Connectivity & uptime

Connectivity is the least glamorous part of a beauty DOOH network and one of the most consequential. A screen that drops offline doesn’t merely miss the slots it was offline for — it can fall below the quality threshold programmatic demand requires and stop getting bids even after it’s back. Uptime, in other words, is a revenue precondition, not an IT detail. This guide is why, and how to engineer for it.

Uptime is a demand gate, not just delivery

The instinct is to treat uptime as “the screen should be on so it shows ads.” True, but it undersells the stakes. In programmatic DOOH, uptime is a quality gate: screens generally must hold above the ~95% uptime threshold that most DSP quality configurations use, or they stop attracting bids (directional). So a screen that drops offline takes a double hit:

It misses the slots it’s offline for — direct lost revenue.
It can fall below the quality bar — and stop getting programmatic demand even after it reconnects, until it rebuilds standing.

That second effect is the one operators miss. A flaky screen isn’t just occasionally dark; it’s structurally less valuable, because demand routes around unreliable inventory. Uptime is therefore a precondition for the fill rate the whole business depends on.

And it churns the advertiser you worked hardest to win

There’s a demand-side cost too. A screen that drops mid-campaign under-delivers — the advertiser paid for impressions that didn’t run. For a young network, that’s the fastest way to lose the first advertiser you fought to land, because under-delivery breaks trust and shows up in the wrap report. So uptime protects both the programmatic standing and the direct relationships — it’s foundational to revenue from both channels.

Salons aren’t data centres — design for that

The engineering challenge is that beauty venues are not controlled environments. You’re deploying into salons with consumer-grade internet, power that blips, a non-technical host who might unplug a router, and a screen that has to “just work” for years. Design accordingly:

Don’t depend on the venue’s network. A salon’s Wi-Fi password changes, the router gets rebooted, the connection is shared with card readers and streaming. Relying on it is relying on something you don’t control.
Expect power and network interruptions and build for graceful recovery — the player should reconnect and resume automatically, cache content locally to keep playing through a brief outage, and resync proof of play when it’s back.
Make it host-proof — minimal exposed cables, nothing the host needs to manage, clear “if it’s off, call us” support.

Cellular as the pragmatic default

For a network of independent venues, cellular (4G/5G) connectivity is often the pragmatic default, precisely because it’s independent of the venue’s network — the operator controls it, it’s consistent across sites, and it sidesteps the salon-Wi-Fi problem. The trade-off is data cost and signal variance, so:

Use cellular as the primary path where venue Wi-Fi is unreliable, with the venue’s fixed line or Wi-Fi as a backup, not the primary.
Cache content locally so the screen keeps playing through short cellular drops, and proof of play queues and uploads on reconnection.
Right-size the data plan — DOOH content is mostly cached and updated periodically, not streamed continuously, so data use is manageable if the architecture caches well.

(There’s no single right answer — venue type and location decide it — but the principle holds: control the connection rather than borrow it.)

You can’t hit uptime you can’t see

Uptime targets are meaningless without visibility. Pair connectivity with monitoring and alerting (the next guide):

Heartbeat monitoring — every player reports in; a missed heartbeat triggers an alert before the host (or an advertiser) notices.
Proof-of-display telemetry — confirm the screen is on and showing content, not just that the player is powered.
Fast response — a known-fast process for a dark screen (remote reboot, host call, dispatch) so outages are minutes, not days.

The networks that hit >95% uptime aren’t lucky — they instrument it, alert on it, and respond fast. (Edge ops at scale is one of the things that doesn’t get easier with size.)

The takeaway

Connectivity and uptime are a revenue precondition, not an IT afterthought. A screen below the ~95% uptime gate doesn’t just miss plays — it loses programmatic standing and churns direct advertisers through under-delivery. So engineer for a hostile environment: don’t depend on the salon’s network (cellular-first is often pragmatic), cache content and proof of play locally for graceful recovery, make it host-proof, and monitor every player so you can hit the uptime that demand requires. Reliable screens are biddable, sellable screens; flaky ones are a liability dressed as inventory.