A version of this post originally was published to my column at The Drum on May 6th. Read the original here.
Most commerce media networks sell the past.
They know you bought diapers last Tuesday, so they serve you diaper ads this Tuesday. The closed loop, from exposure to transaction, is the foundational value proposition of every retail media pitch deck in existence.
Marco Steinsieck thinks there’s an opportunity to flip that.
“Brands want to know what you’re going to purchase, not what you purchased in the past,” Steinsieck tells me. He’s the general manager of Backpack Media, a new commerce media network built on top of Sallie, an education solutions company focused on helping students plan and pay for college. His previous role: running Sephora’s media network, where he grew the business to what he says became more than $200m in revenue within two-and-a-half years, roughly 3% of the retailer’s sales.
Steinsieck argues that transaction data is a lagging indicator. Life-stage data is a leading one. And Sallie, through its direct relationships with students, parents and recent graduates, has life-stage signals that no retailer can match.
The data asset underneath
Christened Backpack Media, Sallie calls this the first education-powered media network in commerce media. Through Sallie’s private student loan business and broader education support ecosystem, Steinsieck says the network reaches two-thirds of college-bound freshmen each year.
That reach produces a specific kind of signal. Sallie knows when someone is planning for college, enrolling, choosing a major, living on or off campus, graduating and making their first independent financial decisions. Layer data science on top of those signals, Steinsieck’s first five hires were all data scientists, and you can start modeling future behavior with some precision.
“If I know when you’re moving into college, when you’re moving into your first apartment, when you’re getting your first car, when you get your first job, I essentially know your future financial trajectory,” Steinsieck told me. On the Middlemen podcast, he framed the same idea in terms of brand strategy: “These are very high-intent windows in a consumer’s life, and they’re very consequential for brand loyalty. Get in early and win long.”
An advertiser selling credit cards, car insurance or furniture doesn’t need to know what a 21-year-old bought on Amazon last month. They need to know that she’s a senior at a specific university, studying engineering, about to graduate and statistically likely to move to a city with a high cost of living. That’s Backpack Media’s pitch.
Mirakl Ads is the ad-tech solution trusted by Rakuten and 50+ global enterprise retailers.
That’s because Mirakl Ads was built with both 3P marketplace sellers and 1P suppliers in mind. Both advertiser audiences demand a seamless advertising journey from onboarding to reporting.
You can offer everything from Sponsored Products to video all in one solution.
Why differentiated relationships matter more than transaction volume
Steinsieck has a theory about what makes a commerce media network work, and it isn’t transaction volume.
“There’s a misconception that we’re monetizing transaction data,” he told me. “What we’re actually doing is giving brands access to our customer relationships. Do you have strong and differentiated customer relationships? Even if you don’t have huge transaction volume, you can scale.”
He points to Sephora as proof. At $7bn in annual revenue, it will never compete with Amazon or Walmart on reach. But Sephora’s relationship with its beauty customers is deep and sticky, and Steinsieck argues that depth, not breadth, drove the media business to 3% of sales faster than most networks usually achieve.
At Sephora, that meant building an invite-only ads business. Steinsieck turned away spending from deep-pocketed brands that didn’t fit the curation model, prioritized direct brand relationships over agency intermediaries, and focused on formats like the Sephora Squad influencer program on TikTok. This all reflected the retailer’s identity as a curator rather than a marketplace.
“I couldn’t just take L’Oréal and Estée Lauder, who have the deepest pockets, and cover my site with them,” he says. “It never would work.”
Sallie operates under the same logic with different stakes. The customer relationships are deep and differentiated. The data is sensitive. And the network has to be selective about which advertisers it works with, because the moments it monetizes, a student’s first financial decisions, are high-stakes for consumers.
Building inside a financial services provider
That sensitivity is the piece of this story that deserves scrutiny. Commerce media networks built on health data, such as CVS Media Exchange, financial data, such as Chase Media Solutions and PayPal Ads, and now education-finance data occupy fundamentally different territory than a grocery retailer serving sponsored product ads.
Steinsieck is candid about the learning curve. “The steepest learning curve for me has been around working within a financial institution,” he said on the Middlemen podcast. “Everything that we do is about privacy. Everything that we do is around trust. We never sell personal data, ever. Everything’s anonymized, everything’s aggregated, brands never see individual-level information.”
Every network faces some version of the question: does advertising detract from the core relationship with the customer? For a grocer, the worst case is a cluttered search page that alienates customers and loses a product sale. For a financial institution handling student loan data, the worst case is a breach of trust at a formative moment in someone’s financial life.
Steinsieck’s answer is relevance as a guardrail. “We have to anchor to: is this offer relevant to this person? And if we can anchor to that, then the rest we can solve.” Whether that holds as the business scales and the pressure to grow revenue intensifies is an open question, and one that every commerce media network building on sensitive data will eventually face.
What do we even call this?
There’s a taxonomic question worth sitting with. The IAB describes commerce media as a practice where companies use their physical assets, digital properties, first-party data and connection to the shopper journey to create integrated media opportunities, with access to real transaction data. Retail media, per the IAB and MRC’s measurement guidelines, is marketing to consumers near their point of purchase or choice between brands.
Backpack Media fits some of that, but not all of it. There’s first-party data and a connection to consequential consumer decisions. But there’s no point of purchase. There’s no closed loop in the traditional sense. When Steinsieck told me, “We’re not monetizing past transactions. What we are doing is creating predictive commerce media,” I asked him directly: are you really a commerce media network?
“I guess it depends on how you define commerce,” he said. “To me, we are a media network.”
What’s being built across the commerce media landscape is outrunning the categories the industry created to describe it. Chase Media Solutions monetizes card transaction data across every retailer, not just one. PayPal and Venmo serve ads on post-transaction confirmation pages. In the UK, Blue Light Card is building a membership-data-led media network targeting verified frontline and public-sector workers. Each fits some part of the IAB definition and not others.
The definitions will catch up, or they won’t. What matters more is whether the underlying data asset creates genuine value for advertisers, and whether the measurement infrastructure exists to prove it. On that front, Steinsieck acknowledges Backpack Media is early. Measurement has to flex by advertiser type. A CPG brand, a financial services company and a graduate school all think about attribution differently. The network is working with third-party measurement providers including Nielsen and Kantar, rather than building proprietary reporting.
Future-proof
My conversation with Steinsieck naturally made its way to the topic du jour, AI-enabled shopping and agentic commerce. Steinsieck’s perspective is that agents making purchases, or at least informing purchase decisions, will disrupt commoditized, low-consideration transactions first. Grocery, in his view, is vulnerable. High-consideration purchases, less so.
“If my media network is built on commoditized transactions that are low consideration, I’m going to be more quickly impacted by agentic commerce,” he said. The kinds of decisions Backpack Media is built around, your first credit card, your first car insurance policy, your first apartment, aren’t decisions consumers will hand to an algorithm anytime soon.
That’s a convenient position for Steinsieck in his current role. But the logic tracks with what I’ve been seeing: the networks most exposed to disruption are those whose entire value proposition rests on being part of the discovery and consideration process.
Backpack Media is early. It’s unproven at scale. And the predictive commerce thesis has not been validated by the kind of independent research that would make it more than a compelling pitch.
But if its thesis holds, that life-stage signals outperform transaction data as a predictor of future purchases, the implication is that the industry’s fixation on closed-loop transaction data may be optimizing for a signal that’s losing its edge. And the networks worth watching might be the ones built on data the rest of the industry doesn’t have access to.
PS - please contribute to our industry!
EMARKETER is teaming up again with Bain to survey retail media network leaders for a unique insight into the forces shaping the industry. The survey is intended for people working in-house on retail media teams, and it is global this year!
Important to know that the survey is also double-anonymized. The survey organizers will not know, nor are they trying to determine, which specific retail media network a respondent works for. The goal is to understand broader retail media trends, not attribute responses to individual companies.
Please take the survey here.
