What is Pmax and should my brand be bidding on it?

'Pmax is a black box so we’ve made a conscious decision to not migrate over from standard shopping’. This is what we constantly hear from a lot of retailers out there.

Now, don’t get me wrong, I fully understand the debate; AI on one side and some form of control on the other. Marketers love to be in control, especially those that have been around for a decade and more. This idea of budgeting for each part of the marketing funnel and having the control over what search term we appear for, where and when but most importantly on which inventory. Pmax disrupted the industry, not to mention that it very quickly replaced smart shopping, leaving marketers with very little choice as to what to do next.

When Pmax was first introduced, it was pitched to all of us as an AI-powered campaign with the ability to drive better customer acquisition, better efficiency, and better growth for retailers. Not to forget that it would encapsulate most of Google’s inventory; Shopping, Search, YouTube, Gmail, Discovery, and Maps. And that was the scary part.

As an agency, we initially A/B tested Pmax versus smart and standard shopping and very quickly realised that the vast majority of our merchants traffic was still going towards the Shopping inventory. Google won’t tell you that but many open Google Ads scripts have been widely shared within the industry. No need for coding experience, it can be implemented in a few clicks and it really gives you an idea of the sort of inventory you’re appearing for. Lack of visibility problem number 1 solved.

Next, what about brand bidding? As part of their strategy some advertisers might not be bidding on their brand core terms on Search. But due to the nature of Pmax, they might no longer be in control and have no choice but to invest into their core terms. In this instance, our strategy is to purely add those branded terms as negative keywords to completely get rid of branded search activity on Pmax. And in order to capture those ready to buy customers, we would launch standard shopping. The 2 main benefits here are to avoid inflating CPCs on your branded terms and to ensure Pmax would still drive incrementality rather than perpetually going after returning users.

Aside from inventory monitoring and brand versus non-brand segmentation, Pmax works in a similar way to smart shopping and can only work with the data that you feed the tool with. This means that we would still recommend that you mirror your product categories and reflect that in the number of Pmax campaigns you want to launch. If you are a Fashion retailer think of what’s your bread and butter category and whether it warrants its own Pmax campaign. Everything else might fall under one other campaign. In any case, ensure that you have as many asset groups as possible, again mirroring your product categories.

Of course, as with all feed-driven activity, the quality of the feed is important. Great feed optimisation allows you to segment the campaigns in the most impactful way, and maximises the amount of pertinent information you are feeding into the Google AI.

Let’s take a step back and be honest for a second. The reality is that Google is no longer putting hours into developing standard shopping so this campaign type will eventually be long gone. Not being agile, testing new Betas and really using automation to your advantage will undoubtedly become a future risk for your business. So please do A/B test both campaigns before you’re left with no choice but to migrate.

So, what are some ecommerce brands doing?

I talked about campaign segmentation earlier on. Similar to smart shopping or standard shopping, the best use of Pmax is still by segmenting campaigns based on your business objectives. These can be based on profitability with margin-based segmentation, pushing for bestsellers or giving priority to sale items. With many of our retailers, our campaigns are segmented based on all the above. There simply isn’t a one-strategy fits all but in every case our Pmax segmentation matches with our partners business objectives.

So, no Pmax isn’t created in 2 clicks, it still requires manual intervention and it still requires a strategy that matches your business objectives. Lack of control is undeniable but machine learning is ultimately here to support your growth so I dare you to A/B test Pmax now. And remember, ‘you’re not competing against AI, you’re competing against retailers that are using AI’.