As AI takes over everything, should it change how we structure ad campaigns? Can the best practices of a few years ago now harm performance?
In the modern advertising world, we see ever-increasing use of machine learning (often casually called “AI”) to control ad campaigns. These learning models require good data to perform well. Each campaign you run is a collection point for a lot of data. The system then uses this data to make decisions about the ads that are shown. Generally speaking, the more data a campaign has to work with, the better it performs.
But in terms of campaign structure, AI has little to offer. Many advertisers and ad agencies lean toward splitting up ad campaigns based on their reporting requirements, or structuring campaigns around their product categories, or building complex campaign structures around advertising funnel concepts from a decade ago.
That can be fine if you’re a really large company with a lot of conversion data in each of those campaigns. But we frequently see scenarios where merchants have a lot of very small campaigns with a long tail of ad groups focused on product categories or sometimes even individual products that aren’t as popular. Structured this way, they just don’t get enough data to ever optimize well.
Splitting Ad Campaigns Impacts Profitability (for Better or Worse)
Depending on what bidding strategies you’re using, all the digital advertising platforms have recommendations on how big a campaign should be to even be considered as a separate campaign. But these recommendations are for the minimum size campaigns should be. To perform optimally, campaigns should be much bigger if possible.
There are all sorts of reasons that people split things up into different campaigns in their ad accounts – and most of them are bad reasons (like the reporting convenience or product categories mentioned above).
But there are good reasons to do so:
- If you need to set different ROAS targets on different campaigns.
- If you need to set different geotargets on different campaigns.
- If you need allocate a different budget to different campaigns.
Every ad platform has different options available at the campaign level versus the ad group or ad set level, so you may find that with a particular business goal you must split up some of your campaigns. But if you don’t actually have a real business requirement that forces you to split out campaigns, you might be better off keeping them combined into a single larger campaign with more data. Combining some smaller or compatible ad campaigns in strategic ways can actually improve the performance of your account and grow your overall profitability in the long run.
The Right Combinations Bring the Best Results
Typically, it’s a very good idea to group ad campaigns together by the profit margins of the products being advertised in the campaign. By tiering the products with different ROAS targets together in campaigns – irrespective of what type of product they are—you align the bidding strategy more effectively with the profit opportunity for that group of products.
If you can bid more aggressively on the highest margin products and tighten up the spend on the lower margin products, that’s the most efficient way to increase your overall profitability.
In Summary
AI can do a lot for your ad campaigns, but without a thoughtful campaign structure and bidding strategy, your ads will likely underperform.