skip the "more profiles" shortcut unless you genuinely have more than one location.
Make the profile do the selling. When calling takes an extra tap, people read first: photos of real jobs, your services, your reviews. How recent your reviews are counts there as well as how many you have, which is why a steady trickle beats an annual batch.
Put a tappable number on your mobile site's first screen. Some customers will reach you through the website button instead.
Test ads against your own numbers. Sterling Sky says clients running ads now get a better return, because ads keep the call buttons organic listings are losing. That is one agency's observation, with no figures published. Judge any ad on booked jobs, not leads.
Do not invent locations. Sterling Sky's answer to a single profile losing traffic is to open more locations, meaning real ones. Google's guidelines allow "one service-area Business Profile" for a business without a shopfront, and a profile per location only with "separate service areas and separate staff". Virtual offices are not eligible. A one-van trade gets one profile.
What to do this week
- On your phone, look up your main job plus your town. Note the label above the map results and whether anyone has a call button. Repeat monthly.
- Pull 12 months of Performance data. Views, website clicks and calls, side by side. Find your row in the table.
- Count real calls for a fortnight. A forwarding number, a call log or a tally of "how did you find us", set against Google's figure.
- Re-read your Search Console graph. Mark September 2025 and late April 2026, then judge by clicks.
- Check Google's UK Local Services list for your trade before anyone sells you ads.
- Change the profile only if your row says so. If calls fell on their own, the profile is probably not your problem.