Hardware checklist for a network

The phrase “hardware checklist” promises a shopping list — a column of part numbers you order, install, and forget. It isn’t. Most of the list is decisions, not parts: what kind of panel, where the brain lives, what your software can manage, what you’ll do when one dies on a Saturday. And the single decision that quietly kills more young networks than any other is the cheapest-looking one — buying a consumer TV where the job needed a commercial display. This guide walks the checklist as a sequence of decisions, marks clearly which ones the vendor spec sheets actually let you verify, and — just as clearly — which ones the spec sheets leave to your judgement. Because a checklist that hides its uncertainty is worse than no checklist at all.

1. The one decision that ends networks early: a commercial display, not a TV

Walk into any electronics store and a 43” 4K TV costs a fraction of a commercial signage panel of the same size. The temptation is obvious, and it is the most expensive mistake on this list. The two are not the same product class built to different price points — they are built for different duty cycles.

• Commercial displays are rated for 16/7 or 24/7 continuous operation — they’re engineered to run sustained, non-stop, often in landscape or portrait, with the thermal design and panel grade to match (vendor display guides — directional).
• Consumer TVs are designed for intermittent, evening use — on the order of 6–8 hours a day. Run one as signage and you’re not just risking premature failure; you’re typically operating outside the 1-year consumer warranty, which generally excludes commercial use outright (vendor display guides — directional).

The practitioner translation: the cheap TV isn’t cheaper, it’s unwarranted. The first time one fails after fourteen months of all-day duty, you replace it at full cost, eat the truck-roll, and explain the dark screen to your venue partner. A commercial panel’s longer warranty and duty-cycle rating is the thing you’re actually buying — price the network on commercial panels, full stop, and fold the difference into your start-up cost model.

2. Brightness and orientation: spec for the room, not the box

The second display decision is brightness, and it’s where “I tested it at home and it looked fine” goes wrong. A panel that looks vivid in a dim showroom can wash out in a south-facing salon at 2pm.

• Consumer TVs typically output ~250–350 nits (vendor display guides — directional). That’s adequate for a controlled, dim living room — and marginal in a bright retail or lobby environment with daylight and overhead lighting.
• A bright indoor space wants a brighter commercial panel. The brighter your ambient environment — window-facing waiting areas, glass storefronts, high-lumen salon lighting — the more nits you need to stay legible. The exact target is a per-site judgement; the verified anchor is only that you need more than a consumer TV gives you (vendor display guides — directional).

Two more room-driven choices belong here, and both are routinely supported on commercial panels but not assumable on a TV:

• Orientation. Commercial displays support portrait as well as landscape — which matters because so much beauty inventory (a slim screen beside a station mirror, a vertical point-of-wait display) is portrait (vendor display guides — directional). A consumer TV mounted vertically may run its panel and electronics in an orientation they were never rated for.
• Glare and finish. A high-gloss screen opposite a window becomes a mirror of the wrong kind. Anti-glare finish is a real spec to ask for — though note we did not independently verify anti-glare/anti-fingerprint claims in this pass, so treat finish as a question to put to the vendor, not a settled fact.

The discipline: spec each screen for its actual location. A back-room display and a sunlit window display are not the same purchase.

3. The brain: external media player vs built-in SoC

Every screen needs something that stores content, runs your CMS client, and drives the panel. You have two genuinely different architectures, and this is a real fork, not a detail.

Option A — external media player. A dedicated box (the category leader being BrightSign) cabled to the display. What the vendor docs verify (BrightSign — primary vendor):

• BrightSign offers commercial players in tiers (Series 4/5/6) so you can match horsepower to need.
• They run BrightSignOS, a purpose-built signage OS — not a general-purpose computer you have to harden, patch, and babysit.
• SoC/player tiers support resolution up to 4K.
• Newer series carry a 5-year warranty — with one caveat worth pinning: that 5-year term applies to Series 5 and above, from January 2025, so an older unit won’t necessarily carry it.

Option B — built-in SoC display. A “system-on-chip” commercial display (e.g. Samsung QMC/QHC running Tizen) puts the player inside the panel and removes the external box entirely (Samsung, BrightSign — primary vendor). One device, one power cable, one thing to mount, nothing dangling behind the screen to be unplugged by a curious client.

How to choose, as an operator:

• SoC wins on simplicity and install — fewer parts, fewer cables, fewer points of failure at the wall, cleaner behind a mirror.
• External players win on flexibility and decoupling — you can standardise on one player OS across a fleet of mixed-brand panels, upgrade the brain without replacing the (expensive) display, and swap a dead $X player far more cheaply than a dead screen.
• A purpose-built signage OS is a feature, not a footnote. Whichever path you take, prefer hardware running an OS built for unattended signage over a repurposed tablet or mini-PC — it’s the difference between a device that reboots into your content and one that reboots into a desktop.

Note what we could not verify: specific RAM/storage/codec figures and any claim about Raspberry Pi-class devices as DOOH players. Treat those as open questions for your shortlist, not settled spec.

4. The beauty-specific display: mirror screens vs lobby screens

Beauty is one of the few verticals where the form of the screen is part of the product, not just the panel grade. Two formats dominate, and they sell to different parts of the room.

• Mirror displays (“magic mirror” / mirror-overlay). Screens that present as a mirror until content plays through the glass. These are explicitly marketed for the hair & beauty setting, and mirror-overlay advertising displays for salons are a real, available product class (Pro Display — primary vendor). This is the format that lets you sell the station mirror itself — the surface a seated client looks at for an entire appointment — as captive-audience inventory.
• Lobby / waiting-area screens. Conventional commercial panels (often portrait) in the point-of-wait zone — reception, the waiting bench, the nail bar. Lower install complexity than a mirror, and the natural place to start a network.

Which to buy first is a network-design question we treat in depth in Mirror vs lobby screens — the short version is that mirrors carry the premium dwell story but cost and complexity more, so most operators prove the model on lobby screens before investing in mirror inventory.

5. The plumbing the spec sheets skip: mount, power, connectivity

Here the research turns honest in a way most “hardware checklists” don’t. We found no source-verified figures for mounting, cabling, power draw, or bandwidth in this pass — so rather than invent numbers, here is the decision framework to take to a qualified installer and your own measurements. Treat everything in this section as judgement, not data.

• Mount & enclosure. Decide commercial wall mount vs. mirror-integrated housing per site; in a public salon, plan for an anti-theft / anti-tamper approach and cable concealment — both for safety and because exposed cabling is the fastest way to look unprofessional in a premium venue. Get the venue’s landlord/fit-out constraints in writing before you drill.
• Power. Confirm a suitable outlet within reach of each screen, who pays for the electricity (a clause for your venue agreement), and whether the install needs an electrician rather than a plug. Don’t assume there’s an outlet where you want the screen.
• Connectivity. You need a reliable path for content delivery and proof-of-play reporting. Wired ethernet is the most stable; salon Wi-Fi is convenient but shared with staff and clients; cellular (4G/5G) is the common failover so a screen stays live and keeps logging plays when the venue’s internet drops. Which combination is right is a per-site call — but plan for failover, because a screen that can’t phone home can’t prove it played the ad you billed for.

The reason to flag all of this rather than quote it: the figures that float around for these items come from installer marketing, and we didn’t subject them to verification. Anyone handing you a confident “$X to mount, Y Mbps per screen” number should be asked where it came from.

6. Cameras and sensors: what they do — and the claim to distrust

If you intend to measure your audience with computer vision rather than buy third-party data, you’ll add a camera with on-device face detection — the basis of Anonymous Video Analytics (AVA), the standard approach to camera-based audience measurement for out-of-home (IAB, arXiv — primary). AVA detects faces to estimate viewers, attention, and coarse attributes on the device.

Now the part that earns this guide its method note. A widespread vendor and field-trial claim is that AVA “records no images or video and collects no personally identifiable information, logging only aggregate statistics.” We put that claim through three independent adversarial verifiers and it was refuted unanimously (0–3) — so we are not asserting it. The honest position: AVA systems process camera imagery to detect faces (that’s how they work), and the blanket “no PII, nothing recorded” reassurance does not survive scrutiny. For a beauty operator this matters twice over: your venues’ clients are in an intimate setting, and privacy regulation around facial processing is real. If you deploy cameras, get specific written answers on what is processed, what is stored, and what notice/consent is required in your jurisdiction — and do not repeat the “anonymous, no PII” line to venues as if it were settled fact.

Ambient-light and occupancy sensors are a lighter-touch alternative (auto-brightness, basic presence) — but we found no verified specifics on them here, so they’re noted, not specified.

7. Don’t buy hardware your software can’t manage

A checklist that stops at the screen forgets that you’ll run dozens of them remotely. Two hardware-adjacent decisions belong on it:

• CMS-agnostic vs. locked. Hardware that only works with one vendor’s CMS ties your whole network’s fate to that vendor’s pricing and roadmap. Favour hardware that speaks open or widely-supported protocols so your content management and your panels can evolve independently — the same decoupling logic that favours external players in §3.
• Remote management & spares. You cannot drive to every salon for every reboot. Prefer devices with real remote management (status, remote content update, reboot, health alerts), and decide your spares strategy up front — how many swap units you hold, and whether a failure means a same-day swap or a week-long gap of dark, unbilled inventory. We found no verified MTBF figures to quote, so size your spares from your own observed failure rate, not a datasheet promise.

So — what do you actually buy?

Not a part number. A short stack of decisions, in order:

• Commercial displays, never consumer TVs — rated for 16/7–24/7, warranted for commercial use, bright enough for that room, in the orientation the spot needs.
• A brain you can manage — an external media player on a purpose-built signage OS, or a built-in SoC display that removes the box; either way, something built to run unattended and update remotely.
• The right beauty format per zone — mirror displays for the premium captive station inventory, lobby/point-of-wait panels to prove the model first.
• Plumbing planned, not assumed — secure mount, real power, and connectivity with failover so every screen can deliver content and prove its plays.
• Sensors only with eyes open — if you add cameras, treat the “anonymous, no PII” claim as unproven and get the privacy answers in writing.
• Software-agnostic hardware and a spares plan — so no single vendor, and no single failure, can take your network dark.

Get this stack right and the cost and unit economics have a foundation that won’t fail in month fourteen. Get the first decision wrong — a TV where you needed a commercial panel — and no amount of good pricing saves you from the truck-rolls.