A click estimate is not a forecast

Every keyword on a Rankgap worklist carries a value, and the value is arithmetic you can check: the clicks a better position would be expected to win, multiplied by what an advertiser pays for that click. Both of those inputs come from somewhere, and neither of them is a measurement of your site.

Where the inputs come from

The search volume and the cost per click are reported by a third-party keyword provider. The click side is a published average click-through curve — what a position tends to earn across the web, not what it earns on your pages, in your market, for your brand.

That makes the number useful for exactly one job: comparing two keywords. If one is worth several times another, the ranking is telling you something real about which to write first. If two are within a few percent of each other, the difference is inside the error of the model and you should pick on something else — how well you can actually write the page, or whether you already have one.

What it is not

• It is not a prediction of your traffic. Your own click-through rate is a property of your title, your brand and the shape of the result page.
• It is not a prediction of your revenue. What an advertiser pays for a click is not what a visit is worth to you.
• It is not a promise that ranking is achievable. The difficulty band beside it is the closest thing to that question, and it is a different measurement.

Once you connect Search Console, the curve is replaced by one fitted to your own impressions and clicks, which narrows the first of those three gaps and leaves the other two exactly where they were. That is the honest version of what this kind of number can do, and it is why the footer of every page here says the same thing.