Walk the floor of any high-traffic category page and the "not enough bidders" theory falls apart fast. The same handful of SKUs often compete for the same head-term keywords every single day, sometimes hard. If anything, the popular real estate looks crowded, not empty. So if bidder scarcity isn't the constraint on retail media revenue, something else is, and it sits one level up from any individual auction.

TL;DR

• Retail media revenue isn't capped by too few advertisers bidding on a given keyword. Popular keywords are often intensely competitive already.
• The real constraint is how many independent demand channels are even allowed to reach the auction in the first place. Most retail media setups draw from a small, closed set of partners: direct sales plus a single vendor's own demand pool.
• Real-time bidding (RTB) doesn't add more bidders to an existing keyword auction. It opens entirely new channels, most significantly Google Search and Amazon Ads, that were structurally unable to compete for that inventory before.
• Most vendors can only solve half of this. An ad server can serve ads within its own closed demand pool. A demand connector can pipe in outside budgets. Doing both, inside one ranking logic that governs where new demand actually lands, is a different position entirely.
• The metric worth tracking isn't "how many bidders." It's "how many independent channels."

‍

Where the "Not Enough Bidders" Idea Goes Wrong

It's an easy assumption to reach for: revenue feels capped, so the auction must be thin. But that's usually not what's actually happening at the keyword level. Head terms in a given category can be genuinely crowded, with the same set of brands returning to bid on the same real estate week after week, because that's where the traffic is and everyone already knows it.

The plateau shows up somewhere else. It's not that any single auction lacks competitors. It's that the entire pool of who's even eligible to compete, across every auction on the site, has historically been small and closed.

‍

The Actual Constraint: Channel Access, Not Bidder Count

Retail media has typically run on a narrow set of connections: a retailer's own direct sales relationships, plus whatever demand happens to sit inside one incumbent vendor's ad server. That's the ecosystem, full stop, for most setups. It doesn't matter how aggressively the existing advertisers in that pool bid against each other for a head-term keyword; the ceiling on total revenue is set by how many genuinely separate sources of demand are structurally allowed to reach the inventory at all, not by how fiercely the ones already inside are competing.

That's a channel constraint, not a bidder constraint. And it's the reason a retailer can have a genuinely competitive, well-run auction on its top keywords and still be leaving real revenue on the table: the ceiling was never about intensity of competition among the advertisers already in the room. It was about how few doors led into the room in the first place.

‍

What RTB Actually Opens Is a Door, Not a Bigger Crowd

Real-time bidding gets described casually as "more competition," which makes it sound like it just adds more advertisers to an existing keyword auction. That's not quite the mechanism. RTB opens a channel: a live connection to demand that was previously structurally unable to bid on that inventory, most significantly the advertiser pools already active in Google Search and Amazon Ads, a combined pool worth roughly $100 billion.

Those advertisers aren't new entrants to advertising. They're already spending, just somewhere else, because there was never a connection letting that budget reach a specific retailer's onsite inventory. Clicks sourced through a newly opened channel like this commonly clear at a real premium compared to clicks served through demand that was already directly connected, because the demand itself is coming from a genuinely separate, independent pool rather than the same closed set of partners bidding against each other again.

This is also why "reducing dependency on a small set of partners" is the more accurate way to describe what's happening, rather than "increasing competition." The advertisers already in the room aren't multiplying. The number of rooms is.

‍

Why Most Vendors Can Only Fix Half of This

This is the part worth being direct about, because it's a real structural difference, not a marketing distinction.

A legacy ad server, built to serve ads inside its own closed demand pool without visibility into the organic ranking around it, can only sell what's already connected to it. Adding new channels means renegotiating and rebuilding that connection from inside a system that was never designed to be open. A layer built to sit on top of that system instead of replacing it can add new channels without that rebuild, which is a different starting position entirely.

On the other side, a pure demand connector, something built only to pipe outside advertiser budgets into a site, has the opposite problem: it can open the door, but it isn't the thing deciding what happens once new demand walks through it. If that new demand isn't governed by the same ranking logic already balancing paid and organic results, opening the channel can just as easily cannibalize relevance as expand revenue, because nothing is unifying where the new bids actually land against everything else on the page.

Doing both at once, unifying the ranking and mediating multiple demand channels within that same ranking logic, isn't a feature checkbox. It's a structurally different position from occupying either half of the problem alone. Most of the retail media stack was built to do one side or the other. Very little of it was built to do both inside a single decision layer.

‍

The Question Worth Asking Instead

If revenue feels capped, the useful question isn't "are enough advertisers bidding on our keywords." For a lot of retailers, the honest answer to that question is already yes on the terms that matter. The useful question is narrower: how many genuinely independent demand channels, beyond direct sales and whatever's native to the current ad server, are actually able to reach this inventory today.

That's a smaller, more answerable question, and it's the one that actually explains where the next unit of revenue is going to come from.

‍

Key Takeaways

• Head-term keyword auctions in retail media are often already competitive. Bidder scarcity at the keyword level isn't usually the reason revenue plateaus.
• The real ceiling is channel access: how many independent demand sources, beyond direct sales and a single incumbent vendor's own pool, are structurally able to reach the auction.
• RTB opens new channels rather than adding bidders to existing ones. The advertiser demand behind it, primarily Google Search and Amazon Ads, was already active elsewhere and simply had no connection into a given retailer's onsite inventory before.
• Most retail media vendors solve one half of this problem: either serving ads within a closed demand pool, or connecting outside demand without governing where it lands. Doing both inside one ranking logic is a structurally different position.
• The metric worth tracking is how many independent channels can reach the inventory, not how many advertisers are bidding against each other once they're already in the room.

‍

Frequently Asked Questions

Doesn't retail media already have a lot of advertisers bidding on the same keywords? Often, yes, especially on popular head terms within a category. That competition is real, but it's separate from the constraint on total revenue, which comes from how many independent demand channels can reach the inventory at all, not how many advertisers are competing within the channels already connected.

What's the difference between adding bidders and adding demand channels? Adding bidders means more advertisers competing within an already-connected pool of demand. Adding a demand channel means connecting an entirely separate source of advertiser budget, one that was previously unable to bid on that inventory at all because no connection existed, regardless of how many advertisers are already competing elsewhere.

Does real-time bidding increase competition on existing keyword auctions? Not in the sense of adding more advertisers to a keyword that's already being bid on by the same set of brands. RTB opens a new channel, most significantly to advertiser pools already active in Google Search and Amazon Ads, that were structurally unable to reach that inventory before.

Why can't a legacy ad server just connect more demand sources on its own? A legacy ad server is generally built to serve ads within its own closed demand pool, without visibility into the organic ranking happening around it. Connecting a new external demand source usually means rebuilding that connection from inside a system that wasn't designed to be open, and even if connected, there's no guarantee the new demand is governed by the same logic balancing paid and organic results elsewhere on the page.

What does "reducing dependency on a small set of partners" actually mean? It means a retailer's revenue is no longer entirely dependent on the advertisers inside one incumbent vendor's closed demand pool. Opening genuinely independent channels, governed by the same unified ranking logic, reduces how much of total revenue rests on that one narrow relationship.

‍