Programmatic shopping for is simply getting extra (and increasingly more) automated and agentic. However extra accountable? That’s a special story.
Exact.ai, a startup based by a pair of advert tech vets, is making an attempt to vary that by making use of an method known as “decisioning economics” to programmatic, which entails measuring the associated fee and contribution of all the pieces that goes into shopping for an impression as an alternative of solely trying on the finish consequence.
Companies and types get loads of efficiency metrics from their companions, stated Spencer Potts, CEO and co-founder of Exact.ai. What they normally don’t get is a transparent sense of which concentrating on, bidding and optimization decisions truly helped and which simply added value.
On Wednesday, Exact raised a $7 million seed spherical co-led by Blockchange Ventures and early-stage VC agency Lasagna, with participation from Michael Kassan’s 3C Ventures and Click on Ventures, the VC agency based by Mediaocean CEO Invoice Clever.
The pitch resonated with traders, Potts stated, as a result of a lot cash is transferring by automated programs with little or no visibility.
“Everybody has consequence measurement,” Potts stated. “What they don’t have is the contribution worth of each aspect that went into a choice earlier than they purchased an impression and a method to see precisely what’s working and what’s not.”
Laying the muse
However Exact didn’t begin out as a device for choosing aside media choices.
Throughout its first act, the startup labored on privateness and consent infrastructure that used blockchain to create a transportable, verifiable report of how information moved by the advert provide chain.
In 2024, Qonsent’s privateness and consent instruments had been mixed with blockchain-based information infrastructure from Valence Labs – an R&D store based by Adam Helfgott, now Exact’s co-founder and chief architect – to type Exact.ai, with a concentrate on getting permission to gather information, monitoring how that information was used and making a verifiable report of the place it flowed.
(As a result of it’s a small world in spite of everything, Helfgott and Potts additionally beforehand labored carefully collectively at TV advert platform Madhive. Helfgott co-founded Madhive and nonetheless stays on the board, and Potts was an exec there, together with as CEO between 2023 and 2025, when he left to tackle the chief government position at Exact.)
Over time, Helfgott realized that the identical provenance layer he was constructing for privateness might additionally function a “determination accountability” layer, a method to log which choices had been made on what inputs and with what influence on efficiency and price. Exact modified its focus and wound down the Qonsent facet of its enterprise.
“Having that form of energy over supply-chain economics was at all times the imaginative and prescient, even within the early days of Madhive,” Helfgott stated. “If you understand the worth of one thing in reference to all the pieces else, you possibly can incentivize gamers that present actual worth, however perhaps they couldn’t show it prior to now.”
Exactly the way it works
Exact now makes use of that very same record-keeping basis to trace and consider the selections taking place inside stay campaigns.
The platform ingests log-level information from DSPs and different advert tech programs, all the pieces from which segments had been used and what bid modifiers and costs had been utilized to which optimizations and provide paths had been chosen and the way the impressions had been attributed.
“The extra information we are able to take a look at from a runtime perspective the higher,” Helfgott stated.
From there, Exact runs a contribution evaluation on every of these inputs to quantify how a lot every variable is definitely serving to (or not) and surfaces that info in a dashboard.
“You might need 5, six, 10, even 15 inputs that go into the choice to purchase a specific impression for a specific marketing campaign at a specific time,” Potts stated. “What we’re doing is splitting these aside to determine the very best mixture to get the very best consequence.”
Though the ambition is to attain each parameter in a transaction, for now Exact is beginning with comparatively easy parts, like viewers segments.
“If a DSP is appending 45 totally different segments or no matter it’s, it is best to be capable to know whether or not and which of them are contributing,” Potts stated. “That’s just a few very low-hanging fruit.”
Ultimately, the identical math could be prolonged to different inputs like provide paths, bid costs and even artistic.
“Within the provide chain, you decide to purchase an impression and then you definately measure the result,” Potts stated. “We’re sitting upstream to interrupt that down virtually to the atomic degree.”
Testing the thesis
Exact is doing six pilots, together with one with a big holding firm, and assembly with Mediaocean’s partnership group in Cannes subsequent week to speak about testing.
In the meanwhile, the main focus will likely be on proving that decisioning economics can transcend reporting and affect how campaigns run and funds will get allotted. The $7 million seed spherical is supposed to offer Exact sufficient runway to try this with a small set of huge patrons earlier than scaling extra broadly.
The Exact group itself is lean – simply eight folks – and among the new funding will go towards making a couple of senior product and industrial hires.
However the concept is to remain small even because the shopping for panorama will get extra agentic and sophisticated. When media-buying choices occur at machine pace, companies have much less and fewer visibility into how these calls are getting made.
“And it’s solely going to get quicker because the brokers get smarter,” Potts stated. “We need to be an impartial intelligence layer that holds everyone answerable for proving their contribution worth to a marketing campaign.”
