Search Top Impression Share Exposed!

When Google retired the average page position metric in their search ads, we had a few initial concerns. Would the new metric, search top impression share (IS), poised to replace it tell the same story? And how would we create an internal benchmark for these metrics, to gauge performance and visibility?

At DSA, our internal benchmark for impression share is 70%. FYI, this metric is different from search top IS, even though they sound similar. While this isn’t always achievable, it’s a strong goal to continuously optimize towards. For those who don’t know, impression share is how often your ads show in the auction versus how often they potentially could’ve shown. It’s a strong first indicator of how visible your ads are in a competitive marketplace.

For average page position, the benchmark was a bit more broad. Google stated that up to 4 ads could be shown at the top for desktop, and 1-2 for mobile. We would strive to get a page position around 2 where we could, to try and show above the organic results as often as possible.

So where does that leave search top IS? This metric is classified as how often your ads showed at the top of the page above organic results versus how often it potentially could’ve shown. When first considering this, I assumed we’d see a strong correlation between search top IS and impression share, but all the examples researched showed quite a variation. This frustrated our team when trying to find a benchmark for performance of search top IS. If one reports how often your ads showed in the auction, and one reports how often they showed at the top, why do they not match or at least fluctuate the same?

This lead me to visit the one other metric still in the platform, that also reports on ads showing at the top. Within the auction insights report, there is a metric called “top of the page rate” (TPR). This metric reports how often your ad shows at the top of the page. Simple, right? After reviewing these metrics we assumed there must be a similar factor that connects these three. Here are five examples after reviewing five different campaigns from five different industries:

Client by Industry Impression Share Search Top IS TPR*
Food Services 33.70% 26.67% 78.81%
Telecomm 50.59% 40.62% 78.45%
Banking 75.95% 66.02% 86.60%
Real Estate 40.83% 30.31% 70.89%
Government 47.73% 38.69% 79.48%


*Please note that the TPR metric will be slightly off from the other two. The TPR is pulled from the auction insights report, which is accumulated from between 70-80% of the available impressions, so not the campaign total, while the impression share and search top IS are campaign total numbers. All these numbers were calculated over the past six weeks of data.

After reviewing these three metrics for the five different campaigns, it’s very easy to spot a pattern. The search top IS is never above the impression share. This makes sense because if impression share is how often your ads showed in the auction, and search top IS is how often they showed at the top, one would have to supersede the other. So how does Google come to this conclusion? Here’s our theory:

If we take the TPR number and divide it into the impression share, we should in theory get a number very close to search top IS, which is exactly what happens. So how do we find a benchmark for search top IS? Looking at the average TPR among five different industries, we get a mean average of 78.85%. The five campaign examples are all doing well for performance in terms of CTR % and CPC. Rounding this up to 80% for easier calculations, we can safely say for now (we recommend reviewing this again after 12 months) that this is a good benchmark for TPR. Ultimately, we could say that if we view impression share as a whole number, and divide that by 80%, this is a strong benchmark for the search top IS.

Do you have a different opinion, a different calculation or a general question? We’d love to hear it! Please reach out to me at