See what your Google profile actually earns you.

The screen most owners open and then close again

A physiotherapy clinic in Indiranagar, Bengaluru, ran three Google posts a month for a year and never once looked at what those posts did to calls. The owner's logic was reasonable enough: the phone rang sometimes, business was fine, why dig further. Then a new clinic opened two lanes away and the calls thinned out over eight weeks, slowly enough that nobody noticed until a slow Tuesday turned into a slow month. Insights & Performance exists for exactly that gap — the period between "something changed" and "I know what changed," which is usually where a business loses the most ground because nobody was watching the number that would have told them sooner.

The four numbers that matter — profile views, phone calls, direction requests, website clicks — are not interesting on their own. A view is not a customer. What's actually diagnostic is the ratio between them, and how that ratio moves. Rising views with flat calls means people are seeing the listing and not acting on it, which usually points at photos, hours, or a category that doesn't match what they searched for. Falling views with flat calls means fewer people are finding the listing at all, which is a different problem with a different fix — one lives in Rank OS, the other lives in the Demand Clusters read of what people are actually searching.

Reading the branded-versus-discovery split without fooling yourself

Branded searches — people typing your business name — will always look healthy if you've been open a while and have regulars. That's not a signal of growth, it's a signal of existing loyalty holding steady. The number worth watching is discovery: people who found you through a category or service query without knowing your name yet. A clinic whose branded searches are flat but whose discovery searches are climbing is actually winning new entity authority in its area. A clinic where discovery is flat or falling is invisible to anyone who isn't already a customer, no matter how good the branded number looks.

A false alarm worth naming up front: a single bad week, especially around a festival or a monsoon washout in a service-area business, is noise, not a trend. Google's own performance data lags by roughly a month, so a dip you notice today already happened weeks ago, and the fix for a lag like that is a custom date-range comparison against the same period last year or last quarter, not a panic reaction to last Tuesday.

What a healthy month looks like next to an unhealthy one

Put two months side by side and the difference is usually about movement, not size. A healthy month has views, calls and directions rising or falling together, roughly in proportion — that's a listing behaving the way you'd expect it to. An unhealthy month has one number moving alone: views up while everything else sits flat, or calls dropping while views hold steady. The second pattern is the one worth acting on, because it means the top of the funnel is fine and something further down is broken — a wrong phone number, hours that don't match reality, or a category mismatch that gets someone to click through and then realise this isn't quite what they needed. A single number moving on its own is a louder signal than the whole set drifting together, even when the drift looks scarier on a graph.

A worked read: a diagnostic lab in Ahmedabad

Picture a diagnostic lab in Navrangpura, Ahmedabad, running home sample collection alongside its walk-in centre. Profile views were up 22% quarter over quarter, which looked like good news until the owner pulled the month-on-month search breakdown export and saw that direction requests hadn't moved at all — people were finding the listing and not planning a visit. Website clicks, on the other hand, had climbed sharply. That combination usually means the audience wants to book online rather than walk in, which for this lab meant the fix wasn't more visibility, it was making the online booking link impossible to miss on the profile itself — a change that runs through Edit Location, not through Insights. The metrics didn't hand over the fix directly. They ruled out three wrong guesses and left one right one.

A second decision this data actually settles

There's a recurring argument inside multi-location businesses about which outlet needs the marketing budget this quarter, and it usually gets settled by whoever argues loudest rather than by the numbers. Comparing the branded-versus-discovery split across locations settles it more honestly: a location with strong branded search and weak discovery has a loyal base but isn't winning new customers, while a location with the reverse pattern is attracting new people but hasn't built repeat demand yet. Those need different fixes — the first needs more posting and category work to get found, the second needs the kind of reputation work Reviews & Brand is built for, since repeat business tracks with trust more than with visibility. Treating both locations the same because they're both "underperforming" wastes budget on the wrong lever at whichever one gets it wrong.

Where this doesn't help

Insights & Performance reads what Google's own Performance API exposes for the profile — it does not see website analytics, ad spend, in-store footfall, or calls that came through a number printed on a hoarding rather than the Google listing. If a business relies mainly on WhatsApp enquiries or a call centre number that isn't the one attached to the profile, this dashboard will systematically understate real demand, and no amount of staring at the export will fix that; it needs to be paired with whatever the business already tracks off Google. It also won't tell a multi-location brand which single location is dragging down an average — for that, the comparison has to happen at the Agency OS portfolio level, filtering listings individually rather than reading one aggregate number. And it measures what happened, not why an AI-generated answer chose to mention a competitor instead — that's a separate, structural question the measuring AI search visibility pillar covers in more depth than a performance dashboard ever could.

What it feeds

Freshness is one-fifth of the Rank OS score, and Insights is the evidence that freshness work is landing — a post that doesn't move views probably wasn't the right post. The branded-versus-discovery split is also the first clue that a demand cluster is scaling before the keyword volume itself catches up, since discovery searches often rise a few weeks ahead of a cluster getting flagged. And because the numbers here are first-party and GBP-native rather than modelled, they're the honest check against everything else the platform reports — including Geo-Grid claims about where you rank, citation fixes meant to lift trust, and local ranking factors generally, none of which mean much if the ranking isn't converting into calls and directions. For agencies, the same month-on-month export becomes the backbone of a white-label report a client can read without needing an explanation, and for anyone learning to read the raw numbers for the first time, the reading GBP performance data guide is a reasonable place to start before diagnosing anything.

If there's one habit worth building from this page: check the date range before believing the number. A month-old lag hiding inside a "last 7 days" view has talked more owners into the wrong fix than any single bad metric ever has.