Smart shopping learning period

Hi all -

Long story short: I was contact by an adwords rep with some advice to set up a smart shopping campaign as it would out-perform our existing shopping campaign. They helped me set it up over the phone, and i kept my existing budget of the original campaign for the new one, as they said they couldn't run at the same time. What they didn't inform me about was the learning period, saying it would take 15 days for it to learn and in that time performance wouldn't be great. So I complained that i was losing lots of money, and they advised me to lower the budget. However this then affected the campaign, and i called back and was informed changing the budget caused it to 'relearn' again as the budget was changed.

I can't find any documentation on this, and have spoken to several google agents now who have all told me contrasting things. Put simply, does a smart shopping campaign always have to 'relearn' every time a change is made, and we just have to put up with the negative impact and money loss during this period? To me this seems a terrible flaw which only benefits Google.

 

Hi Dylan,

Let me start by saying that Smart Shopping definitely CAN run alongside Standard Shopping Campaigns. In fact, I don't recommend switching to Smart Shopping altogether, but rather using it to target under-performing categories/product types from your Standard Shopping Campaigns while keeping your Standard Shopping Campaigns on (and excluding these product types/categories from Standard Shopping to avoid traffic cannibalization).

Here is the last 30 days' worth of data from one of our clients to support this:



A couple of reasons against full reliance on Smart Shopping:

1. There's no search term optimization. As you used to manage standard Google Shopping campaigns, you know just how irrelevant some of the traffic can get. With Smart Shopping, there's no optimizing that.
2. There's no bid optimization for the best return on investment on specific categories/product types (Example: Product Type B is below your target ROAS, but has plenty of searches and eats up the bulk of your budget at a high CPC? Too bad. Your only option is to separate it into its own Smart campaign with its own budget and hope it has enough data for the algorithm to learn).
3. There are no insights for feed optimization. Feed optimization(and specifically title optimization) works, but you need to know which keywords/search terms to optimize for.

As for relearning periods for Smart Shopping, I haven't noticed any drastic changes following budget changes across any of the clients, to be honest, but I wouldn't be surprised if that's the case and it restarts the learning period. I think the most important thing during the learning period is to provide the algorithm with enough conversion data. I can't find it in the official Google Shopping support articles, but it used to say that Smart Shopping requires at least 20 conversions over the past 45 days across existing Shopping campaigns, meaning that its performance could suffer if there wasn't enough conversion data.

Hope this helps. Happy to explain any of this further if you have questions.

 

Google is trying to push all advertisers to smart campaigns. They claim this is best for you. I prefer keeping control but that's me and I know what I'm doing (I manage client accounts after all). I understand it's not for everyone and most advertisers would rather set-and-forget and let the system do it's thing but that means putting your advertising and potentially your business into the hands of an algorithm.

As for your specific question, I don't think changing the budget would cause it to relearn. Most other changes would however, primarily a change of ads. You should test different ads with the understanding that you are testing and it may not turn out the way you want. Frankly, this relearning period should not factor into your decision of testing.

Unfortunately, as with getting information on forums, you'll get differing recommendations from Google reps that are contradictory. What I can say is that Art above is correct as far as I know (he seems to have more experience with Shopping campaigns than I do).