You're looking at BYOD audio guides. The vendor says visitors use their own phones. Great, no hardware to buy, no devices to charge. Then someone on your team asks the question: "Is our Wi-Fi good enough?"
This is where the conversation often stalls. Museum directors picture a massive networking project. IT teams start spec'ing enterprise-grade access points. Someone mentions a six-figure infrastructure quote. And suddenly the whole audio guide project feels heavier than it needs to be.
The truth is simpler. Digital audio guides aren't bandwidth-hungry applications. You don't need the kind of network that streams 4K video to hundreds of devices. You might not need Wi-Fi at all. What you need depends on your building, your visitor numbers, and the specific guide you're deploying.
Audio guides use less bandwidth than you think
An audio stream runs at about 128kbps. That's kilobits, not megabits. A single Netflix stream in standard definition uses about 25 times more bandwidth.
For a museum with 200 visitors using the audio guide simultaneously (which would be a very high adoption rate for most institutions) the total audio streaming load is around 25 Mbps. A basic business internet connection handles that without breaking a sweat.
AI-powered features use more during active interactions. When a visitor asks a question and the system generates a response, there's a brief exchange of data. But these interactions are sporadic and short-lived. Not every visitor is asking a question at the same moment. The peak load from AI features in a museum of 200 concurrent users might add another 10-15 Mbps in bursts.
Add it up and you're looking at 40 Mbps or so during your busiest hour. That's not nothing, but it's a fraction of what most people assume when they hear "hundreds of phones streaming audio."
The bottleneck, when there is one, is rarely total bandwidth. It's concurrent connections.
Concurrent connections are the real constraint
Most consumer-grade Wi-Fi routers cap out at 30-50 simultaneous devices before performance degrades. Some start struggling at 20. Your museum probably has staff phones, POS systems, security cameras, and laptops already eating into that number. Add a hundred visitors with phones and the router doesn't run out of bandwidth. It runs out of connection slots.
Check this first. Not "how fast is our internet?" but "how many devices can our access points handle at once?"
Enterprise access points typically support 200+ concurrent connections per unit. You don't necessarily need enterprise pricing to get enterprise connection capacity. There are mid-range options designed for exactly this scenario: lots of low-bandwidth devices in a shared space. Hotels, conference centers, and airports solved this problem years ago.
If you're talking to your IT team or a networking vendor, the question to lead with is: "What's our concurrent device capacity during peak visitor hours?" If the answer is vague, pin that down before anything else.
Coverage matters more than speed
A 500 Mbps connection in your server room means nothing if gallery 7 has no signal.
This is the most common Wi-Fi failure in museums. The lobby works fine. The gift shop is great. But the second-floor gallery behind two thick walls and a fire door is a dead zone. And that's exactly where a visitor is standing, trying to hear about the painting in front of them, when the audio cuts out.
Coverage problems are worse in older buildings. Stone walls absorb Wi-Fi signals. Metal structural elements scatter them. Vaulted ceilings create unpredictable reflection patterns. A signal that reaches easily across an open-plan office hits a 16th-century stone wall and dies.
The fix isn't a faster router. It's more access points, positioned where visitors actually are. Repeaters in galleries. Access points in hallways that feed adjacent rooms. The goal is consistent signal across every space a visitor might use the guide, not a powerful signal in one spot.
The simplest test: walk every gallery in your museum with a phone, watching the Wi-Fi signal indicator. Do this during operating hours when visitor phones are competing for bandwidth. Note where the signal drops. Those are your problem areas.
This walk-through costs nothing, takes an hour, and gives you more useful data than any theoretical network plan. Do it before you spend money on anything.
The captive portal problem
You know the screen. You connect to "Museum_Guest_WiFi" and a browser pops up asking you to accept terms, enter an email, or click through a splash page. Captive portals.
They exist for good reasons: legal liability, usage tracking, bandwidth management. But they add friction at exactly the wrong moment. A visitor has just scanned a QR code to start the audio guide. Their phone connects to your guest Wi-Fi. And instead of downloading the tour, they're staring at a terms-of-service page. Some visitors won't complete it. Some won't understand what's happening. Some will assume it's broken and give up.
Every step between "I want the guide" and "the guide is playing" is a point where you lose people. Captive portals are one of those steps.
Options:
- Open guest network: No login, no portal. Simplest for visitors, but your IT team may object to the lack of access controls
- Pre-authenticated network: A dedicated SSID for audio guide users, printed on the QR code signage with the password. No portal, but still access-controlled
- Minimal portal: If you must have one, make it a single tap. "Accept and connect." No email collection, no multi-step process
The right choice depends on your institution's policies. But if you're spending money on a great audio guide and then gating it behind a five-step Wi-Fi login, you're undercutting your own investment.
Historic buildings are hard on radio signals
We've worked with museums in buildings that were built to withstand cannon fire. Those walls do an equally thorough job of blocking 2.4 GHz radio waves.
The challenges are specific and predictable:
- Thick masonry walls (common in museums housed in former palaces, fortresses, and manor houses) attenuate Wi-Fi signals dramatically. A signal that travels 30 meters in open air might not make it through a single stone wall
- Metal structures like wrought iron galleries, steel-reinforced floors, and metal mesh in old plaster create reflections and dead zones
- Basement and underground galleries are often close to unreachable from access points on upper floors
- Listed building restrictions may prevent you from mounting hardware where it would be most effective, or running cables through historic fabric
None of this is unsolvable. It just means you can't treat your 18th-century townhouse museum the same as a modern office building when planning wireless coverage. You need a site survey that accounts for the materials your building is actually made of, not a standard template.
If you're commissioning a network upgrade, insist on a physical site survey with test measurements in every gallery. Any vendor who quotes you based on a floor plan alone is guessing.
The cellular fallback
Many visitors won't use your Wi-Fi at all. They'll use their mobile data.
In a city-center museum with good 4G or 5G coverage, this often works perfectly. The visitor doesn't need to connect to anything. They scan the QR code, the guide loads over cellular, and they're off. No Wi-Fi password, no captive portal, no connection drops when they move between rooms.
Cellular has its own coverage limitations. Thick-walled historic buildings can block cell signal just as effectively as Wi-Fi. Rural museums may have weak coverage from all carriers. But for a modern museum in an urban area, your visitors' data plans may handle the audio guide better than your guest network does.
This isn't a reason to skip Wi-Fi planning entirely. It's a reason not to panic if your Wi-Fi isn't perfect. Between your guest network and visitors' own data, most people will have a connection. The visitors who don't (and there will be some) are the ones who benefit from offline mode.
When the answer is no Wi-Fi at all
Some buildings don't have Wi-Fi and shouldn't get it. A remote heritage site with no power supply. A historic structure where running cables would damage irreplaceable fabric. A venue where the cost of proper wireless coverage exceeds the entire audio guide budget.
For these venues, the answer isn't an expensive network installation. It's an offline-capable audio guide.
At Musa, we're building our offline product around exactly this reality. Visitors scan a single QR code (wherever there's a connection, even just cellular in the car park) and the audio guide downloads to their phone. From that point, everything runs locally. Curated tours work. Suggestions work. Diverging from the tour path works. The visitor never needs to scan another code or touch the internet again.
What doesn't work offline is asking the AI custom questions. That requires a round trip to a language model, which requires a connection. But the guided tour experience, which is what the majority of visitors want, runs fully on the device.
So the Wi-Fi question has a real answer for venues at every point on the spectrum. Good Wi-Fi? Great, visitors get the full experience including custom questions. Patchy Wi-Fi? The core tour still works, and custom questions work in the areas with coverage. No Wi-Fi at all? Offline mode covers the guided experience completely.
Don't over-spec
The temptation with any technology project is to over-engineer it. Especially when IT is involved. Especially when a vendor is writing the spec.
You do not need a full mesh network with centralized management, redundant uplinks, and 802.11ax access points in every room to run an audio guide. You need:
- Enough concurrent connection capacity for your peak visitor count plus your operational devices
- Consistent coverage across the galleries where visitors use the guide. Not blazing speed, just reliable signal
- Sufficient bandwidth for your expected concurrent users, which, as we've covered, is much less than most people assume
For a small museum with 50 visitors at a time, a couple of well-placed mid-range access points might be all you need. For a large institution with hundreds of concurrent users across a complex building, you'll want more, but "more" means more access points for coverage, not necessarily a more expensive network architecture.
The spec that matters is: can every visitor in every gallery maintain a stable connection while the guide is running? If yes, your network is good enough. If not, find the dead spots and fill them.
What to tell your IT vendor
If you're bringing in an outside vendor for Wi-Fi improvements, here's what to communicate:
- The application: Audio streaming and occasional AI text interactions from visitors' personal devices. Low bandwidth per user, high concurrent connection count during peak hours
- Peak load estimate: Your average visitor count during the busiest hour, multiplied by your expected audio guide adoption rate. If 300 people visit during your busiest hour and you expect 15% adoption, plan for 45 concurrent audio guide users (plus all the other devices already on your network)
- Coverage requirement: Every public gallery and circulation space. Not offices, not storage, not the loading dock. Show them the floor plan with visitor areas highlighted
- Building constraints: Material composition of walls, any listed building restrictions on mounting hardware or running cables, locations where power is available for access points
- Captive portal preference: Whether you need one and how minimal it can be
Give them this information and ask for a site-survey-based proposal. Push back on any quote that doesn't include physical signal testing in the actual building.
Start with what you have
The practical first step isn't buying hardware. It's testing.
Walk every gallery with a phone connected to your guest Wi-Fi. Stream music or a podcast, something that approximates audio guide bandwidth. Note where it works, where it stutters, and where it drops. Do this at different times of day, including your busiest periods.
If the coverage is solid everywhere visitors go, you're probably fine. Start a pilot with your audio guide vendor and see how it performs with real visitors. Actual usage data beats theoretical capacity planning every time.
If you find dead spots, you have specific information to act on. Not "we need better Wi-Fi" but "gallery 12 and the ground-floor corridor have no signal, and the east wing is intermittent." That's actionable.
And if you find that your building is a connectivity black hole (thick walls, no cell signal, no budget for wireless infrastructure) that's fine too. It just means you need a guide that works offline. The infrastructure shouldn't dictate whether your visitors get interpretation. The content should reach them regardless.
If you're working through Wi-Fi questions as part of an audio guide project, let's talk through what makes sense for your specific building.