Google Business Profile Calls Fell 11.9% in Q2 While Direction Requests Rose 21.1%. Your Dealership Counts One of Those.

Three numbers came out on August 31 that describe what is happening to the bottom of your funnel.

Comparing the second quarter of 2026 against the same quarter of 2025 across US Google Business Profiles, calls fell 11.9 percent. Website clicks fell 12.5 percent. Direction requests rose 21.1 percent. The figures come from GMBapi, a profile management platform reporting on its own managed locations, and were published in a contributor column at Search Engine Land alongside the fuller writeup on the company's own site. Both carry the same byline, so treat this as one source telling you about its own book of business.

Rankings did not move to produce that. That is the part worth stopping on. The same profiles held their positions and the actions underneath them split in two directions at once.

Here is the more interesting thing about the report, and the reason to take the smaller number seriously. In April, the same company stood on the BrightonSEO stage and declared local SEO was dying, because its early read of Q1 showed that "actions and impressions appeared to have fallen by roughly half" across its US portfolio. It advised some of its agency customers to move budget into paid search on the strength of it. Then Google revised the Q1 data retrospectively, the company rechecked, and the halving vanished. Its own sentence in the August piece: "We were wrong about the scale."

Somebody publishing a correction that embarrassing about their own conference keynote is the closest thing to a quality signal you get in this business.

The corrected picture, and why it reads differently

The rechecked Q1 2026 comparison, drawn across 5,000 to 10,000 US profiles, came out at website clicks and calls each down 15.8 percent, with direction requests up 31.3 percent. Desktop Search impressions were up 12.3 percent, mobile Search impressions down 20.6 percent, and desktop Maps impressions down 17.9 percent.

Q2 tested that corrected baseline and the shape held while the magnitude eased. Calls down 11.9 percent, website clicks down 12.5 percent, direction requests up 21.1 percent, all three decelerating against Q1.

The impression side is where the story stops being about decline. In Q2, US desktop Maps impressions flipped from down 17.9 percent to up 3.2 percent. Mobile Maps impressions were up 30.4 percent. Desktop Search was up 13.9 percent. Mobile Search was down 20.1 percent, and because mobile Search is by far the largest impression category, that single number is enough on its own to erase the gains everywhere else.

Read those together and the pattern describes a business whose visibility moved from one Google product to another rather than one that lost it. GMBapi's read is that US consumers are increasingly going straight to Google Maps instead of Google Search when they want something local, and that the desktop Maps reversal means this is no longer a mobile-only behavior.

Maps used to be the doorway. People searched, saw a map pack, and clicked through to a website or a phone number. Maps is becoming the destination. People open it, search inside it, evaluate inside it, decide inside it, and the only thing they do on the way out is tap the button that starts navigation.

The structural half of it

Behavior explains part of this. Google's own layout changes explain the rest.

Sterling Sky and Jepto's analysis of 179 Google Business Profiles found that AI-powered local packs frequently show two businesses rather than three, often arrive without a click-to-call button, and surface only 32 percent as many unique businesses as the traditional Map Pack. Most rank trackers cannot see any of it, which is how a store stays green on a ranking report while its phone goes quiet.

Two things follow for a dealership specifically.

Going from three slots to two is a harsher cut in automotive than in most verticals. A metro with six same-brand rooftops competing on the same "[brand] dealer near me" query already had a brutal three-seat auction. Removing a seat removes a third of the field.

And the missing call button is a design decision with a revenue direction. GMBapi's own recommendation, stated plainly, is that "Google Ads, specifically Local Services Ads (LSAs), are essential to retain the prominent 'call' buttons that organic listings are losing." That is a vendor telling you to buy the thing you used to get free. Weigh it accordingly, and note that it is also probably true.

What this means for your dealership

Your profile report will look like a loss and will not be one. A store reading a monthly Business Profile summary this fall is going to see calls and website clicks each down roughly an eighth against last year, with no ranking change to blame it on. That is the industry number rather than a failure of execution. Before anyone gets fired over it, read our piece on why dealers keep firing the wrong vendor, because this is the exact shape of chart that starts those conversations.

The action that grew is the one your store cannot see. A direction request is a shopper tapping a button that loads your address into a navigation app. There is no form submission, no phone log entry, no chat transcript, no CRM record. It puts a person on your showroom floor forty minutes later with no digital fingerprint attached. Every other action in your funnel deposits a row in a system somewhere. This one deposits a person in your parking lot.

Which means your attribution is now understating the channel by design. If calls and clicks are down 12 percent and directions are up 21 percent, and your scorecard counts the first two and ignores the third, your Business Profile looks like a declining asset in a quarter where it produced more high-intent behavior, not less. That gap is what happens when the customer's preferred action stops generating a record.

We need to update something we told you in July. In our piece on Google citing itself inside AI Mode, we wrote that if site sessions soften while profile actions hold or climb, you are meeting the shopper somewhere else rather than losing them. That advice needs a revision. Profile actions are no longer moving as one number. Two of the three are falling and one is rising, and reading them as a single total will hide the whole story. Split them on the report.

Directions to the wrong place are worse than no directions. This is the piece that is uniquely automotive. Most dealerships are not one building. Sales, service, parts, collision, and quick lube can sit at different entrances, on different sides of a lot, or at a satellite address entirely. If your pin drops a service customer at the sales entrance at 7:15am, or routes a shopper to the rooftop you closed in 2023, you just converted the fastest-growing action on your profile into a bad first impression. Nobody catches this, because nobody is monitoring the metric.

The counter-case, honestly

This is one company's dataset and it deserves the caveats it states about itself.

The data is "slightly biased towards the home services sector," in GMBapi's own words, and no automotive breakout exists. The clicks-to-call collapse Sterling Sky measured was on Jepto-managed law firms. Nobody has published what any of this looks like for a franchised dealership, including us.

GMBapi also notes that its managed-location count roughly doubles year over year, so the portfolio being compared is not a stable panel, and it says outright that "changes in GMBapi's own customer portfolio likely explain some of the remaining gap." It declines to lean on the mobile Maps figure alone because a handful of large advertisers switching campaigns on or off can move it. There was a Google API bug in June that dented the series.

The overseas picture argues against treating this as a universal law too. In the EU, desktop Maps impressions accelerated downward from negative 16.5 percent to negative 34.7 percent, the opposite of the US reversal, while direction request growth slowed. The EU and UK are on different tracks rather than trailing behind the US on a delay.

So do not adopt these numbers. Use them as a hypothesis to test against your own profile, which is the only dataset that pays your bills.

What to do about it

• Export your own Business Profile performance, Q2 2026 against Q2 2025, and split the three actions. Calls, website clicks, and direction requests, on three separate lines. If your directions line is up while the other two are down, you are in the pattern and you now have the number that proves the channel is working.
• Put direction requests on the monthly scorecard with a stated value. Pick a number your GM will defend, even a conservative one, and hold it constant. An action nobody assigns a value to gets cut in the first budget review.
• Build the bridge from directions to walk-ins. Add a navigation option to your "how did you hear about us" field and make the desk log it. It is imperfect and it is far better than the zero you have now.
• Audit every pin, address, and department entry this week. Sales, service, parts, collision, each rooftop. Drive the route your own profile gives a customer at 8am on a Saturday. This is a ninety-minute job that protects your fastest-growing action.
• Check your hours the same way. A direction request is a customer committing to a drive. Wrong hours turn that into a locked door and a one-star review.
• Take the review changes seriously, because they feed the second system. Google now prompts reviewers with structured tags rather than free text and nudges customers to review places they have visited. Traditional Maps ranking still runs on proximity, relevance, engagement, and prominence. The AI layer sitting on top adds web context, entity matching, brand authority, and review sentiment. Clean review volume is what makes an AI confident enough to recommend you without sending the shopper to check.
• Stop grading the channel on rank. Most trackers cannot see the AI local pack at all. A number one position in a three-pack a shopper never loads is not a result.

The bottom line

The panic version of this story ran in April and its own author retracted it. Local search is not collapsing for dealers.

What is actually happening is narrower and more useful. Google moved the local shopper's journey inside Maps, trimmed the seats, removed some of the buttons that generated records, and left standing the one action that hands you a customer without telling you it did. Your profile is producing more intent and less paperwork.

The stores that lose the next four quarters will be the ones that read a falling call count as a falling channel and defunded it. The stores that win will be the ones that fixed the pin, valued the directions, and kept the profile clean enough for Google's AI layer to answer with confidence.

If you want to see what that layer actually knows about your store right now, across your profile, your site, and everything feeding both, run your dealership through our AI compatibility test and let's look at the real numbers instead of last year's scorecard.