Digital Signage ROI: How to Work Out What One Screen Actually Earns
Every proposal for a screen network contains the same slide: a big number, a hockey stick, and the phrase "unlocking new revenue streams." Almost none of them contain the arithmetic. If you own or manage screens — in a hotel lobby, a serviced-apartment block, a clinic waiting room, a shuttle bus — the only question that matters is whether the screen earns more than it costs. That is a calculation you can do on one page, and you should do it before anyone sells you anything.
Start with the only three numbers that matter
Screen revenue reduces to three inputs: how many people see it, how often an impression converts to money, and what you keep after costs. Everything else is decoration.
Daily impressions. Not footfall. Impressions are people who were plausibly in a position to look at the screen for more than a second or two. A lobby with 400 guest movements a day does not produce 400 impressions if the screen is behind the concierge desk at hip height. Walk the space at the hours it is busiest and count honestly. Halve your first guess. That is usually closer.
Yield per thousand impressions (CPM). This is the rate an advertiser pays for a thousand plays in front of an audience. Programmatic digital-out-of-home inventory in mid-tier venues generally clears in the low single-digit dollars per thousand. Premium, tightly-defined audiences — business travellers in a five-star lobby, patients in a private clinic — clear considerably higher because the advertiser is buying context, not volume. Directly-sold inventory to local businesses almost always beats programmatic rates, because you are cutting out three intermediaries who each take a cut.
Fill rate. The share of available ad slots that actually sell. This is where most projections quietly cheat. A new network does not run at 100% fill. It runs at 15% in month one and climbs. If a vendor's model assumes full inventory from day one, the model is marketing, not maths.
The calculation, done properly
Take a single screen in a boutique hotel lobby. Suppose 250 genuine impressions per day. A loop of ten 15-second slots means each slot plays four times an hour; over a 16-hour operating day that is 64 plays per slot. Impressions per slot per day are therefore roughly 250 — every person passing sees the full loop once, on average.
At a $6 CPM and 30% fill across ten slots, monthly revenue per screen is: 250 impressions × 10 slots × 30 days × 0.30 fill = 22,500 paid impressions, at $6 per thousand = $135 per month. One screen. That is the honest number, and it is why single-screen deployments almost never pay back on advertising alone.
Now change one variable. Sell two of those ten slots directly to local businesses — a restaurant, a car service — on a flat monthly retainer of $200 each. Monthly revenue becomes $400 plus roughly $110 of programmatic fill on the remaining inventory. Same screen, same audience, 3.8× the return, because direct sales price the relationship rather than the impression.
The costs people forget
Hardware is the cost everyone budgets for and the one that matters least over a five-year horizon. The costs that actually erode returns are these:
- Content refresh. A screen showing the same six creatives for four months is furniture. Someone has to produce new material, and either you pay them or you do it badly.
- Downtime. A black screen earns nothing and quietly damages the venue's impression of the whole project. Networks without remote monitoring routinely run 5–10% dark without anyone noticing for weeks.
- Sales cost. Direct advertising revenue does not arrive by itself. If you are paying commission or salary to fill inventory, that is 20–35% off the top.
- Revenue share. If the venue is not yours, the venue takes a cut. Standard splits run 30–50% to the property.
- Bandwidth and power. Small per screen, meaningful across fifty.
Why hardware-free deployments change the maths
The reason most screen networks fail is not that the advertising doesn't work. It is that the payback period on media players, mounts, installation and cabling pushes break-even out past two years, and few operators have the patience or the sales pipeline to get there.
Removing the hardware line removes the cliff. If a platform runs on the smart TVs a property already owns, the capital cost approaches zero and the calculation becomes purely operational: does the monthly revenue exceed the monthly software and content cost? At $135 a screen in passive programmatic revenue and no capex, a fifty-screen estate clears roughly $6,750 a month before costs, and it does so from month one rather than month twenty-six. That is the structural argument for software-only deployment, and it is the model behind the AdVision Connect platform.
Non-advertising revenue is usually the bigger number
The ROI conversation fixates on third-party advertising because it is easy to model. In practice the larger return in hospitality and residential settings comes from selling your own inventory to yourself.
A hotel screen promoting the rooftop bar does not need a CPM. It needs an incremental cover count. If the screen drives four extra covers a week at an average spend of $45, that is $720 a month from one property, with no advertiser, no fill rate and no revenue share. The same logic applies to spa bookings, late checkout, airport transfers, and room upgrades. These are high-margin lines, and the screen is a zero-marginal-cost sales channel sitting in front of a captive audience who have already paid to be there.
When you build the model, run two columns: third-party advertising and internal upsell. In most venues we have looked at, the second column is two to five times the first.
A sane payback test
Before signing anything, insist on three things. First, a pilot on a small number of screens with the actual audience — not a case study from a different market. Second, impression measurement you can audit, not a number the vendor reports to itself. Third, a written assumption for fill rate in months one, six and twelve, so that when reality diverges you can see by how much.
If a supplier will not put a ramp curve in writing, the ROI slide is not a forecast. It is a hope with a chart around it.
Work out your own three numbers first. Then let the vendors argue with your arithmetic instead of supplying their own.