Technology & Implementation16 min read
The audio guide decision is really about architecture. You're choosing between two fundamentally different models: devices you own and manage, or software that runs on phones visitors already have. The choice ripples through your budget, your operations, and what you can actually measure about visitor behavior.
Let's be direct: for most museums, phone-based software wins. But there are real exceptions. Understanding both sides means understanding your own constraints.
When you commit to dedicated audio devices, you're buying into a full hardware business. That sounds obvious, but the implications aren't always clear at the decision stage.
You purchase devices upfront—real capital expenditure. A typical setup ranges from €50 to €300 per device depending on sophistication. For a museum guiding 500 visitors daily across 350 operating days, that's a 175,000-device-year footprint. Even at lower end hardware, you're talking €8.75M in capex before you deploy a single tour.
Then the devices age. They get dropped. Batteries degrade. Screens crack. The electronics that cost €150 today cost €200 to repair in three years when the manufacturer has moved on. You need warehousing, insurance on inventory, staff trained in device management and maintenance. You establish a returns process, a cleaning protocol, a charging system. You dedicate roughly one person per 1,000 devices in your active base.
Lost devices are expensive, not just in replacement cost but in operational friction. A visitor loses a €150 device, and now you're tracking serial numbers, investigating claims, potentially chasing them for payment. Some devices will walk out. You account for 8-12% annual attrition as normal.
Scaling geographically gets expensive fast. If you operate five sites, you maintain five device inventories, five sets of charging infrastructure, five trained teams. Moving 500 devices between locations for seasonal adjustment is logistics. It's trucks. It's coordination.
The upside: visitors with high friction toward phones (elderly demographics, visitors with language barriers, people intimidated by touchscreens) often find dedicated devices more accessible. The interface is designed specifically for one task. No notifications interrupt. The experience is bounded and controlled.
Phone-based guides are fundamentally different economics. There's no device inventory. Visitors bring their own hardware—a device they already know, already trust, already carry.
Your infrastructure is pure software. You deploy to app stores or publish a web app. There's no manufacturing, no returns logistics, no device graveyards. Your cost structure is operational: cloud hosting, API bandwidth, development, support. It scales linearly. 500 visitors or 50,000 visitors, your per-device cost trends downward as fixed costs spread.
The tech is simpler on your end. A single codebase runs everywhere. You fix a bug once. You deploy once. You don't debug why device A crashes while device B works on the same content.
Analytics are orders of magnitude better. You see exactly where visitors pause, which artworks get skipped, drop-off points in tours, how long people spend on each section. You see it in real time. Device-based systems? They show sync usage after visitors return devices, buried in spreadsheets. The difference between real-time visitor insights and batch reports, a week later, changes how you iterate.
The visitor data is cleaner. You see which exhibitions attract repeat visits, which demographics spend most time in which galleries. Hardware systems give you rough numbers—devices checked out, returned at this time. Software systems show you the actual behavior inside.
Internationalization becomes native. Your platform supports 40+ languages? That's a software feature. Hardware systems require multiple physical devices, multiple languages, or devices that switch languages via menu. Or you deal with inventory hell, managing French devices separately from German devices.
The downside is dependency on visitor connectivity. A visitor needs to scan a QR code or visit a URL. They need a smartphone, even if it's an older one. For populations without phones—a real constraint in some geographies or for some demographics—the software model creates barriers.
Let's compare a 100-visit-per-day museum (modest size, open 350 days annually) rolling out guides.
Hardware path (five-year horizon):
Software path (five-year horizon):
The hardware path looks cheaper if you stop counting at deployment. But if a device fails at year three, you've underestimated replacement costs. If your team has to manage devices manually (common), you've underestimated labor. If you want to change tour content, you're manually syncing devices—that hidden cost isn't in the budget.
The software path front-loads development but locks in marginal costs. By year ten, the gap widens enormously. Hardware device replacement compounds. Software costs grow with features, not with visitor volume.
For larger museums (500+ daily visitors), the hardware costs become prohibitive. Inventory alone becomes a significant operational burden. Software scales indefinitely with nearly flat marginal cost.
Hardware devices appeal to a particular visitor: older, less comfortable with phones, or visiting from a country where phone-based services are unreliable. This is real.
But modern phone-based guides can offer richer experiences. A good software guide -- particularly an AI audio guide -- uses the phone's sensors for positioning and orientation to know where the visitor stands and what direction they're facing. It shows contextual information. It can push information to their pocket. They can resume later. They can share favorite sections. They can see a map.
Hardware devices are locked. You step through slides. Next, next, next. The interface is identical for the Rembrandt and the neolithic pottery. It's simple, which some visitors love. But it's rigid.
Phone guides, if designed well, adapt. A 40-minute tour compresses to 15 minutes if the visitor moves quickly through rooms. Multilingual navigation is a button press. You can segment by age, interest, duration. The visitor controls flow.
The quality difference depends entirely on implementation. A poorly designed app is worse than a well-designed device. But a well-designed app (addressing visitor needs, not engineer convenience) beats dedicated hardware on flexibility and insight.
Accessibility is worth isolating: phone-based guides actually serve accessibility better when built thoughtfully. Voice control, text sizing, language selection, captions—these are software features that cost nothing to add. Hardware devices? Voice control is rare. They come in single language variants. Text sizing is constrained by physical buttons.
Hardware audio guides aren't dead. They make sense in specific contexts:
Very elderly populations. If your core audience skews 75+, and a significant portion has zero smartphone habit, dedicated devices remove friction. The trade-off is worth accepting the operational burden.
High-security environments. Some governments, military sites, or sensitive cultural institutions restrict what phones visitors can bring. You provide hardware. You control what's on it. You audit it. The security requirements justify the expense.
Ultra-thin connectivity. Rare, but real: some remote sites genuinely lack reliable cellular or WiFi. A device with preloaded content works. A phone app doesn't.
Exclusive positioning. A few ultra-luxury museums deploy high-end devices as part of the experience design. The device becomes a souvenir, a design artifact. Cost is absorbed into admission pricing. This is aesthetics, not practicality.
For everyone else: software wins on cost, flexibility, analytics, and maintainability.
Here's where the decision gets existential: what you can learn.
A hardware system tells you utilization (devices checked out per day) and rough demographics if you ask at the desk. That's all. You don't know if visitors actually listened. Which sections did they skip? Where did they pause? Did anyone reach the end of the audio trail? You'll get broad behavioral data a week after collection, if you sync devices and export CSVs.
A software system shows you everything in real time. Heat maps of where visitors spend time. Sections with zero engagement (remove them). Sections that stop visitors cold (redesign). You see which demographics prefer which content. You see repeat visitation patterns. You can A/B test a new description and measure its impact within days.
This data advantage isn't superficial. It's how you iterate. Museums that can measure learn fast. They optimize tours quarterly. They kill content that fails. They double down on what works. Hardware systems lock you into guessing.
If you're serious about using technology to improve visitor experience—not just check a box, but actually measure and evolve—you need software analytics. Hardware can't give you that.
The hardware path means sourcing, shipping, warehousing, distributing, collecting, cleaning, charging, and eventually recycling devices. It's supply chain work. It's real friction in daily operations.
The software path means building, deploying, supporting, and iterating on code. It's engineering work. It scales with your capabilities as an organization.
Which fits your operation? If you have strong technical staff, software is easier. If you have strong logistics and physical operations teams, hardware is less foreign. Most museums are stronger at neither and should default to the model with fewer moving physical parts.
The hidden cost of hardware is context switching: your curatorial team thinks about exhibitions, your ops team thinks about device inventory. Your developer time is spent on device quirks, not on tour quality. Software lets discipline stay clean: curators curate, engineers engineer.
Large urban museum (1,000+ daily visitors): Software only. Hardware inventory becomes unmanageable. Costs explode. Analytics justify development investment.
Mid-size cultural institution (200-400 daily): Software is likely right. If a significant portion of visitors are 70+ and uncomfortable with phones, consider a hybrid: software for the core, a small fleet of loaner devices for accessibility.
Small specialized museum, heritage site: Depends on visitor demographic. If visitors are young and mobile-native, software. If visitors are elderly or disconnected from tech, hardware. Or hybrid.
International sites with poor connectivity: Hardware wins. Preload content. Control the experience.
Organizations with strong operations teams and limited development capacity: Hardware might be more practical in the short term, but you pay for it later in opportunity cost and inflexibility.
The decision really isn't about hardware vs software as abstract categories. It's about what your visitors actually are, what your team can maintain, and whether you want to measure and evolve or deploy and hope.
Can I migrate from hardware to software later? Yes, but you lose existing investment. Device inventory has no residual value—hardware depreciates fast. The time your team invested in device management doesn't transfer. You don't lose content, though; tours designed for one system can be adapted for another. Plan the transition as a sunk cost if you know it's coming.
What about accessibility for visitors without phones? Hybrid works: deploy software as your primary system, but maintain a small loaner device fleet for accessibility. It gives you the best of both worlds without the burden of managing 500 devices. You might loan 5-10 devices per day to visitors without phones, turning accessibility into a staffed feature rather than a hardware requirement.
If I go software, do I need a native app? No. A responsive web app (PWA) runs on any phone with a browser, eliminates app store review cycles, and reduces distribution friction. Native apps are better if you need offline content or device integration (Bluetooth audio triggers, for instance). For most tours, a web app works. Musa builds both web and native routes depending on your needs.
How do I measure ROI on audio guides? Software systems give you real metrics: unique visitors per day, time spent in each gallery, tour completion rates, repeat visitors. You can compare engagement before and after guides. With hardware, this measurement is nearly impossible—you'll get device checkout counts, but not visitor engagement. If ROI measurement matters to your board, software is the only practical choice.
audience: b2b
For most museums, the software question answers itself once you look at economics and scale. The hardware model made sense when phones were unreliable and apps were novel. Neither is true anymore.
If you're evaluating systems, the real question isn't which technology is "better" in theory. It's which one lets you focus on what you do best—delivering experiences that matter—without building an equipment logistics company.
At Musa, we've built specifically for the software model: no devices to manage, multilingual out of the box, real-time analytics, full visitor control. If you'd like to talk through your specific situation—whether software makes sense, how to handle accessibility constraints, or how to measure impact—we're here. Get in touch.