How to Measure AI Advertising Campaigns

Pair AI With Incrementality

Now that machine learning tools are available to manage ad campaigns, the options of what to optimize for are even more important. If automated campaigns are not well understood by marketers, the wrong optimization strategy can highlight flaws in KPIs more intensely than if investments were allocated manually. For example, if a Google ad presents a pair of shoes to someone who was already searching to buy that exact item, it will have a high conversion rate and, ultimately, ROI. If that purchase would have happened organically regardless of the ad, that investment is not necessary. If a platform like PMAX is not managed with this in mind, it can overvalue an ad reporting high conversion and reallocate more ad dollars to scenarios like these without recognizing that it’s an inefficient investment. To invest ad dollars efficiently, it’s important to steer these algorithms towards more high-level investment philosophies like incrementality. Incrementality is focused on the truly causal impact marketing has on a desired action and is one of the main pillars of our unified measurement metric, Enterprise Marketing Return. So, how can you test for incrementality in an automated campaign?

Testing for Value

Measuring the effectiveness of a campaign is one of the most important aspects of AI adoption. That is a question that companies need to have answered when they consider the value of ASC and PMAX. Often it can be an efficient choice to incorporate within Meta and Google itself. For instance, mix an ASC Campaign with Meta’s DABA tool or mix PMAX with tactics bidding directly to Google Ads to reach an optimal allocation can be difficult.

We believe these two technologies will not become the only way to run digital marketing. Rather, mixing this platform with more hands-on management systems will cover some of the opportunity gaps of such a unified approach.

One of the ways we have done so is by using it not only to support the fine-tuning of our KPIs but also as a useful tool to rigorously understand campaign efficiency for a particular brand. That would mean removing media from a specific marketing channel in one business unit and comparing it to another audience segment as a control. That allows you to compare the effects of media spend and understand the true value of your marketing tactics. In-market tests can inform when it is best to leverage these platforms and how much to invest in them. That way, you optimally distribute that investment within your marketing channels.

Testing is one of the most useful foundations of our company’s analytical structure. We believe more marketing firms could benefit from testing machine-learning technologies like ASC and PMAX against control groups. By doing so, a marketer can clearly compare these tools with other offerings and arrange the best combination for their use.

So, What’s A Marketer To Do?

Even though AI campaigns are extremely powerful, they still need to be paired with a smart measurement and testing strategy. Implementing a measurement strategy can unearth insights that might not be found by leaving measurement up to AI.

It’s important for marketers to have a measurement team that understands the ever-changing landscape. Measurement is at the core of what we do at Ovative. We’re constantly thinking about measurement and how it can future proof your business for uncertain times. We don’t just set and forget our campaigns; we are constantly testing to ensure that media tactics are successful.

Reach out to learn how our measurement team can ensure you’re spending your budget more efficiently when running AI campaigns.