On-Site Ads and Retail Media: When Is It Worth It for Your Store?

Every few months someone forwards me an article about retail media being the highest-margin business in retail, and asks if they should build an ad network on their own store. My answer is almost always the same: not yet, and here's the trap. Retail media is genuinely one of the most profitable things in commerce. It's also the single most premature thing a small store can chase. Both of those are true, and the gap between them is where a lot of operators waste a year. I've sat on both sides of this. I've spent money as an advertiser inside someone else's retail media, watching ROAS like a hawk. And I've looked at building sponsored placements into a store's own search results. The lessons from those two seats are completely different, so let me separate them cleanly, because "retail media" gets used to mean two very different decisions.

What retail media actually is

Strip away the jargon and retail media is just advertising that happens on a retailer's own property: sponsored products in search results, banner placements on category pages, that "sponsored" tag next to a listing. The retailer sells ad space to brands who want their products seen. On-site means it runs on the retailer's website. Off-site means the retailer uses its shopper data to target ads elsewhere. The reason everyone's obsessed is the margin. Forrester data cited across the industry puts on-site retail media margins at 70% to 90%, versus the 3% to 4% net margin a typical retailer scrapes from selling actual products.[1] You're selling pixels, not pallets. Off-site is lower, usually 20% to 40%, because you're buying media to resell.[2] That spread is why a $100 ad sale can leave $60 to $70 of profit.

The story in the numbers

The money is following the margin. US advertisers will spend more than $62 billion on retail media in 2025, over $10 billion more than the year before, per eMarketer.[3] eMarketer projects $69.33 billion in 2026, and pegs the category's growth at roughly 20% in 2025 against just 4.3% for the total ad market.[4] Retail media is growing about five times faster than advertising overall.

Chart by author. Source: eMarketer retail media forecasts, 2024–2026. 2024 figure ($52B) and 2026 ($69.33B) are eMarketer estimates; 2025 ($62B) reflects eMarketer's 2025 spend figure. But look at who's capturing it. eMarketer reports Amazon Ads and Walmart Connect alone are taking $9.42 billion of the $10.53 billion in incremental retail media spend heading into 2026.[4] Walmart Connect grew about 30% for the full year to reach a global ad business around $3.4 billion.[5] The lesson isn't "retail media is easy money." It's that retail media is a scale game, and the giants already have the scale.

On-site retail media is the digital version of paying for the end-cap. Photo: Pixabay via Pexels (Pexels License).

The two decisions, compared

*Estimates. ROAS/ACoS benchmarks vary widely by category and account maturity; margin figures are Forrester-cited industry estimates.

When it's worth it, and when it's not

Let me be blunt about the two paths, because they get confused constantly. Buying ads on a retail media network (like Amazon Ads) is worth it early. If you sell on Amazon, sponsored placements are basically table stakes. Industry benchmarks put average Sponsored Products ROAS somewhere around 3x to 4x, with ACoS for established accounts often landing in the 18% to 22% range, though new launches can spike to 40% to 60% before optimization (label these as estimates, they swing hard by category).[6] The rule I use: if your margin survives a 30% ACoS and you still profit, advertise. If it doesn't, fix the product economics first. Building your own on-site ad network is almost always premature. Here's the uncomfortable truth: those 70–90% margins only exist if you have enough shopper traffic that brands will pay to reach it. Amazon and Walmart can sell ads because millions of high-intent shoppers show up daily. If you're doing a few hundred orders a month, you don't have an audience worth advertising into, you have a store that needs more customers. Selling ad space to your own suppliers when you have thin traffic just clutters your site and annoys the buyers you fought to get. So what's the honest threshold? I'd look for a few things before even thinking about on-site sponsored placements: consistent, meaningful traffic; a marketplace model with many third-party sellers or brands competing for visibility; and enough catalog that a "sponsored" slot actually changes who wins the sale. If you're a single-brand DTC store, you are the only advertiser, so there's no network to build. Your version of "retail media" is just merchandising, and you should do that for free. If you do reach that scale, the build is non-trivial: auction logic, placement rendering, billing, reporting for advertisers. A platform that already handles catalog, search and seller management gives you a foundation to add sponsored placements on top, rather than bolting an ad server onto a storefront that wasn't designed for it. That's the kind of problem Sellarix is built around, though I'd only reach for it once you've genuinely got the traffic and seller base to monetize.

The takeaway

Retail media is a fantastic business at scale and a distraction before it. As an advertiser, lean in early: ads on Amazon and similar networks reach high-intent buyers and the ROAS math is usually workable if your margins are healthy. As a publisher, wait. You can't sell access to an audience you don't yet have. Build the traffic and the seller base first; the 70–90% margins will still be there when you're actually ready. So before you chase the shiny ad-network revenue, ask yourself: do I have an audience brands would pay to reach, or do I just have a store that still needs more customers? Be honest. The answer tells you exactly which retail media decision you're actually allowed to make.

Sources

• eMarketer — $10 billion incremental ad spending into US retail media: https://www.emarketer.com/content/10-billion-incremental-ad-spending-will-flow-us-retail-media-2025
• eMarketer — Retail Media Ad Spending Forecast and Trends H2 2025: https://www.emarketer.com/content/retail-media-ad-spending-forecast-trends-h2-2025
• The Retail Exec — High-margin retail media networks (Forrester 70–90% margin): https://theretailexec.com/retail-operations/retail-media-networks/
• Mirakl — How do retail media networks make money: https://www.mirakl.com/blog/how-do-retail-media-networks-make-money
• Marketing Dive — Walmart Connect retail media growth: https://www.marketingdive.com/news/walmart-q2-2024-earnings-retail-media-networks/724470/
• Triple Whale — Amazon Ads benchmarks (ROAS/ACoS): https://www.triplewhale.com/blog/amazon-benchmarks