All articles
billups Turns Two Decades Of OOH Buying History Into A Globally Queryable Planning System
Shawn Spooner, Global Chief Technology Officer at billups, explains how the company made two decades of buying records queryable in any market, and why AI agents may soon read out-of-home inventory.

Make brandbeat one of your go-to sources on Google
No matter where you are in the world, the context of the out-of-home itself has a huge impact on how effective it is. We built some systems to quantify that worldwide.
billups has spent the past year rebuilding how its planners work across markets. The company launched audrai in November 2025, an agentic system built for audience planning, creative performance testing, and predictive campaign insights. Early deployments compressed complex geospatial mapping from one or two days to under two minutes and cut planning cycles from days to hours. Now, billups is applying the same systems to link inventory, measurement, and campaign outcomes across every market where it works.
Shawn Spooner, Global Chief Technology Officer at billups, was named the company's first chief technology officer in 2022 after serving as Chief Scientist, and his research background covers computer vision and privacy-preserving models of human mobility. He oversees technology and product across North America, EMEA, and APAC, where billups works in more than 20 countries. His work covers the planning tools buyers use and the measurement models underneath them, and reconciling local market knowledge with global systems is the problem that occupies most of his time.
"No matter where you are in the world, the context of the out-of-home itself has a huge impact on how effective it is. We built some systems to quantify that worldwide. If you're looking at a wallscape in Atlanta and you want to compare it to a digital bulletin in London, you can fairly do that based on the size, the lighting, the viewability, how much attention each one gets," says Spooner. A brand that runs a strong campaign in one city and wants the same result somewhere else has to know why the original worked. Creative and audience figures answer part of that question, and billups built its systems to find inventory in the second market that matches everything else.
Local knowledge, now queryable
Out-of-home remains one of the most fragmented supply chains in advertising. A single national campaign can pull a planning team into dozens of separate supplier relationships, and that supplier base runs from national operators down to family-owned billboard companies. Purchase orders, availability checks, and proof of posting still travel by email and phone call across much of it. "In the US there's something like 1,200 suppliers that we could work with," Spooner notes. "Worldwide, it's three to five thousand of them, at least that I know of. If you're doing a big campaign across the US alone, you might hit two or three dozen suppliers, and not all of them have what we call modern systems."
billups answered that fragmentation by making its own buying records searchable to every planner in the company, whatever market they sit in. A plan that used to stall until a colleague in the right time zone came online can now be started on one continent and finished on another. "All of the operational history is now queryable, and that works globally," Spooner explains. "I don't know London very well. I can go ask audrai what I should do in London, and it will pull what our London team has done and say this is what you should consider. Here's the local patterns that matter in London."
A company working across three continents runs into a problem a single-market operator never faces. Clients in different regions ask for different proof of performance, and each region judges a campaign against the measurement its market has had available. "The type of results that the US wants is quite a bit different than Europe, quite a bit different than APAC," Spooner adds. "Even inside of APAC, it depends on the country you're in, what they're looking for, what access they have. The bar is set in different places, which can be quite challenging because you might clear it in the US and completely miss it in Indonesia or vice versa."
Context travels between markets
Digital display runs in a controlled environment where a thousand impressions look broadly alike, and a media plan can treat them as interchangeable. Out-of-home gives planners no such guarantee, which is where the measurement problem that has followed the medium for years begins. Spooner has been building computer vision systems to quantify the differences, unit by unit and market by market. "If I buy a showing of 100 posters in Los Angeles, the quality of each individual poster is going to vary wildly," he says. "The sizing might even vary. The context around them varies."
Unit selection inside a single market carries as much weight as the total volume bought. Two advertisers can buy the same number of boards along the same thirty-mile stretch of highway and produce completely different results for the same spend. "It matters a lot based on which ones you pick," Spooner explains. "Which ones you pick controls how many people see it, how many of them see more than one, how often they see them, and when it happens on their journey. If you're really strategic and you pick just the right ones in a small area, you could get very little overlap between who sees them."
Spooner applies the same modeling to creative rotation. Advertisers who buy most of the boards along a highway corridor tend to run many versions of a single message, a practice the industry has followed for years without much evidence behind it. "We can tell you that you need to have the versions of the creative because you have so much frequency you're going to burn out after so many days," he notes. "We're trying to model out what that burnout looks like, what your optimal creative refresh strategy looks like, where you roll that out, how you make it measurable."
Agents at the buying desk
Spooner returns often to what AI systems will be able to do with out-of-home once they start placing media themselves. An agent can only evaluate a channel that publishes structured inventory, pricing, and performance data in a form it can read, and Spooner sees little of that in out-of-home today. "One thing I think people are maybe undervaluing is making our space legible to agents," he adds. "So much of what we do is human centric and has been human centric since the beginning that people have an aversion to having systems talk to other systems."
Agents could begin transacting a meaningful share of media spend within roughly two years, by Spooner's estimate, which leaves the industry a narrow period to prepare. He points to connected TV as the precedent worth studying, since out-of-home never competed for the budget that moved there. "If agents continue to grow and begin to surface for many buyers, open web display is going to decrease in value. That's my hypothesis," Spooner concludes. "If AI can drive that and out-of-home can be more legible and more transactable, we might stand to gain some of that. If we don't, I think we'll stand to stay the same and we'll see those dollars go elsewhere."





