Can we trust Google Smart Shopping?

Campaign Planning

Sorry, Google, this blog’s a pop at you. 

 

If you run ads via Google you may have noticed there have been two bugs recently. One on November 13th and the other on August 20th, both affecting shopping, and both probably costing retailers millions. 

So first, let’s discuss the most decent bug. 

On Saturday, November 13th, CPCs went through the roof, with costs increasing anywhere from 50% to 200%.

So, over the course of the bug, depending on your current spend – let’s say it averages at £100 per day – costs would have increased to £150 - £300-plus. But, let’s be honest, most accounts are a lot higher than that.  

Google posted a message on Twitter, letting us know they had an ‘unexpected error’. However, they haven’t told us what this error was, why it happened and the likelihood of it happening again. 

 

A Google spokesperson announced that they’ll be getting in touch with all accounts affected and offering a refund by December 3rd.  

However, we highly recommend that you calculate the money Google owes you (using previous data from Fridays/Saturdays to work out an average), and contact them with this figure. 

This might prevent them from potentially offering you a figure a lot less than you’re truly owed. 

 

With this happening twice in the past two months, can we really trust smart shopping? 

 

The honest answer: We really don’t have much choice. 

If we want to continue using Google shopping (we’re not likely to switch to Bing, are we?), it’s something we’re going to have to learn to live with. 

Google has made it very clear that they are increasing their machine learning and artificial intelligence capabilities and are seriously encouraging all users to opt towards machine learning when creating campaigns. But with glitches such as these two recent examples, the public trust in Google software might be decreasing. 

Along with bugs causing issues, there are several features that, if not very closely monitored, will end up costing you £££s where they don’t need spending. 

Although not shopping specific, auto-apply suggestions (on search campaigns) has shown us exactly how AI can get out of control if not closely monitored. 

The auto-apply suggestions feature may seem handy, you know because AI is so smart, but in reality, and we’ve seen this on our accounts, the feature could be adding expensive keywords to your ‘approved’ lists resulting in huge fees. 

For example, if you had a client that sold car mats, auto-apply could add in things like ‘yoga mats’ and ‘bath mats’, before moving onto things like ‘shower mats’, ‘shower tray’ etc which end up looking more like B&Q than the actual store. 

Note, this happened. 

One piece of advice we will leave you with is this: while Google is still making all of their upgrades to machine learning smart shopping, we highly recommend that you are extremely vigilant with your ads and check them daily. Missing another bug may cost you £££s.