There have been a lot of changes to Google Ads in the last few weeks.
As the dust starts to settle, here’s a recap on what they are, and what they mean.
1) ETAs are on the way out
As of the end of June next year, no more new Expanded Text Ads! Responsive Search Ads will be our only option within standard search campaigns.
This is yet another step in reducing advertiser control and although this move has been looking likely for a while – it is one of the biggest changes we’ve had in that direction.
One of the reasons that PPCers are unhappy about this (which they are…) is that RSAs often underperform ETAs in practice (not always, but often enough for ETAs to play an important role in all of our accounts).
A few silver linings:
1 – RSAs have huge potential, and should improve
2 – We can still pin assets -and if you choose to pin H1, H2, H3, D1 and D2… you’ve effectively made an ETA… (ignore the low Ad Strength that Google will assign when you do this)
3 – Data and reporting around RSAs is – currently – not great. Ad strength is a fairly meaningless metric – not tied to anything of importance – while the visibility into RSA asset performance is non-existent – making even basic message testing absurdly inconvenient…
That will surely (surely??) have to change, with RSAs becoming the only text ad type available.
2) A change to match typing
When you have the same keyword on multiple match types – The tighter match type will always be preferred when eligible.
Phase one of this move was revealed in a very low key way, back in February… That exact match instance of a keyword would always be favoured when its matching search term comes up, regardless of ad rank.
But on 23rs Sept it was further announced that a similar system in place favouring Phrase match over Broad.
One important upshot is that removes the key benefit of structuring my match type – as when it comes to making sure that the right search term goes to the most precisely matching instance of a keyword.. that is taken care of without internal negatives or bid manipulation.
3) Search term visibility
Search term data seems to have fluctuated over the past few months – since the big cull back in September 2020.
Note that a lot of the new search terms appearing are for zero-click search terms (not completely value-less – but certainly of more limited interest than the clicks..) and in many cases a high proportion of clicks are still coming from hidden search terms.
4) Defaulting to Data-Driven attribution
Data driven attribution is soon to become the default model for new conversion actions.
While many of Google’s recent changes have been unwelcome – in this case, we all have more to gain than lose. Last Click is rarely the best option, and if you do have a good reason to use it – you’ll still be able to switch back.
The really major change here is that DDA will no longer have its current, high data threshold, being available from the start.
It may not work quite as well without any (internal) pattern history behind it… but it won’t do any worse a job than Last Click.