Pmax for shopping campaigns is now the default on Google ads for retailers with over 84% of shopping advertisers using Pmax as the default.
Around 42% of all Google ads spend is sitting in Pmax, so it’s fair to say it’s the go-to campaign type for ecom.
Google positioned Pmax as an all in one campaign type which leveraged AI learning. It certainly delivered in the all in one category. Pmax increased Google’s YouTube revenue by over X billion in 12 months alone.
Let’s get real, that wasn’t advertisers suddenly opting into YouTube. They did so unwittingly.
Many sources still cite that Pmax spends most of its budget in the PLAs (shopping ads) but this is horribly untrue. We have uncovered instances of Pmax spending <10% of ad spend in PLAs and the advertiser is sat wondering why performance is tanking.
In a parallel situation, we’ve also seen Google eat the majority of the budget on ‘undisclosed’ ad inventory, only to spend the rest on brand search – thus giving the impression that Pmax is delivering a ROAS in some capacity. Meanwhile, the advertiser is again, left scratching their heads at why sales aren’t increasing.
Pmax also prioritises easy conversions. It’s so tediously predictable to audit any Pmax and literally, physically, count how many products have received >10 clicks – even on a sizable sku inventory. The same products, round and round and round.
Inventory stagnation, stock holding costs rising.
Maybe you’re getting the impression that I’m not such a big fan?
I am a fan. I use Pmax day in and day out. What I don’t do is blindly trust that it’s delivering results.
Prior to the helpful introduction of AI assets (don’t get me started) Pmax could be efficiently policed by running Pmax without assets. However, thanks to the introduction of unauthorised AI assets, Google will simply solve for your lack of assets by providing their own, off brand asset pack.
You are likely to have AI text, images and videos, all loitering in your ads account, serving on dodgy bargain basement ad inventory and some 4 year old’s Peppa pig video.
My point? Pmax can be great, but it can be bad in equal measure. What you cannot do is trust that it will perform and continue to perform at the same pace.
You cannot trust that Pmax isn’t delivering a solid ROAS, while underhandedly serving up a good proportion of your budget on branded search to inflate the stats.
So, yes, do Pmax, but police it rigorously and also test manual alongside.
Good old manual shopping with a fully optimised feed will do exactly what you’ve asked it to do and you perhaps won’t have an overinflated ROAS based on brand search – but do you want that anyway?
Manual testing for search and shopping allows advertisers to make sure they are represented in the PLA auctions that matter. You can flex between manual CPC strategies and investments to make sure you have a balanced outcome – targeting lower priced inventory and also bidding for higher inventory in others. You don’t have to compete against yourself with the right structure.
If you’re running Pmax or ROAS based search, you need to move the doors to manual and maybe, just maybe some sales will fall out of the overhead locker.
