By Spencer Potts, CEO, Precise.ai
Every few years, programmatic produces an idea the industry mostly agrees with and then mostly leaves on the table. Supply path optimization is the cleanest recent example.
The thesis was correct. Supply paths in programmatic advertising carry different costs, fees, levels of transparency, and relationships to the publishers on the other end. The choice of which path to use is an economic decision with consequences for the buyer.
That insight was correct in 2017, when SPO emerged as a concept, and it was correct in 2020, when the major holding companies and DSPs operationalized it. It remains correct now, and yet in 2026 most agencies will tell you their SPO practice has plateaued. The early gains were captured, the conversation went quiet, and the next phase, the one where SPO would transform programmatic economics, stayed on the runway. SPO asked the right question and stopped one level too early.
The question SPO asked
Which supply paths are earning their share of the budget, and which are duplicating cost without producing proportional value? That is a good question. It is, in fact, the question programmatic should have been asking from the beginning. The industry arrived at it later than it should have, and once it did, the early implementations did real work. Some agencies and brands found genuine efficiencies, the DSPs that built credible SPO offerings differentiated themselves, and the cleanest paths were rewarded with more spend.
Then the conversation shrank.
SPO became framed as a procurement exercise. Success became a count of how many intermediaries you could strip out of the path between buyer and publisher. The harder question, whether each remaining decision was actually earning its place, sat underneath the simpler one of cutting hops to save money.
That framing capped the ceiling. Once you eliminate the obvious redundant resellers, you are out of easy cuts and the supply path question goes back on the shelf. The industry decided SPO was about cleanup and walked past the follow-up. The path is cleaner. Are the decisions inside it creating value, or simply producing fewer line items?
The cautionary version of the same story
There is a second reason SPO stalled, and it matters more for whatever comes next.
The verification category is the cautionary version of the same story. When the major verification vendors went to market in the mid-2010s, they had a defensible thesis. Fraud was real, brands were paying for impressions that lived nowhere, and someone needed to call it out. The execution framed the agencies running the campaigns as the problem’s source. The report went to the brand, the brand asked the agency why it was buying fraudulent inventory, and the agency had little to say in response, because it lacked the tools to see the problem the way the vendor saw it. Adoption was inevitable. Repairing the agency relationship took years, and in some corners that repair is still happening.
SPO had a milder version of the same dynamic. Some of the early conversations positioned the agency’s supply choices as the thing under review, which made adoption feel like an audit on the agency rather than a tool for it. SPO stayed on the right side of that line, mostly, and the friction was still real enough to slow everything down.
The lesson sits in plain sight. A category built on grading the agency caps at the agency’s tolerance for being graded. A category built for the agency compounds at the speed of the agency’s own ambition.
The bigger question SPO was pointing at
Step back from the supply path for a moment, and the underlying question SPO surfaced applies to everything programmatic does.
Every campaign is a population of decisions. Audiences, publishers, supply paths, optimization signals, data segments, and fees. Each decision carries a cost. Each one carries a contribution to the outcome. For two decades the buy side has paid for those decisions in the aggregate and judged them in the aggregate, because the decision-level view underneath was beyond reach.
Measurement companies and attribution companies have done meaningful work at that altitude, and both of them operate one level above the decision itself. They describe the cloud. The individual drops live somewhere else.
SPO was the first time the industry asked the cloud-versus-drop question out loud, and it asked it about supply paths. The answer was instructive. It was also incomplete. The same question applies to the audience layer, the data layer, the optimization layer, and every fee sitting alongside them.
That is the category. Decision-level contribution measurement.
The reframe is simple enough to repeat in a hallway. Measurement says what happened. Attribution allocates credit. Decision-level contribution measurement evaluates each decision inside the activation chain against what it cost and what it contributed, while the campaign is still running. Three altitudes, three jobs, each one earning its keep.
When a buyer operates at that altitude, two things happen at once. Offense, where investment moves continuously toward the decisions creating value and away from the ones consuming it. And defense, where every recommendation carries evidence, so the agency walks into a client review, a procurement conversation, or a board meeting with the same kind of receipts a CFO carries into a quarterly close. Offense and defense together become a discipline, and over time a discipline becomes a standard.
That is the category SPO was reaching for and stopped short of.
Built for the agency
The lesson from SPO and from verification is the lesson for whatever comes next. A category that works in programmatic is built for the agency, by people who understand what the agency is trying to accomplish and the leverage it needs in front of clients, finance, procurement, and the platforms it buys through. Granularity at the decision level is the first real cheat code agencies have had in programmatic, and it functions as operating leverage rather than as another audit.
Build that conversation with agencies and brands on the same side of the table, and the part of programmatic SPO was reaching for becomes solvable in a way it has stayed beyond reach until now.
The question was always good. The answer comes from the right side of the desk, and at the right altitude.
The cloud has been measured for years. The drops are finally legible.
