More commerce data providers and retail media networks are coming on the scene every day, and the category is now considered the data gold standard for good reason. Marketers have discovered commerce data solves many challenges in today’s ad tech landscape, but the quality, coverage, and transparency of the data sources can vary greatly by partner.

When talking to any commerce data vendor, marketers shouldn’t hold back on the hard-hitting questions to validate the source, seed methodology and longevity of the data they sell. Here are 5 questions you can commit to memory and help the cream rise to the top.

1. Where and how is the data sourced?

Just because data is from a commercial source doesn’t automatically imply that it’s sourced fairly or explicitly. There are many companies that aggregate data from disparate sources with differing coverage and quality, repackaging it as their own. There are also many third-party data brokers selling so-called “past purchase” data; Data in which the source and its quality is completely opaque, not to mention reliant on cookies. Instead, look for partners who sell their own first-party data with explicit permission from consumers to do so. If they can’t tell you where their data is sourced, that’s a red flag.

2. Is the seed audience deterministic?

Would you rather make a decision using someone else’s best guess, or a verified fact? It’s a simple question to answer, yet many seed audiences use some form of probabilistic modeling to grow the seed audience, which dilutes the very definition of what a seed should be. Though a probabilistic data set might give you a larger seed audience, it will only magnify any inaccuracies when it comes time to lookalike model. Smaller seed audiences that are using extremely high-quality data, coupled with a solid modeling methodology, will be more performant and efficient compared to a massive audience relying on probability. As the saying goes, garbage in, garbage out.

3. Is the identification method deterministic?

Prioritize data providers who are able to use first-party PII as a digital identifier, such as permissioned email addresses and phone numbers. In these instances, PII is fed into a device graph and matched with the larger programmatic universe. PII is a stable source of identification with high match rates. And of course, this data needs to be sourced with appropriate consumer consent. Providers who opt for probabilistic identifiers rely on a patchwork of ephemeral signals, such as IP address and operating system, in an attempt to identify a consumer. That cobbled-together identifier has to be matched to a programmatic device graph, inviting another element of uncertainty into your strategy: will you even be able to find the consumer you have worked so hard to define?

4. What’s the scope of the data variables offered?

The data industry tends to be siloed by verticals, especially when it comes to commerce data. Generally speaking, retail media networks have limited insight into their customers beyond what they’ve purchased from the retailer. Working with providers that have access to enriched data profiles, beyond the core data set, such as demographics, psychographics, and purchasing behavior, can open the door to more precise targeting strategies and less waste.

5. How do they prove the value of their data?

Commerce data providers are sitting on the holy grail of closed-loop measurement. The data used to target can, in turn, be used to measure campaign success with actual sales.  Look for providers that are willing to use their data to also measure and prove the efficacy of their audience segments through sales attribution or incrementality.

When assessing any commerce data provider or RMN, the key is to remember that quality trumps quantity, and transparency is non-negotiable. Ensure you partner with providers who prioritize deterministic methods and can demonstrate the tangible value of their data. At the end of the day, your data is only as powerful as its source, so make your choices wisely.