The CPG holding company Kellanova, which owns Pop-Tarts, Pringles, Eggo and many other well-known grocery brands, is perusing data suppliers in pursuit of purchase data to snack on. The person who’s in charge of taste-testing those retail data suppliers is Paul Loukes, Kellanova’s senior director of data-driven marketing.
“It’s a job I convinced the CMO that she should create,” he told AdExchanger.
Naturally, some of Kellanova’s newest potential data partners involve generative AI, he said. But there is also a growing number of data suppliers with different approaches to measuring retail store sales, such as Becausal (formerly Scanbuy), which collects data from a consumer shopping app and other retail loyalty programs.
AdExchanger caught up with Loukes to discuss what Kellanova is looking for in a purchase-based data supplier and how that data is put to use.
AdExchanger: For your role, are programmatic log files or the Chrome Privacy Sandbox brouhaha on your radar? Do they even ring a bell?
PAUL LOUKES: That is part of my role.
Most brand marketers are more focused on their day-to-day brand strategy and briefings to ad agencies, things along those lines. They typically don’t have the time to be obsessed about person-level log-file data and things like that.
I straddle that realm with the advanced analytics and measurement team. So I have a peer that I work very closely with to test all kinds of different measurement techniques and explore avenues to try to match these purchase signals to that log-level data.
When did you start working with Becausal?
A few months ago. I was interested because it seemed like they have data sources that were a bit different from the ones we had been using.
How so?
To start I’d say our mantra with data-driven marketing revolves around the concept of full-funnel marketing. For us it starts with finding people who are at various states of relationships with our brand. We’re trying to place people within either a Retain, Reclaim or Recruit bucket, depending on what their purchase behavior is and history with our brands.
For example, with our Recruit bucket [Note: Recruit is Kellanova’s term for conquest marketing] we’re always trying to find ways to improve our odds. If you just target people who don’t buy your product, you’re wasting a ton on people who never will or perhaps are rejectors of your brand for whatever reason.
So we use data from a source like Becausal to try to find people based particularly on what they last bought in the category or shoppers who commonly switch among brands. Those are ways we used purchase-based data to improve our odds of conquesting, rather than randomly throwing out ads to people that aren’t buying our stuff today.
Is Becausal part of a roster or category of data providers you work with?
We would call Becausal a purchase-based or PBT [purchase-based targeting] data provider. Obviously, there are other industry partners out there. But they all have different types of purchase data and different sources of the data. Sometimes it’s a traditional in-store grocery data seller, or they use shopper cards or app purchases, or combinations of consumer panels.
We’re constantly evaluating different sources and looking for opportunities to identify new subsegments of people that are buying our products, rather than just one lump sum. We can get really fine, rich insights on those insights we’re getting from Becausal for instance.
What about the data is particularly granular or allows fine controls?
Some of the retail purchase data suppliers are just a mechanical stream of data. Like it might just tell you, “These are people who bought X product.”
What Becausal brings to the table is we can get a more balanced set of signals. It’s not just one data point. They can add value on top of just the baseline signal. Which means we can add the data to our consumer profiles and use it to create lookalike audiences.
Are there particular media channels or ad formats where you’re applying that data?
In terms of media, it’s the usual suspects.
CTV continues to be an important area that we’re testing our way into as linear becomes less and less part of our mix. And the kinds of audiences and data we’re creating with Becausal becomes really important in a channel like CTV because of the higher CPM costs involved. Though we’re not testing Becausal with CTV at this time.
For instance, let’s say we have an audience and we don’t want to spend money in CTV on people that just purchased the product. Those kinds of signals allow us to optimize, so we’re not spending CPMs on people that aren’t in the purchase cycle.
I don’t think of Kellanova brands as having a clear first-party data pipeline, since I imagine it’s almost entirely sold in stores or online retailers and the retailers get the data rather than the brand.
It is true that most CPG companies don’t have first-party data or are really building it from nothing. I feel very fortunate, actually, with Kellanova. If you think back to the days of cereal and mailing in box tops for prizes, we’ve actually had a very long relationship with consumers that’s not at the point of purchase, per se. It actually goes back to the history of cereal.
We also used to have a shopper loyalty program called Kellogg Family Rewards where people can do things like show receipts with a proof of purchase to enter into a sweepstakes.
So we actually are a bit of a unicorn in the CPG space, in that we have a really nice first-party data asset.
This interview has been edited and condensed.
