Google Search Console: the free tool most businesses ignore

Almost every business either has it unverified or has never opened it, while paying monthly for rank tracking tools that estimate what Search Console reports directly.

It is not a full picture. It covers organic Google search only, it samples and filters, and it will not tell you about revenue. But inside its boundary it is first-party data from the source, which puts it in a different category from almost everything else you will look at.

What can Search Console tell me that nothing else can?

Four things, and the first two are the reason to bother.

The actual queries that showed your site. Not keywords a tool guessed you might rank for. Real searches, with impressions, clicks, average position, and click through rate. This is the closest thing you have to a transcript of what your market types.

The gap between showing and being clicked. A query with many impressions and few clicks means you appear and people choose somebody else. That is a title and description problem, or a review count problem, and it is fixable without ranking higher.

Which pages Google indexed and which it refused. The Pages report lists every page it declined to index and why. Businesses regularly discover half their site is not eligible to appear at all.

Whether Google can crawl what you published. Broken pages, redirect chains and blocked resources surface here before they surface as a traffic decline.

The queries report is the one to open first, and almost nobody does. Sort by impressions rather than clicks. The top of that list is demand you are already visible for and not yet capturing.

How do I read the queries report without getting lost?

Set the date range to the last three months and add the comparison to the previous period. Then run three passes.

Pass one, brand versus non-brand. Filter for your business name. Everything matching is people who already knew you. Everything else is discovery. The ratio between those two is the single most useful number in the tool, and it is the number most reports quietly blend together to make organic performance look stronger.

Pass two, positions four to fifteen. Filter for average position greater than three and less than sixteen. These are queries where you are visible but below the fold. Improving a page you already rank for is far cheaper than creating demand for a new one.

Pass three, high impressions and low click through. Sort by impressions, look down the click column. Where it is near zero, read your own title and description for that page as a stranger would.

Take that list into your analytics rather than treating it as separate. A small set of GA4 reports will tell you what those visitors did after they arrived, which is the half Search Console cannot see.

How do I turn pass two into a ranked list of jobs?

With your own click-through rates, not with anybody's published curve. Here is the arithmetic.

Query A. 1,240 impressions, 18 clicks, average position 6.4. That is a click-through rate of 1.5%derived.

Query B, on a page you already rank well for. 320 impressions, 44 clicks, average position 2.1. That is 13.8%derived.

The model. If Query A moved to the same average position as Query B and earned the same click-through rate, 1,240 impressions at 13.8%derived would be 171 clicks, a gain of 153 a quarter over the 18 you get now.

Now state the model's limits out loud, because they are large. You are using one site's observed rate at one position, on a different query, assuming position is the only thing that differs. It is not a forecast. It is a way of ranking candidate jobs so the biggest one goes first.

Run it across your pass two list. Impressions times your own best observed rate, minus current clicks. Sort descending. The top three are your quarter. Everything below that is a distraction, and you now have a defensible reason for saying so in a meeting.

Why can nobody tell me what a top local pack position is worth?

Because the data to answer it does not exist, and the tables in circulation were invented. This is the single most useful thing to know about the boundaries of Search Console.

Start with what the platforms publish. Search Console has no local pack dimension at all. Google Business Profile performance reports publish views, searches, calls, website clicks, direction requests, messages and bookings, and publish no rank and no position. So neither of Google's two free reporting surfaces can pair a local pack position with a click.

That means the well known organic CTR studies structurally cannot produce a local pack table. Backlinko and Advanced Web Ranking both use real Search Console data, and both are organic positions only, because Search Console gives them nothing else. Backlinko now cites FirstPageSage in its own introduction, which makes that citation circular.

Then look at the numbers quoted at you:

The one traceable study is BrightLocal's Local Services Ads click study from October 2018, which found the local pack taking 28.8% of clicks with LSAs present and 32.3% without. Its method, in BrightLocal's own words, was asking 5,500 Mechanical Turk testers to look at five search result pages. Five queries, all San Francisco, desktop only, in 2018, on a layout that has since absorbed local pack ads and AI-powered local packs.

The operating rule: do not convert a ranking change into a traffic forecast using any published CTR curve, because there is not one. Measure the client's own Business Profile actions across a ninety day before and after, pair it with grid rank data, and state the two series as correlated rather than as a formula. Saying that out loud is a differentiator, because every competing proposal will contain a table from the list above.

What are the limits of the data?

Real limits, and knowing them is what stops you from over reading it.

It filters low volume queries. Rare searches are withheld for privacy, so query totals never sum to click totals. That is by design.

Average position is an average. A position of 6.2 might mean you were consistently sixth, or third in one city and eleventh in another. Local businesses see a lot of that, and a change made in one location behaves differently in another partly for this reason.

It covers Google organic only. No Maps interactions, no paid, no Bing, no direct traffic.

It is delayed. Two to three days, so it is not a monitoring tool.

It does not report AI answer inclusion in any usable form. Google Search Central announced Search Generative AI performance reports in June 2026, giving impressions by page, country, device and date. Clicks, click-through rate, position and the user's prompt are not available, AI Mode is not separated from AI Overviews, and the data also remains inside the overall performance report, so it is a separate view rather than a separate dataset. Do not double count it. This is the honest position and it matters: you cannot use Search Console to prove AI visibility. Measuring AI search visibility honestly requires sampling and server logs, not this report.

Sixteen months of history is the maximum. Export regularly if you want anything longer, which you will the first time you try a real year over year comparison.

Can I see the effect of AI Overviews at all?

Indirectly, with one method that has been validated, and it is worth doing because impressions and clicks have visibly come apart.

Ahrefs compared 150,000 keywords showing AI Overviews against 150,000 matched informational keywords without them, using aggregated Search Console click-through rate for March 2024 against March 2025, and measured a 34.5% decrease in clicks. A later cut of the same work revised the figure to 58%, and SparkToro's panel work with Similarweb puts the drop around 60% when an AI Overview is present. Quote the range, never a point. The spread reflects different baselines, dates and query mixes, and all of it is vendor research.

The usable method inside your own account is Ahrefs' proxy. Isolate the queries most likely to trigger an AI Overview, meaning informational, question-starting and non-branded, then compare their click-through rate before and after a known rollout date while holding average position constant. Validated against confirmed AI Overview keywords, the guessed set showed a 42% relative click-through drop and the confirmed set 44%, a difference of only 2% relative. That is close enough to be useful and it costs nothing but an export.

Keep the local context attached. Ahrefs' analysis of 146 million search result pages found AI Overviews on 20.5% of all searches but only 7.9% of local ones, and Sterling Sky independently observes AI-powered local packs on roughly 7% of tracked keywords. For a local business this is a real effect on informational content and a small one on money queries.

How should I use it month to month?

Lightly, and with the same discipline you would apply to any small sample.

Look at non-brand impressions as your visibility trend and non-brand clicks as your capture trend. Those two lines, tracked monthly, tell you most of what search is doing for you.

Compare to the same period last year rather than to last month wherever your business has a season. Seasonality distorts adjacent month comparisons enough to reverse the apparent direction of a trend, and search demand is seasonal in almost every industry.

Resist the urge to react to individual queries moving. Position data is noisy at the query level. Watch the aggregate.

And when you make a change to a page, use Search Console to see whether that page's impressions and clicks moved, understanding you are looking at a before and after rather than a controlled test. Genuine A/B testing does not work at most local business traffic levels, so a before and after with a long enough window is often the best evidence available to you. Say that out loud rather than dressing it up as a test.

What to do this week

Verify the property if you have not. Domain property rather than URL prefix, so it covers every subdomain and both protocols. Verification is a DNS record and takes ten minutes.

Add yourself as an owner, not a user, and make sure ownership sits with your business rather than with an agency or a former developer.

Open Queries, set three months, filter out your brand name, and read the top thirty. Write down three queries where you have impressions and almost no clicks.

Run the ranked list arithmetic on those three, using your own best observed click-through rate. Pick one.

Open Pages under Indexing and read the not indexed list. Fix anything important sitting in there.

Then export the data. A monthly export into a folder you own gives you the long history the interface will not.

Questions about your Search Console data?

Email me at eric@seod.com with a screenshot of your Queries report, three month range, sorted by impressions, brand name filtered out if you can manage it. I will tell you which three queries you are closest to winning and what specifically is standing between you and the click.

I answer these myself. It takes a few minutes and there is no followup attached to it.

More on measurement sits in the analytics library.