PPC Search average position is a weighted average and must take into account the number of impressions
For example, if you have raw data that shows the average position that an ad appeared at across two sites within a campaign the temptation is to add the two positions up and divide by two to give the overall campaign average – and this would be wrong
Intuitively the above seems fine. But the ad was seen in position 1 twice as many times as it was seen in position 7, so clearly the average should not be half way between the two.
Instead, you must calculate a new field by multiplying the number of impressions by the average position in order to include the underlying weighting (think of this new metric as the sum of all of the positions, if that helps).
Now we need to find the sum of ‘position x impressions’ and the total number of impressions
Finally, the average position across the whole campaign can be calculated by dividing the total ‘position x impressions’ by the total impressions
(i.e. 450,000 / 150,000 in the example) which is….3.0!
This answer is closer to 1 than 7, which makes sense when we know there were twice as many impressions in position 1.
The same principle would apply if you were looking at average position across months, campaigns or keywords, as well as across different sites
So remember – don’t add up and divide to get overall average position – take the number of impressions into account…