What is a good top impression share? The bands we work to, and the four columns that say why you miss them

These are Riibon's working thresholds, set in our Search diagnostic framework, not a Google figure and not a benchmark measured across client accounts. On Search, we treat a top impression share of 50 per cent or better as healthy and under 30 per cent as a flag. On searches for the brand's own name, we also hold the absolute top impression share to 60 per cent or better, and treat under 40 per cent as a flag.

They are starting points. Our practice is to replace them with the account's own history as soon as there is enough of it, because the right band depends on the market and on how many advertisers bid on the same searches.

Google defines Search top impression share as the impressions you've received among top ads divided by the estimated number of impressions you were eligible to receive among top ads. Search top impression rate, labelled Impr. (Top) %, is a different column: the percentage of your ad impressions that are shown among top ads. Our bands apply to the share.

The difference matters most when you change a bid. Google does not recommend using the rate columns as a target to set your bids, because sometimes they may decrease as bids increase: higher bids may allow you to enter more competitive auctions in a worse location. It lists the share columns as the ones you can use as a target if you want more of your ads to show among top ads or in the first position among top ads.

This is our reasoning, not Google's. Someone searching for the brand's own name has already chosen it, so a place below another advertiser on that search risks handing that searcher to someone else. On non-brand searches the advertiser is one option among several, and paying to be first on every one of them is rarely the cheapest way to win the sale. So brand searches carry an extra band on the absolute top, the very first ad, on top of the general band for the top of page as a whole.

Google reports the shortfall in four columns, each saying how often your ad missed a position and why. Search lost top impression share (rank) estimates how often your ad didn't show anywhere among top ads due to poor Ad Rank. Search lost top impression share (budget) helps you to understand how often your ad didn't show anywhere among top ads due to a low budget. The two absolute top columns estimate how often your ad wasn't the very first ad among top ads, due to poor Ad Rank and due to a low budget respectively.

Our reading of the split: a rank loss and a budget loss point at different constraints, so check which column carries the loss before you touch a bid or a budget. On rank, Google says it considers ad quality in establishing your Ad Rank, and that the amount you bid may affect your Ad Rank but doesn't impact the assessment of your ad quality. A rank loss is therefore not only a bid problem.

Google's general tips that can help improve impression share include increasing your campaign budget, increasing your bid and, for Search Network campaigns, improving your ad quality. They are a general list for impression share. Google does not attach a tip to each loss column, and we read the columns as naming the constraint, not prescribing the fix.

Google divides both top impression share and absolute top impression share by the same estimated number of impressions you were eligible to receive among top ads. Our arithmetic from those definitions: for the same rows and dates, the absolute top share cannot be higher than the top share, and the gap between them is the share of top of page opportunity where you showed but not first. The gap in points cannot be bigger than the top share itself, so on its own it says little when the top share is low. The better read is the ratio: absolute top share divided by top share is how often you were the first ad when you were among top ads at all. We read a low ratio as winning a place among top ads without winning the first slot.

Search partners don't distinguish between 'top' and 'others', Google notes, and won't be included in location segmentation or impression share data. Our reading: a campaign that also runs on search partners is being judged here on its Google Search traffic only.

And do not average the share across campaigns or days. Riibon's metric layer marks both top shares as ratios that are never summed or averaged across rows, because a plain average weights a row with a few eligible impressions the same as one with thousands. How we roll impression share up across days is its own post.