Reading Search Console as a business owner: which data helps you decide

Search Console is one of the most useful tools a company already has available, and one of the least used: you open it, look at a chart going up or down, and close it. The difficulty isn't technical but one of reading: four numbers placed side by side seem to say one thing, while they actually describe different stages of the same path. This guide helps you draw observations from it that you can bring to a meeting and turn into checks.

The dashboard shows what happens on Google's results: how many times you appear, how many times someone chooses you, for which searches. It doesn't say what happens after the click, and that's the first thing to keep in mind so you don't credit or blame it for something it isn't responsible for.

Four numbers and what they really say

Taken in order, the four numbers tell a sequence: appearing, being chosen, and the position where this happens.

• Impressions: how many times one of your pages appeared among the results. It doesn't mean it was read, or even seen: it means it was on that results page.
• Clicks: how many times someone chose it. It's the only one of the four that corresponds to a person who decided something.
• CTR: the percentage of times that, having appeared, you were chosen. It depends mainly on position and on how well the title and description shown answer the query.
• Average position: the average of the positions where you appeared. It's the most misunderstood number, because it's an average across different searches: you can improve a lot on one important search while the average gets worse, and vice versa.

From this comes the most useful reading rule: average position should only be looked at for a single search or a single page, never as an overall indicator for the site. In aggregate, it doesn't describe any real situation.

Brand searches and service searches

This is the separation that changes the reading of the data more than any other. Whoever searches for your company's name already knows you: they saw you on a van, someone gave them your name, they're a customer looking for your phone number. Whoever searches for "industrial door repair" doesn't know you.

• Mixed together, the two categories produce an average that describes neither of them. Brand searches have high positions and high click-through rates, and they hide weakness on service searches.
• Filter out searches that contain your name, even misspelled, and look at the two groups separately. It's the first thing to do every time you open the dashboard.
• Growth made up only of brand searches isn't visibility earned: it's advertising, word of mouth, or name recognition working somewhere else. It's good news, but not on this channel.
• On service searches, impressions matter before clicks: if you don't appear, there's no title to improve.

A sample report with three observations

A report that's readable for someone who doesn't do this job fits on half a page: three observations, each with the check that goes with it. The examples are illustrative.

The third row is the one that saves the most time. When impressions are nearly absent, any work on titles and descriptions is wasted: first you need to understand whether the page is known and whether it says anything about that query.

Comparing periods without being misled

Comparing two periods is where the most wrong decisions get made, and the precautions are always the same.

1. Compare comparable periods: same number of days, same season. Many industries have dead months, and August against October measures the calendar.
2. Check whether the most recent period is complete. The last few days may still be partial, and the drop you see at the end of the chart often doesn't exist.
3. Be wary of small numbers. On a few dozen clicks, two more or fewer visibly change a percentage without anything actually having happened.
4. Before interpreting, ask what changed at the company and on the site: new pages, removed pages, a redesign, a campaign. Half of the swings are explained this way, and no data point will tell you that in your place.

From observation to check

A data point leads to a decision only if it already has a recipient. Typical observations fall into three groups, and each one goes to a different person.

• Content: searches that find no answer on your pages, questions that keep coming back, services you name but don't explain. These go to whoever writes.
• Technical: important pages with no impressions, duplicate addresses, pages that appear instead of the ones intended. These go to whoever maintains the site.
• Commercial: searches that reveal demand you're not serving, or customers looking for you with words you don't use. These go to whoever decides what you offer.

The right pace for most companies is monthly, with a broader review every three months. Looking at the data every day produces reactions to noise, not decisions.

What this guide doesn't cover

This guide covers performance on Google Search, up to the click. The link between this data and customers actually acquired — what can be attributed and what can't — is covered in a dedicated guide, because it requires other sources and a lot of care. Presence in the answers of AI-based assistants, which doesn't show up in this dashboard, has its own guide.