PPC used to be all about segmentation - if like me you were around ten years ago you’ll remember the halcyon days of even fairly small accounts having 100+ campaigns, each with 50 ad groups. For many of us, it's crazy to think of that now - the past is indeed a foreign country! So let’s look at some of the main changes in PPC for retail over the past few years.
Nowadays, the way in which we interact with digital media has changed, with infinitely more touchpoints at which we could engage with a brand. And in a cookieless world, there will be a lot more modelling involved (or “educated guesswork”) in tracking those engagements.
For PPC, we are now rejecting that “segmentation, segmentation, segmentation” approach and fully embracing automation - grouping together your keywords, ad groups, campaigns, and giving machine learning the data it needs to do the dirty work. Your agency is no less necessary, but what you demand of them will have evolved, and the way they need to structure your accounts will look very different.
One major change Google has made over the past few years is with keyword match types. Exact match nowadays operates in essentially the way that phrase match used to, with phrase match itself now closer to a version of broad match modifier, which Google discontinued in July 2021. For our generic campaigns in particular, we operate with a much smaller, core base of keywords in just exact match and broad match. With the right safeguards in place, this enables us to allow the AI to then maximise our performance potential.
More generally, with the introduction of cross-channel campaign types like performance max and demand gen, your marketing channels are less distinct from each other than maybe they used to be - for example Pmax can serve across all Google platforms, so is it lower-funnel? Mid-funnel?
Here at Genie Goals, we look at things through the lens of the classic marketing funnel of awareness, consideration, conversion. But the traditional PPC model of video - display - shopping - search, defining by channel, has evolved to Demand Gen - Performance Max (maybe Shopping) - Search - so more defined by campaign type. Embracing this, and mapping it to your account objectives, will get you a lot further than trying to translate one directly to the other.
There are also some previously commonly-used features that are not compatible with automated bid strategies - location adjustments, device-level adjustments, scheduling. We’ve seen many examples of agencies not understanding this and spending what must be an inordinate amount of time trying to apply adjustments to campaigns that cannot use them. Far more important these days is first-party data and refined audience management, now that Google has reduced access to third-party data with the tightening of data protection regulations that we’ve all had so much fun with recently.
All of these changes have reduced the time we spend on repetitive, labour-intensive PPC tasks, which is great news for everyone. But as AI takes a more prominent role alongside those changes in consumer behaviour, we are continually having to learn new ways to understand and influence the performance of our accounts.