I recently returned from Path to Purchase Institute’s Future Forward conference in New Orleans, and the #1 topic of conversation (and #2 and 3 as well) was all things retail media network (RMN): advancements in targeting, partnerships, and new capabilities. Above all else, though, people wanted to talk about measurement, ROI, and the optimization of RMN advertising.

For the RMN wranglers out there, measurement is a tricky subject: every network has different metrics, different definitions of their metrics (share of voice, anyone?), diverse ways of accessing data, differing levels of support, and so on. At the same time, through joint business planning exercises, RMNs are requiring a greater and greater portion of marketing budgets be spent with retailers.

So, what’s a marketer to do? I would suggest these five guidelines when considering how to evaluate retail media performance:

1.    Focus on incrementality: One of the biggest criticisms of retail media is that it is claiming credit for sales that would have occurred anyway. But you don’t have to take the retailer’s word for it; you can measure incrementality yourself. Choose an approach such as test and control or regression analysis that can be applied consistently across RMNs for cross-network comparisons.

2.    Standardize across networks: The IAB has enacted a standardization committee—of which I am a part—to tackle metric uniformity across the industry. It could take years for the major networks to adopt these standards, though, if they ever do. In the meantime, you are on your own. 

But don’t fear! There are things you can do. First, agree on a few key KPIs you really care about and that you can get the raw ingredients for from most/all your networks. I would suggest impressions and incremental Return on Ad Spend (iROAS). Next, focus on standardizing what you can, such as consistent lookback windows. It may not be perfect, but chances are it is better than what the RMNs are currently giving you.

3.    Create contextually appropriate comparisons: Everyone always wants benchmarks: How am I performing vs my competitors/the category/other business units etc? However, when it comes to optimizing retail media performance, other comparisons are more appropriate, and competitor/category benchmarks may not be available anyway. 

There are two types of comparisons that will be most useful to you in optimizing RMNs: performance between RMNs, and performance within an RMN. For the former, metric standardization is critical, because you have to be sure you’re comparing apples-to-apples. For the latter, note that many RMNs let you download historical data. Creating self-referential benchmarks within an RMN (e.g., How has this brand performed on this tactic in the past?) is an effective way of evaluating performance.

4.    Tap into AI: You didn’t think I was going to skip this, did you? The beauty of AI is that it can process enormous amounts of data and identify insights that no human could. For example, we use Google’s Cloud Vision to help us identify trends in top performing creative linked to the actual conversion. We also leverage 3M’s VAS software to predict where consumer’s attention is likely to be drawn. 

5.    Push back: Not a measurement tip per se, but perhaps the most important. I was talking to the head of one RMN and their advice was that advertisers have more power than they think. According to them, “advertisers don’t just have to accept what RMNs tell them. If they need additional data or want to negotiate better terms, they can. After all, they pay the bills.” Don’t be afraid to use your leverage.

To mix metaphors, RMN measurement is the current Wild West of marketing, and retailers are all trying to get in on the gold rush. Fundamentally, however, there is nothing different about measuring retail media than any other marketing: focus on the details, standardize, be consistent, and you’ll be fine.