Commerce Media Has Officially Outgrown Retail
Retail media started as a retailer-only conversation. Today, it spans across a number of industries, including airlines, banks, fintech platforms, and beyond. On a recent episode of the Aperiam Podcast, Kevel founder and CEO James Avery joined hosts Corey Ferengul and Joe Zawadzki to unpack exactly how the category has matured, where the biggest opportunities still lie, and what separates world-class commerce media networks from everyone else.
If you work in retail media, ad tech, or commerce strategy, this episode is essential listening. Here is a breakdown of the most important takeaways.
What Is Kevel and Why Does It Exist?
Kevel's mission is straightforward: help any retailer, bank, or travel company build a commerce media network that rivals Amazon's. James started the company more than 16 years ago with a vision he describes as "Twilio for ads", pure API infrastructure that any company could build on top of.
The platform covers the full stack: ad serving and ad decisioning for promoted listings and native ads, audience management that turns loyalty and order data into actionable targeting, self-serve brand interfaces, and increasingly, off-site reach extension across platforms like Meta and Google using first-party retailer data.
What Kevel does not do is equally important. The company is exclusively supply-side focused. Kevel does not sell to brands or act as an ad network and Its customer is always the retailer or commerce platform.
From a Box of Legos to a Pre-Built Infrastructure
James is candid about Kevel's early positioning challenge. When the company led with APIs and developer infrastructure, some retailers chose competitors simply because those competitors would write them a check. Kevel’s value proposition, while technically superior, was harder to grasp.
"Instead of a box of Legos with no instruction set, we wanted to be a pre-built toy. A pre-built infrastructure made out of Legos that you can rearrange and customize, one that is more approachable to the rest of the market," James explained.
Kevel’s evolution has also included building demand connections to partners like The Trade Desk, Pacvue, and Skai, so retailers using Kevel's infrastructure can now access automated revenue streams without sacrificing tech quality.
New Markets vs. Legacy Lift-and-Shifts
Kevel serves two distinct customer profiles today. Larger retailers typically arrive with existing tech, often a Criteo or Citrus stack, or something built in-house, and need a structured migration. Newer commerce media players in verticals like travel and fintech, or up-and-coming regions like Latin America, are often starting from scratch, which means faster deployment and cleaner architecture from day one.
SAS Airlines and Nubank are two recent examples of greenfield deployments. James noted that launching a commerce media network from scratch can take as little as a few weeks when there is no existing infrastructure to unwind.
The legacy challenge James describes most often is fragmentation: retailers who cobbled together three or four providers for promoted listings, banners, and native ads separately, and then struggled to give advertisers unified, consistent reporting. Consolidating on a unified stack solves that immediately.
Selling Success, Not Software
One of the most interesting pivots James shared involves how Kevel thinks about customer success. A year ago, the team reframed its mission from selling software to selling outcomes.
"We have to sell success, not software," James said. "If you just sell somebody some software and they use it and churn, it does not help anyone."
Kevel hired a team of commerce media strategists, with extensive expertise in building commerce media networks, unlocking new revenue streams, scaling go-to-market operations and driving commercial impact, to work directly with customers on go-to-market planning. That includes helping with media kits, identifying high-value brand partnerships, and building out sales org strategy. Kevel does not charge extra for this work. The bet is that customer success drives retention and growth, so the investment pays for itself.
The Four Ps of a World-Class Retail Media Network
Reflecting on what separates high-performing retail media networks from underperformers, Joe from Aperiam crystallized James's thinking into a framework: the four Ps.
- People: The right hires make the difference. Dollar General's results accelerated dramatically after bringing in a strong GM. Chewy built a standout sales org with deep industry experience.
- Playbooks: Clear, repeatable go-to-market strategies that identify top brand partners and prioritize high-value relationships.
- Philosophy: The best retail media networks treat advertiser dollars as something to earn, not something owed. They compete to be a better investment than Facebook, not just a trade obligation.
- Product: A unified, first-party-data-native stack that gives brands consistent reporting, clean attribution, and real proof of performance.
Offline Attribution: The Quiet Unlock
One of the most underappreciated topics in the conversation was offline attribution. James described working with a European retailer where digital ads initially showed a weak return on ad spend. Once Kevel connected in-store purchase data through loyalty IDs, the picture changed entirely.
Consumers were seeing ads digitally, researching products online, and then completing purchases in-store. Without connecting those touchpoints, the ads looked ineffective. With the full attribution picture in place, brands saw the real ROI and increased their spend accordingly.
"In-store offline attribution has a huge impact on the ROI that brands see," James said. "And then they want to spend more, because they see that it is actually working."
This is a significant unlock for any retailer with a physical footprint, and it is one that Kevel's infrastructure is specifically built to support through loyalty ID matching and server-to-server integrations.
Travel and Fintech: The Next Big Commerce Media Frontier
James is most excited about two verticals: travel and fintech. Both sit on rich first-party data that has historically been under-leveraged.
He used a simple example to illustrate the opportunity: logging into a Delta seatback screen mid-flight and being served a generic Delta ad. The airline knows exactly who that passenger is, where they are going, and a great deal about their consumer behavior. Brands would pay a significant premium to reach that person at that moment. The infrastructure to make that happen is still being built.
The same logic applies to financial platforms like Nubank, which hold detailed transaction data and behavioral signals that are genuinely differentiated from what any other media channel can offer.
AI and Machine Learning in Commerce Media
James made an important distinction between machine learning, which has always been at the core of ad serving for relevancy scoring and auction optimization, and the newer wave of large language model(LLM) applications.
On the LLM side, Kevel recently launched an AI reporting agent that lets users query campaign performance in natural language instead of generating 17 different spreadsheets. They also launched an MCP server so customers can use Claude to manage campaigns directly, like turning off a set of campaigns with a single instruction rather than manually checking boxes in a UI.
"There are tons of places where AI is improving the workflow for the customer," James said. "Making their lives easier and making it easier to scale."
The DOJ Antitrust Case Against Google: James's Take
James was a witness in the DOJ's antitrust case against Google's ad tech business, and his take on the experience is characteristically direct: he found it fun. His core argument in the case was that if publishers could access demand from Google's AdX exchange through fair, competitive means, companies like Kevel could compete more effectively against Google Ad Manager on merit.
Whether or not the case produces the outcome James hoped for, it reflects Kevel's broader conviction that a more open, competitive ad tech ecosystem benefits publishers and advertisers alike.
The Human Behavior Argument Against AI Maximalism
James hosts his own podcast, Unlocking Retail Media, and he noted one trend he consistently pushes back against: AI maximalism. The argument that AI agents will soon handle all purchasing decisions runs counter to what he observes about actual human behavior.
"What I have learned is that human behavior does not actually change. The tools and technologies change, but the behavior does not," he said.
People still go to malls. They still walk into Home Depot and pick out a refrigerator without prior research. They still buy things at gas stations at a rate that exceeds Amazon. The technology may evolve, but commerce remains deeply human, and that is ultimately why retail and commerce media continues to grow.
Key Takeaways at a Glance
- Commerce media has expanded well beyond grocery retail into travel, fintech, and other data-rich verticals.
- Kevel's differentiator is its supply-side-only focus, API-first architecture, and first-party data infrastructure.
- Offline attribution is one of the biggest unlocked ROI opportunities in the space right now.
- The best retail media networks compete to earn advertiser dollars rather than relying on trade obligations.
- Travel and fintech represent the next major wave of commerce media growth.
- Human shopping behavior is more resilient than AI maximalists suggest, which is good news for commerce media long-term.
