Unit economics: revenue per screen & payback

Everyone wants the same two numbers before they buy a screen: what does it earn a month, and when does it pay for itself? The uncomfortable truth is that neither is a number you can look up. We went looking for a credible public “revenue per beauty screen” benchmark and a “DOOH payback period” benchmark, and neither survived verification — the figures that circulate are either someone’s marketing, or a billboard average wearing an in-venue costume. What you can anchor is the price of an impression and the haircut between what an advertiser pays and what lands in your account. Everything else you build bottom-up, from a ceiling set by the venue itself. This guide shows you the model, the two inputs you can trust, and the inputs that are honestly just your assumptions — so you can see exactly where your own answer comes from.

1. Why “revenue per screen” is a formula, not a number

There is no shelf you can pull a beauty-screen revenue figure off. So instead of quoting one, hold the chain that produces it:

Monthly net revenue per screen = impressions/mo × fill rate × (gross CPM ÷ 1,000) × (1 − supply-chain take) × (1 − venue share)

Then subtract the monthly running cost to get what actually services the hardware:

Monthly contribution = monthly net revenue − OpEx/mo → Payback (months) = CapEx ÷ monthly contribution

Every term on the right is either something you can anchor to a source or something that is honestly your assumption. The value of writing it out is that it forces the question onto the right inputs — and it makes plain that a single weak assumption (say, an optimistic fill rate) propagates straight into a fantasy payback. Let’s take the terms one at a time, hardest-to-fake first.

2. The impression ceiling — the input you don’t control

Start with the input most people get wrong by treating it as “more screens = more impressions.” It isn’t. The cleanest statement of the rule comes from a supply-side platform that forecasts exactly this: “1,000 daily visitors translate to a total opportunity of 10,000 impressions per day… You can add more screens and capture more of the audience but you can never capture more than 10,000 impressions. There is a limit!” (Vistar Media — primary). Past an optimal screen count, each additional screen in the same room loses money.

So a venue’s impression supply is a ceiling, set by:

• Foot traffic — how many people pass through, which for a salon is essentially its booked clients plus companions.
• Dwell opportunities — how many ad plays each of those people has the chance to see while they’re in the exposure zone.

This is where beauty’s structural edge lives: a salon client is captive for the length of an appointment, not a billboard-second, so each visit yields many creditable audience impressions rather than one (see dwell time benchmarks). But it is still a ceiling — high dwell raises it, it doesn’t remove it. Your model’s impression term is foot traffic × plays-seen-per-visit, and plays-seen-per-visit comes from your loop design (dwell ÷ loop length), which is a choice you make, not a benchmark you cite. We found no trustworthy public loop/slot-structure benchmark for beauty venues — so treat your impressions figure as a venue-measured input, never a borrowed one.

3. The price: CPM — the one number you can anchor

CPM is the price per thousand impressions, and it’s the only input with a real, primary-sourced benchmark. The clearest series comes from one large programmatic OOH marketplace:

• $7.16 (H1 2024) → $7.62 (H2 2024) → $6.53 (H1 2025) (Place Exchange — primary).

Two honest caveats, both load-bearing. First, that average is billboard-heavy — roadside and large-format dominate the spend behind it — so it is not an in-salon rate. Second, broader guidance puts programmatic DOOH in a wide $2–$15 band (StackAdapt — vendor blog), with place-based in-venue screens (malls, gyms, offices) cited higher at $8–$30 (AdQuick — directional). Does beauty command a premium? Directionally yes — the same marketplace noted that “Healthy/Beauty” venues saw higher CPMs in H2 2024 (Place Exchange — primary) — but it was grouped as a secondary-tier increase, and no public beauty-specific dollar CPM exists. The discipline here: anchor your model to the mid-single-digit marketplace average, treat the $8–$30 place-based range as upside to argue for, and never plug $30 into a spreadsheet as if it were a fact.

4. The haircut: gross to net

Here is the line that quietly halves more business plans than any hardware overrun: the CPM is gross. What reaches your account is smaller, eaten in two bites.

Bite one — the ad-tech supply chain. Historically, only about half of programmatic ad spend reached the media owner: a landmark audit found publishers received 51% (range 49–67% by deal type), with a 15% “unknown delta” that couldn’t be traced to anyone (ISBA/PwC 2020 — primary). Combined DSP+SSP fees average around 35% of the bid, and 22–45% for half of all impressions (Adalytics — primary). Two crucial mitigations, though: that audit is open-web display/video, not DOOH; a 2022–23 follow-up lifted the publisher share to ~65% with the unknown delta cut to ~3%; and programmatic DOOH runs a lighter chain — total supply-chain take is more like ~10–25% via private marketplaces, up to ~30% open market (Mi3 / ANA — medium confidence), and direct in-venue sales can skip the SSP/DSP stack entirely. Net: a DOOH operator likely keeps more than the open-web 50% — call it ~70–90% on programmatic-DOOH, more if sold direct — but treat the exact figure as a range to verify, not a constant.

Bite two — the venue’s share. The host wants paying for the wall, and it comes out of your net, not the advertiser’s gross. This is the term with the weakest public data of all: we found no trustworthy benchmark for the operator-vs-venue split — whether it’s a percentage of net ad revenue or a flat rent. It is a negotiation, it varies per venue, and it directly suppresses your per-screen take. Model it explicitly (and conservatively); don’t assume it away. See the operator playbook in how to launch a beauty DOOH network.

5. Fill rate: the swing factor and the great unknown

Fill rate — the share of your ad slots actually carrying paid creative — is where payback is won or lost, and it is the input we most wanted to benchmark and most decisively couldn’t. Multiple circulating young-network fill figures (a widely cited “41% fill / 61% bid-rate” claim among them) were refuted under verification — they didn’t hold up. So there is no number to borrow; there is only a discipline:

• An installed, unsold screen earns nothing — it’s pure cost with a glowing house promo on it.
• Early fill is low and ramps slowly. A new network has thin demand until it has scale and a sales motion; assume months of sub-break-even fill, not an instant full loop.
• Fill is multiplicative in the revenue formula, so it scales the entire line. Halve fill and you halve revenue while OpEx stays fixed — which is exactly how a screen tips into permanent loss.

Because fill is both the largest swing factor and the one with no public anchor, conservative fill modeling is the single most important honesty test of your plan.

6. Payback: putting it together — and what the shape tells you

Now the second question: when does the screen pay back? Payback = CapEx ÷ monthly contribution, and we can populate the cost side from live, sourced ranges even though no payback benchmark survived:

• CapEx: a generic fully-installed signage screen runs roughly $1,800–$3,500 in year one (CrownTV — vendor estimate); an in-mirror/bezel-free beauty unit runs higher and is quote-based (full breakdown in how much does it cost to start a network).
• OpEx: CMS SaaS is the one cleanly priced line at $8–$30 per screen/month (Yodeck, ScreenCloud — primary), plus connectivity and content (no reliable public per-screen connectivity figure exists).

Notice what’s not on this list: a “DOOH pays back in N months” figure. We won’t invent one, because the honest output depends entirely on the two inputs with no public anchor — your fill rate and your venue share. What we can state are the two structural truths the arithmetic guarantees, and they matter more than any single payback number:

1. There is a breakeven fill rate below which payback is never. Because OpEx is a fixed monthly drag, monthly contribution goes negative whenever revenue (already small, at a mid-single-digit net CPM) falls under it. Below that fill, a cheaper screen doesn’t help — the screen loses money every month it’s lit, forever.
2. Above breakeven, payback is governed by the fill ramp, not the panel price. Net contribution is small per screen, so halving fill roughly doubles payback while shaving 20% off the hardware barely moves it. The lever is sold impressions, not cheaper hardware.

The practical reading: a single beauty screen is marginal economics — modest impressions × a mid-single-digit net CPM, against a fixed monthly cost, makes one screen a slow or losing proposition. The unit economics only work as a network, where a sales motion lifts fill across many screens and shared OpEx and overhead amortise. That’s the same conclusion the cost guide reaches from the spend side: cheap to build, expensive to keep alive until it sells.

So — what does a screen earn, and when does it pay back?

The honest answer, in the only form the evidence supports:

• Earnings per screen = (foot traffic × plays-seen-per-visit) × fill × net CPM ÷ 1,000 × (1 − venue share). Two of those inputs are anchorable (CPM ~mid-single-digits gross; supply-chain take ~10–30% on programmatic DOOH); the rest are your venue-measured assumptions. There is no public per-screen revenue benchmark to shortcut this.
• Payback = CapEx ÷ monthly contribution. CapEx is a knowable few-thousand dollars (more for a mirror unit); the denominator is set by fill and venue share, so payback ranges from “a couple of years” at strong fill to “never” below breakeven. There is no public payback benchmark; compute it, don’t cite it.

Build the model, plug your own live quotes and your own conservative fill ramp, and watch the breakeven fill line — not the hardware price. That number tells you whether you have a business.