SWYM.ai CEO: The Largest DSPs See Only 40% of Available Ad Supply

The largest demand-side platforms see only roughly 40% of available ad supply, and smaller DSPs see just 20% to 25%, according to Ravi Patel, CEO and co-founder of SWYM.ai. The reason, he says, is that DSPs throttle traffic to keep their cloud costs down, leaving buyers to pick inventory from a partial view of the market.

Patel laid out his case on this week's episode of AdExchanger Talks, drawing on a grocery-store metaphor he favors: running a programmatic campaign is like walking into a store with a budget you must spend before closing time, with no way to tell the lemons from the cherries until you have already paid. Worse, the store learns from your purchases: buy a lot of lemons, and it assumes you love them and keeps stocking more.

Why do SSPs and DSPs keep serving buyers the same bad inventory?

The structural problem, Patel argues, is that supply-side platforms and demand-side platforms "talk at each other and not to each other." There is no feedback loop between them. An SSP sees only which bid requests a buyer won, not whether any of those impressions helped the buyer achieve its goals. The consequence, in his words: "If you bought a lemon yesterday, you're getting a lemon today." That is why filtering has to happen earlier in the transaction, he says.

A caveat follows the promise. Patel's figures on supply visibility are his own estimates, not audited data, and SWYM has a commercial interest in arguing that pre-bid filtering is broken. The episode itself is a promotional format; no independent measurement of the claimed visibility gap appears in it.

What is supply shaping, and how is it different from curation?

SWYM's product is AI-powered supply shaping, delivered through containerization — in plain terms, its software runs inside an SSP's own systems rather than sitting outside them.

Patel is quick to draw a line between his approach and curation, which historically has meant packaging an audience, attaching a deal ID and adding a markup. Supply shaping, he says, judges each bid request individually and only on whether it is likely to lead to an outcome. "I don't care that user 12345 lives in New York City and loves to snowboard," Patel says. "I just care whether that potential bid request can lead this client to a sale."

Notably, the segment does not specify how SWYM measures or attributes those outcomes — the mechanism behind the 'likely to lead to a sale' judgment goes unexplained.

Does the proof point hold up?

The episode cites one concrete client result: SWYM helped the paper and packaging giant Georgia-Pacific cut its SSP roster by 80% without losing reach. That figure is attributed to Patel and SWYM, not to Georgia-Pacific directly, and no baseline metrics accompany it.

The conversation also covers the rise and fall of Patel's previous startup Ad/Fin, his view that AI-generated sites are not necessarily bad if real humans visit them, and his skepticism that the open web is dying — a death, he notes, that has been predicted before: "Hey, they said the same thing about linear, and it's still here!"

The company's name, fittingly, is short for "Stop Wasting Your Money." Whether supply shaping becomes the industry's standard filter or stays a vendor's pitch will depend on outcomes data SWYM has not yet put on the table.