Average Position. What Are We Losing?


In February this year, Google announced the sunsetting of the average position metric.

Cue animated debate and rantery among the PPC community.


Average position is one of the most useful metrics for one particular purpose: indicating how much scope there is ‘push a keyword harder’ by its raising bid.

‘Pushing harder’ here means raising its position and consequently a) its chances of generating impressions b) (almost always) its click through rate… both leading to an increase in traffic.

I often filter for keywords with low avg pos AND low CPA, to find ‘low hanging fruit’, where a bid hike will almost definitely boost results.

Historically, the closest alternative to this metric has been Impression Share (in what % of searches where the keyword was ‘eligible’ to produce an impression, did your ad actually receive an impression). Or better – Impression Share ‘lost to rank’ (how often did your keyword/s fail to generate an impression specifically because of your ad rank, as opposed to your budget…)

This is useful, but it’s a cruder measure than average position… not just because the definition of ‘eligible’ is less than 100% clear… but also because this only tells you how often an ad appeared (somewhere) – nothing about its position relative to the competition… or relative to its potential.

Last year, we were given a set of more granular Impression Share metrics:

Search lost top IS (rank) = how often did rank prevent your ad from showing above the organic results


Search lost abs top IS (rank) = how often did rank prevent your ad from showing in the absolute number one spot in the search results.

These improve our understanding of how rank is affecting your ad’s status – but they still fall short of the clarity given by average position.

Here’s an exchange I had with a fellow PPC pro shortly after the announcement.

The point about average position not giving a definitive insight into your absolute position on the page is worth remembering. However it doesn’t devalue avg pos in terms of its useful function: revealing unused potential.

The fact that average position only measures your placement against other ads (which may or may not appear above the organic listings) actually keeps it pure as a measurement of how much higher you could appear by raising your rank.

Whether or not ‘any’ ads appear above the organic results is outside of your control – average position tells you how much more you can gain within the area that you can control… i.e. your position vs competing ads.

It actually counts against the replacement metrics (‘top’ / ‘absolute top’ Lost Impression Share) that they take no account of the fact that not all auctions serve ads at the top of the page… This means they tell you more about how you’re doing vs Organic results than how you’re doing against the other ads. Your ad could be first among all ads, and still not be either ‘absolute top’ or even ‘top’…

The decreased clarity on unused potential may not be an entirely unintended consequence of this change… with Google’s continuous push towards automation and control of advertisers’ activity showing no sign of slowing down.

All this to say…. Average position will be missed.

A few weeks ago – I had grand plans. I set out to cobble together a proxy for average position which could live on after avg pos itself dies.

The plan was to find a reliable enough relationship between average position and other existing metrics, for creating a calculated metric (custom column) for a reasonable ‘estimated average position’.

I exported the values of a group of ‘related metrics’ (Search Lost top IS (rank); Search Lost abs top IS (rank), and search Lost IS (rank)) –  along with average position, – for over 2937 keywords from dozens of campaigns across 4 accounts, to see if I could find a reasonably close and reliable predictor of average position using any one or a combination of these metrics…

(I took a bunch of precautions to reduce skewed or misleading conclusions which I won’t go into now because….) It didn’t work out.

Here’s a sample of the results:

Metric Correlation to Average Position (Pearson’s R) Correlation Strength
A. Search Lost top IS (rank) 0.35 Moderate
B. Search Lost abs top IS (rank) 0.52 Moderate to Strong
C. Search Lost IS (rank) 0.3 Weak to Moderate
A x B 0.39 Moderate
A x C 0.37 Moderate
B x C 0.32 Weak to Moderate


Nothing came out strong enough to use as even a rough indicator of average position… though one metric did stand out as being by far the closest: Search Lost abs top IS (rank), with a correlation of over 0.5.


On this basis, I now prioritise Search Lost abs top IS (rank) as the most worthy heir to avg position once the king is dead…

Being ‘absolute top’ of the page is not often a worthwhile aim in itself… but how often it happens may be the best indicator we have of that more useful piece of information: how much unused potential is available, and attainable by raising bids.

For now I recommend enabling all the lost-to-rank columns (along with the enigmatic ‘Click Share’) to look for ‘corroborating evidence’ where multiple metrics show low results… and note that Search Lost IS (rank) has special significance, because how often your ad is showing ‘at all’ is a particularly pressing question…

My PPC colleague was right that we’re losing the convenience of a single metric. It’s unfortunate that we’re also losing some data evidence that this metric holds and will take with it to the grave… data that we may not find anywhere else, and that notches up the advantage Smart Bidding algorithms hold over us human PPCers…

On which note, if you’re not already, it’s time to get friendly with smart bidding. More on that soon.

How to Get More Useful Info…


Join my free Facebook group


Check out my PPC resources


Book a one-to-one call

Share this post

Lastest Posts

Google Ads Updates - Q1 2024 Roundup

No slowdown in the pace of change with Google Ads updates in Q1 2024… Here’s…

Aggregation vs Articulation: Competing Virtues of PPC Management

Three ‘A’s Two concepts have been central to the recent evolution of Google Ads: aggregation…