Booking a Billboard Takes Weeks. We Measured Ours at 72 Seconds.

There's a decent argument going around that out-of-home advertising has a back-end problem: inventory scattered across disconnected systems, availability nobody can confirm on the spot, campaigns that can't be changed while they're running. It lands on a line that's hard to argue with. If inventory can't be surfaced and bought in the moment, it doesn't get bought.

We agree. But the discussion mostly treats this as plumbing, a thing for media owners to sort out among themselves. For the person buying, it's a latency problem. Latency is a product, not a pipe.

So we timed ourselves. Across 727 paid bookings over 60 days, the median time from a customer's payment going through to their ad physically appearing on a screen was 72 seconds.

The short version

• Booking a billboard the traditional way takes days to weeks
• Programmatic was meant to fix that, and it's still only ~7% of global DOOH spend
• We measured payment → on-screen at a median of 72 seconds, 90% inside 5 minutes
• The auction was never the bottleneck. Approval and delivery are
• A minute and a fortnight are not the same product at different speeds

What the industry calls fast

The usual route to a billboard is an insertion order. You brief, they scope, you haggle over a CPM, someone drafts an IO, creative goes for approval, artwork gets produced, and eventually it runs. Booking guides across the industry put lead times at roughly 10 to 15 business days for a standard buy, 2 to 3 weeks for smaller digital campaigns, and 4 to 6 weeks for anything larger or needing print.

Treat those numbers as directional. They come from vendor and agency pages, not an audited study, and as far as we can tell no such study exists. Nobody in the industry disputes the shape of it, though. Weeks, not minutes.

Programmatic DOOH was supposed to collapse all that. It has, a bit, but it's still a rounding error. Global programmatic DOOH spend was $1.4bn in 2025, about 7% of all DOOH spend (WOO / PwC). Flagging this honestly: that figure is self-reported by eleven supply-side platforms and PwC explicitly states it "has not independently verified" it. It's the best number available, not a good one. Either way, 93% of digital-out-of-home is still bought the slow way.

The medium itself is doing fine, mind. UK out-of-home hit a record £1.44bn in 2025, two-thirds of it digital (Outsmart, Route/PwC-verified). The screens went digital. The buying didn't.

What we measured

Every paid booking on FameCake over the last 60 days that wanted to run immediately: 727 of them. Time from the payment clearing to the ad's first recorded play on a physical screen:

Payment → on screen
Fastest 25%	under 44 seconds
Median	1 min 12 sec
75th percentile	2 min 24 sec
90th percentile	5 min 24 sec
95th percentile	24 min

90% are on a screen within five minutes. 96% within the hour.

Widen the lens to the whole job (open the app with nothing, make the ad, see it on a screen) and the median is 4 minutes 55 seconds. That's the number we care about most, because it's the one a customer actually sits through.

And the tail is the tail. The 95th percentile is 24 minutes, and a handful take longer than that. Screens have operating hours, some are fast asleep at 3am, and a post flagged for safety review waits for a human being to look at it. Speed is the norm here, not a promise.

The auction was never the bottleneck

Here's where we think the back-end conversation goes wrong. When people talk about making OOH faster, they talk about the trade: SSPs, DSPs, real-time bidding, unified inventory. That's the part programmatic already fixed.

Between "I've bought this slot" and "my ad is on that screen" sits everything that actually eats the clock:

1. Creative: does the ad exist yet, at the right aspect ratio for that specific screen?
2. Approval: has the media owner agreed to show it? Has anyone checked it's safe?
3. Delivery: has it reached the player, and has the rotation picked it up?

An exchange owns none of those. It brokers a transaction and hands off. The creative pipeline, the approval chain, the CMS, the player and the rotation all belong to other people, and every handoff is a place for hours to go and hide. You can make the auction instant and still wait three days for the screen, which is roughly what "programmatic DOOH" means today for a lot of inventory.

We didn't out-engineer the auction. Nobody here is a genius. We're faster because we own the whole path: the ad gets made in the flow, safety-checked automatically, approved against rules the media owner set in advance rather than case by case, and pushed straight into the rotation. There's nothing to hand off, so there's nothing to be slow at.

No exchange can say that, and it isn't a dig at exchanges. It's a structural difference. A marketplace that connects buyers to other people's systems runs at the speed of the slowest of those systems.

Why minutes and weeks are different products

Cut a three-week lead time to two and you've made the same product slightly better. Cut it to a minute and you've made a different one, because entire categories of advertising become possible that simply weren't:

• The pub with a quiet Tuesday, putting something up at 4pm for tonight
• The barber with three empty chairs this afternoon
• The restaurant with an over-ordered special, advertising it before it goes off
• The shop reacting to weather, or a local event, or the place across the road

None of these can be planned three weeks out, because in three weeks the reason for the ad has evaporated. Reactive advertising is demand the old lead time made structurally impossible to serve, rather than a quicker version of planned advertising. It shows up in nobody's market data, because it never got bought.

That's what "if inventory can't be bought in the moment, it doesn't get bought" really means. Not lost share. Lost market.

The honest caveats

• Our fleet is a mix. Plenty of it is premium roadside: digital 48-sheets on the Tyne Bridge, the A19, the A127 Southend Arterial, and we're adding more. Plenty of it is high-street and shopping-centre screens. A fiver on an indoor screen is not a fortnight on an arterial road, and we won't pretend it is. What's changed is that both of them are now bookable in about a minute.
• 727 bookings is our data, not the industry's. We can tell you what we do. We can't tell you what anyone else does, because nobody publishes it.
• The lead-time figures we're contrasting ourselves against are vendor-sourced and unaudited. We'd love a real study to check ourselves against. There isn't one, which is convenient for us and not very satisfying.
• And the 24-minute tail is ours to own. When we're slow, we're properly slow.

The takeaway

The back-end problem is real and the diagnosis is right: inventory that can't be bought in the moment doesn't get bought. But better plumbing between the same slow systems won't fix it. What fixes it is shrinking the distance between deciding to advertise and being on a screen, until it's short enough that a small business can act on a Tuesday afternoon for a Tuesday night.

For us that's currently 72 seconds from payment, and about five minutes from a standing start. The bar was on the floor. You can see what a screen costs and book one without talking to a single human being.

---

Timings measured across 727 paid FameCake bookings, May to July 2026, from payment confirmation to first recorded play. UK market figures via Outsmart (Route/PwC-verified). Global programmatic figures via the World Out of Home Organization's PwC-aggregated study, June 2026, self-reported by participating SSPs and explicitly not independently verified. Lead-time ranges are from industry booking guides and are not independently audited; we could find no primary study.