A million small placements, bought like direct mail

Out-of-home had a measurement problem it had stopped noticing

Traditional outdoor advertising was sold on a proxy: how many cars passed a structure. The number was real, and it answered almost nothing an advertiser cared about. It could not tell a film studio whether the people passing were opening-weekend moviegoers. It could not tell a packaged goods brand whether the households nearby bought the category. It could not tell a local services business whether the impressions were falling inside its service radius or two towns over.

Meanwhile, direct mail had been targeting at block level for decades, and layering psychographic segments on top of that. The data existed. It had simply never been joined to physical placement, because outdoor inventory was sold by the structure and the structure was a fixed, expensive, highway-side object.

What Mira built

Three things, and the combination is the whole story.

A national network of small-format placements. Roughly a million of them, spread across the country — placements at the height and distance where someone is standing still and paying attention, rather than driving past at sixty miles an hour. Each one was small and cheap relative to a highway board. That is not a lesser version of outdoor advertising; it is what makes the rest possible. Cheap and dense means the network can be assembled at a scale where it is worth selecting from rather than bought whole.

Geographic and psychographic targeting joined to it. A million discrete placements, each with a known address, stops being a media property and becomes a queryable database. An advertiser could describe an audience the way a direct marketer describes one — where they live, and what kind of household they are — and get back the subset of placements that sat inside it.

The software that made the other two mean something. This is the part that was actually hard, and it was built in-house. A million locations joined to third-party targeting data is a data problem before it is a media problem: reconciling addresses, resolving them to geography, attaching segment data, and then answering an advertiser’s audience description fast enough to be usable in a sales conversation. Without that layer the network is a list. With it, the network is a buy.

That inverts the transaction. You no longer choose placements and hope. You describe who you want, and the network tells you where they are.

Why entertainment, CPG and local services

Those three categories bought it hardest because their problems were the ones density plus targeting actually solves.

Entertainment lives and dies on a release date and a specific moviegoing audience concentrated in specific neighbourhoods of specific cities. A campaign that can flood the right blocks for ten days and then stop is worth far more than one that runs for a month on the wrong freeway.

CPG needed proximity to the store. The useful question was never “how many people saw this” but “how many people saw this within a mile of a retailer that stocks it.” Block-level selection answers that directly.

Local services had the simplest version of the problem and the worst options. A business with a fifteen-mile service radius was being sold impressions from a metro that was ten times too large. Selecting placements by block let them buy only their own market — which for the first time made out-of-home affordable to them at all.

The pattern underneath

This is the same move as the retirement product a decade earlier, made in a different industry. The out-of-home business had accepted its own measurement conventions as a fact of nature. The constraint wasn’t inventory and it wasn’t creative — it was that nobody had connected the physical network to the data layer the direct response world had been using all along, and nobody had built the software to make that connection answerable in real time.

Most categories are not waiting for a better version of the thing they sell. They are waiting for someone to connect it to something that already exists somewhere else.

Where this shows up in the work now

Every audience and channel engagement we run starts with the same question that made Mira work: what is the smallest unit you can actually target, and what data could you join to it that nobody in your category is joining yet? For a dealer group that is a rooftop’s catchment. For a health system it is a service line and a ZIP code. For a DTC brand it is a cohort and a creative.

The technique moves. The categories change.