Programmatic buying is only getting more (and more and more) automated and agentic. But more accountable? That’s a different story.
Precise.ai, a startup founded by a pair of ad tech vets, is trying to change that by applying an approach called “decisioning economics” to programmatic, which involves measuring the cost and contribution of everything that goes into buying an impression instead of only looking at the end result.
Agencies and brands get plenty of performance metrics from their partners, said Spencer Potts, CEO and co-founder of Precise.ai. What they usually don’t get is a clear sense of which targeting, bidding and optimization choices actually helped and which just added cost.
On Wednesday, Precise raised a $7 million seed round co-led by Blockchange Ventures and early-stage VC firm Lasagna, with participation from Michael Kassan’s 3C Ventures and Click Ventures, the VC firm founded by Mediaocean CEO Bill Wise.
The pitch resonated with investors, Potts said, because so much money is moving through automated systems with very little visibility.
“Everyone has outcome measurement,” Potts said. “What they don’t have is the contribution value of every element that went into a decision before they bought an impression and a way to see exactly what’s working and what’s not.”
Laying the foundation
But Precise didn’t start out as a tool for picking apart media decisions.
During its first act, the startup worked on privacy and consent infrastructure that used blockchain to create a portable, verifiable record of how data moved through the ad supply chain.
In 2024, Qonsent’s privacy and consent tools were combined with blockchain-based data infrastructure from Valence Labs – an R&D shop founded by Adam Helfgott, now Precise’s co-founder and chief architect – to form Precise.ai, with a focus on getting permission to collect data, tracking how that data was used and creating a verifiable record of where it flowed.
(Because it’s a small world after all, Helfgott and Potts also previously worked closely together at TV ad platform Madhive. Helfgott co-founded Madhive and still serves as executive chairman, and Potts was CEO between 2019 and when he left last year to take on the chief executive role at Precise.)
Over time, Helfgott realized that the same provenance layer he was building for privacy could also serve as a “decision accountability” layer, a way to log which decisions were made on what inputs and with what impact on performance and cost. Precise changed its focus and wound down the Qonsent side of its business.
“Having that kind of power over supply-chain economics was always the vision, even in the early days of Madhive,” Helfgott said. “If you know the value of something in connection with everything else, you can incentivize players that provide real value, but maybe they couldn’t prove it in the past.”
Precisely how it works
Precise now uses that same record-keeping foundation to track and evaluate the decisions happening inside live campaigns.
The platform ingests log-level data from DSPs and other ad tech systems, everything from which segments were used and what bid modifiers and prices were applied to which optimizations and supply paths were chosen and how the impressions were attributed.
“The more data we can look at from a runtime point of view the better,” Helfgott said.
From there, Precise runs a contribution analysis on each of those inputs to quantify how much each variable is actually helping (or not) and surfaces that information in a dashboard. “You might have five, six, 10, even 15 inputs that go into the decision to buy a particular impression for a particular campaign at a particular time,” Potts said. “What we’re doing is splitting those apart to figure out the best combination to get the best outcome.”
Although the ambition is to score every parameter in a transaction, for now Precise is starting with relatively simple components, like audience segments.
“If a DSP is appending 45 different segments or whatever it is, you should be able to know whether and which ones are contributing,” Potts said. “That’s just some very low-hanging fruit.”
Eventually, the same math can be extended to other inputs like supply paths, bid prices and even creative.
“In the supply chain, you make a decision to buy an impression and then you measure the outcome,” Potts said. “We’re sitting upstream to break that down almost to the atomic level.”
Testing the thesis
Precise is doing six pilots, including one with a large holding company, and meeting with Mediaocean’s partnership team in Cannes next week to talk about testing.
For the time being, the focus will be on proving that decisioning economics can go beyond reporting and influence how campaigns run and budget gets allocated. The $7 million seed round is meant to give Precise enough runway to do that with a small set of big buyers before scaling more broadly. The Precise team itself is lean – just eight people – and some of the new funding will go toward making a few senior product and commercial hires.
But the idea is to stay small even as the buying landscape gets more agentic and complex. When media-buying decisions happen at machine speed, agencies have less and less visibility into how those calls are getting made.
“And it’s only going to get faster as the agents get smarter,” Potts said. “We want to be an independent intelligence layer that holds everybody responsible for proving their contribution value to a campaign.”