Google has announced that the Average Position (in some quarters referenced as “Average Rank”) metric, already relegated to a “Competitive Metric” value from its prime “Performance” metric status, will be killed off from September this year.
Is this a good or bad thing?
There were supporters of it but, like many averages, it was nearly impossible to pin it down in such a way that it was genuinely actionable. Therefore, it is being substituted by “Impr. (Abs. Top) %” and “Impr. (Top) %.” - these are the share of your impressions that were above the Organic results or on top of all other ads.
In addition, you can see “Search top IS” and “Search abs. top IS” as well as “Click share” for a sense of how competitive your bids are.
The beauty of these newer metrics is that they not only give you a far better sense of share of voice/visibility but also how much you can expect in terms of impression growth if you improved “rank” on search. Click Share is being provided on Search for the first time (previously it was only on Shopping campaigns).
Average position would give you some sense of this but when bidding on Non-branded traffic, you were generally less concerned with position than with profitability and so you would bid down if parts of the account were not profitable and bid up if they were beating your efficiency targets.
What was wrong with Average Position?
The main problem with an average figure like Average Position (but also CPC and CTR) is that it illuminates very little about the true performance of an account or segment of an account. This is partly because at any given moment for any given search, you may be positioned in 3 or 6 or 8 and yet the average position for the day may be 2.1.
To illustrate a further weakness, if two keywords in an ad group had the following numbers:
You will see that the good performance of one keyword is masked by the other. This is because Average Position is weighted by impressions. So the more impressions a keyword (or other entity) has the more they impact the average. For the record, the equation is (Entity Imps x Position) ÷ Total Imps.
But worse still is that the below scenario throws up almost the exact same Average Position top number as above but the average position for both keywords individually is worse than that shown in the first table.
Therefore, an absolute percentage share of the total potential as a metric (e.g. Top of Page Impression Share) is far more preferable because it has equal fidelity and transparency at any level against which it is shown.
Effects and Possibilities
The addition of automated bidding strategies that let you target a percentage share of “Absolute Top Position IS” and the like is a welcome opportunity for advertisers to control their share of voice (better than targeting a “position”) whilst it continues to keep the Google auction still pretty blind.
In addition, the new share-based metrics will make for fewer metric-related disruptions should Google choose to radically change the layout of search pages and where they place the ads.
As long as you run a report with the Impressions metric, you will be able to aggregate these numbers quite easily and reveal how many impressions (including Top and Abs Top) you’re missing out on using any given dimension/segmentation (such as Day, Device, Keyword etc.). Here’s how:
- Create new columns in a report and populated them for each line with below equations/formulas:
- Impressions x Impr. (Top) % = Total Top Impressions
- Impressions x Impr. (Abs. Top) % = Total Absolute Top Impressions
- Total Top Impressions ÷ Search Top IS - Total Top Impressions = Lost Top Impressions OR Total Top Impressions x ( (1 ÷ Search top IS) - 1) = Lost Top Impressions
- Same as above but using “Total Absolute Top Impressions” in place of “Total Top Impressions”
In this table, the Grey columns are metrics provided by Google and the others are established through the above calculations:
It’s not yet known if Bing Ads will also sunset the average position metric or whether they will wait and see if it is or isn’t missed by Google advertisers when it dies. If history is any guide, however, it’s highly likely they will follow suit.
There will be those who do mourn the passing of Average Position in September but for many, it will be forgotten as they get to grips with leveraging the more useful new metrics now available.
It has been commented within the industry that it makes PPC more of a programmatic buy, like a DSP. To some extent that is correct - and this seems to broadly be the direction PPC is heading as a logical conclusion sometime in the not-so-distant future.