Unit economics for investors

The unit economics of a beauty DOOH network are simple to write down and easy to mis-diligence. Most pitches lead with screen count and a CPM; both are the wrong place to look. This guide sets out what actually drives margin, the inputs to interrogate, and why fill rate is the variable that makes or breaks the model.

The model, written down

A network’s revenue decomposes cleanly (framework; full detail in the revenue-per-screen model):

net revenue ≈ Σ screens × foot traffic × fill rate × effective CPM ÷ 1,000 − ad-tech take − venue share − costs

Costs split into per-screen CapEx (screen, player, install) and network fixed costs (platform/CMS, sales, ops). Margin is what’s left. The point of writing it out is to see which terms an investor can actually move on in diligence — and which the pitch tends to hide.

Fill rate is the variable that decides it

Of all the inputs, fill rate swings the model hardest — and it’s the one most often glossed over. An installed screen earns nothing on unsold slots, and early fill is low and lumpy: the no-bid stage is the biggest loss, and demand concentrates in premium markets, so non-tier-1 screens fill worse. Hold everything else constant and the model swings from loss-making (low fill) to attractive (healthy fill) on this one variable. So the first diligence question isn’t “what’s the CPM” — it’s “what fill rate are these screens actually achieving, and is it rising?” A network’s economics live or die here.

Why CPM is the wrong anchor

Two reasons the headline CPM misleads:

It’s gross, not net. A quoted CPM is what the advertiser pays; the ad-tech chain takes a real cut before the network, and the venue takes a share after. What lands is the number that matters — model net.
There’s no beauty benchmark. No audited beauty CPM exists — the only verifiable anchor is a blended, cross-venue ~$7.62 that isn’t a salon rate. So any model built on “the beauty CPM” is built on a fabricated input. Demand the network’s realised effective CPM from actual transactions, not a market figure.

CPM matters, but only as a net, realised number on sold slots — never as a gross sticker on installed capacity.

What scales and what doesn’t

Margin behaviour as the network grows is asymmetric, and it’s central to the investment case (see economics at scale):

Scales (improves margin with size): fixed costs (platform, sales overhead, ops) amortise across more screens; one programmatic integration serves the whole fleet; track record compounds demand.
Doesn’t scale (stays roughly linear): per-venue direct sales, venue acquisition/signing, and edge ops (connectivity, uptime, hardware failures across a larger fleet).

So the network gets more efficient on fixed costs but not on the human, relationship-heavy work. The investment question is whether the fixed-cost leverage and demand compounding outrun the linear sales/acquisition/ops cost as the network grows — which they can, but only if demand keeps pace with screens.

The CapEx and payback dimension

Per-screen CapEx (and its payback) is real but secondary to fill. Note two things in diligence: there’s no benchmark hardware cost (a commercial mirror costs very differently from a standalone panel — get the network’s real quotes), and payback is dominated by fill, not hardware price — a cheaper screen that doesn’t sell pays back slower than a pricier one that does. So a low-CapEx story with weak fill is worse than a higher-CapEx story with strong fill.

How to diligence the inputs

The discipline is to interrogate the inputs, because the output (a revenue projection) is only as good as them:

Fill rate — actual, by screen and cohort, and its trend. Stress a low-fill case.
Effective net CPM — realised from transactions, net of the haircut and venue share — not a market CPM.
Foot traffic — measured, per venue type and market tier.
The cost stack — per-screen CapEx (real quotes), fixed costs, and how they amortise.
The split base — is the venue/operator share of gross spend or net after the stack? The base matters as much as the percentage.
Actual paid history — the single most valuable artifact: real paid impressions and dollars per screen, not a calculator.

The takeaway

The unit economics of a beauty DOOH network are a transparent model dominated by one input — fill rate — that most pitches under-emphasise in favour of screen count and a gross CPM. Diligence the inputs, not the output: realised net CPM on sold slots, actual fill and its trend, the cost stack and the split base, backed by real paid history. The model can be genuinely attractive at scale (fixed-cost leverage plus compounding demand), but only if sold demand keeps pace with installed screens. Underwrite the fill, and you’re underwriting the right thing.