A director sits across from a procurement officer with a renewal quote on the table. The handsets still turn on. The content still plays. The vendor is asking for another three years at a number that's gone up 11% since last cycle. The question, said out loud or not, is the same: do we keep limping this thing along, or do we rip it out?
The honest answer in most cases: rip it out. Not because the new options are shinier, but because "still works" is doing a lot of work in that sentence, and what it usually means is "still boots, hardly anybody uses it."
This piece is about how to make that call without either falling for vendor inertia or making an expensive lurch toward the next thing. Not how to migrate — that's covered in how to replace rented audio guide hardware. Whether to.
The sunk-cost trap, named plainly
The single biggest reason museums stay on a tired system is a feeling, not a number. The feeling is: we paid for this. We trained on this. We have a wall of devices in a charging rack and a person whose job partly involves maintaining them. Throwing that away feels wasteful.
It isn't. It's already thrown away. The handsets you bought five years ago are off the books. The training your staff did is part of their general competence and most of it transfers. The charging rack is a piece of furniture. None of that money comes back if you sign the renewal, and none of it costs more if you don't.
The procurement instinct — "we already invested, we should keep using it" — is the exact mistake economists named sunk-cost fallacy. Past spending shouldn't influence forward decisions. The only numbers that matter for the next three years are the numbers for the next three years.
When we've sat with museum finance teams and laid out two columns — keep vs. switch, future cash only, no historical spend included — the decision usually flips. The hardware sitting in the rack stops feeling like an asset and starts looking like what it is: a depreciation line that's already been zeroed out.
What "still working" usually means
"Still working" is a description of the equipment, not the program. The two are different.
We've looked at adoption data across museums running mature hardware systems — the kind that's been in place for five to ten years. The pattern is depressingly consistent. Paid hardware guides at general-admission museums hit somewhere between 4% and 12% of ticketed visitors. The high end is a museum where staff actively upsell the guide at the counter. The low end is a museum where the guide is technically available but nobody mentions it.
A program that touches under 10% of your visitors is not a program. It's a line item. There's more on this in our breakdown of audio guide adoption rates, but the headline is: hardware guides almost never recover from a slow adoption decline. They flatten and stay flat.
So when somebody says the system is still working, ask the second question: working for whom? If the answer is "the 7% of visitors who rented it last quarter," that's not a working program. That's a system that's still capable of working, being used like a wall plaque most people walk past.
A working program has a number behind it. 30% adoption. 45%. Some BYOD deployments get past 60% in the right conditions. Those are working programs. 8% is a system that's running.
The hidden costs of staying
The renewal quote is the cost everybody sees. It's also usually the smallest line in the actual operating cost of an aging hardware program.
Here's what doesn't show up on the invoice:
- Repair and replacement runs at roughly 8–15% of fleet value annually past year three. Screens crack. Batteries swell. Earpieces snap. Charging contacts wear. By year five, you're paying meaningful money to keep a fleet at operating capacity, often through the same vendor who sold you the hardware and prices spares accordingly.
- Content rewrites are quoted as separate projects. Your collection has changed. Three works on the audio tour aren't on the floor anymore. A new gallery has zero coverage. The vendor will happily produce new segments at €800–€2,500 per stop with a 4–8 week turnaround. You either pay it or accept a guide that doesn't match the museum.
- Language lock-in is the quietest cost. Your hardware program covers, say, English, French, German, Spanish, and Mandarin. Your visitor data shows growing Italian, Korean, Portuguese, and Arabic demand. Adding each language means recording, mastering, syncing across the fleet, and quality-checking. Most museums on legacy systems freeze the language list and absorb the lost engagement.
- Counter staffing is real labor. Someone hands the device out, someone takes it back, someone wipes it down, someone troubleshoots, someone reconciles missing units. At a museum doing 200 rentals a day, that's most of an FTE just on the device counter — money that doesn't appear in the audio guide budget at all because it's spread across general admissions staffing.
- Lost adoption is the cost that hurts most and gets named least. Every visitor who walks past the counter because the queue is too long, who tries it and quits because the language they wanted isn't there, who can't be bothered to swap an earpiece — that's a visitor who didn't get the guide. Multiply by your annual visitation. Then ask whether your current program is a $30,000 vendor invoice or a $300,000 missed opportunity.
The renewal quote on the desk is one of those numbers. It's almost never the biggest one.
Signals it's time to switch
Some specifics. None of these alone is decisive. Two or more, and the decision usually makes itself:
- Adoption has fallen for two consecutive years with no operational change to explain it. That's not a blip. That's the program ageing out.
- Repair invoices have crossed half the rental fee in any single year. You're paying twice for the same fleet.
- The vendor has raised the renewal price at or above 8% per cycle. You're being optimized for, not partnered with.
- You can't add a language without a multi-month project and a five-figure quote, and your visitor mix has shifted.
- Your content audit reveals more than 15% of stops out of date — moved works, wrong attribution, dead loans, gallery changes the audio doesn't reflect.
- Your hardware is out of warranty and the vendor has stopped manufacturing replacement units for your model.
- Staff have a workaround culture around the system — informal printed maps, "actually, here's what I'd recommend instead," tours that quietly route around it. Staff route around things they no longer believe in.
- Your accessibility and operations teams are quietly using their phones to do what the official guide should do. Visitors notice.
Three or four of these and the question isn't whether to replace. It's how fast.
Signals to wait
Lean yes is not blanket yes. There are honest reasons to sign one more renewal and revisit in 18 months.
You're inside 18 months of a fresh capex. If you spent serious money on new handsets two years ago, the operational case for ripping them out is weaker. Run them another cycle, get the depreciation closer to zero, plan the replacement deliberately. Don't double-spend in two budget years.
Your contract has a punitive exit clause and renewal is far out. Some legacy contracts make mid-term exits more expensive than the remaining term. Read the contract before deciding. If exiting costs more than running it out, run it out — and start the procurement work now so you're not making this same decision panicked next time.
A larger digital strategy review is in flight. If the museum is six months from a board decision on a CRM, ticketing, app, and audience data overhaul, the audio guide question belongs inside that review, not ahead of it. Replacing audio in isolation now might lock you into integrations you'll regret in a year.
You genuinely don't have the staff bandwidth. Migration is a project. It needs a content owner and an operations owner, and a few weeks of their attention. If both of those people are running an exhibition opening, the right answer is probably September, not now. Don't half-do this.
Outside these four cases, "wait" is mostly a vocabulary for "decide later." That's fine for a quarter. It's expensive for a year.
What "replacement" actually costs now
Most museum operators are working from a price model that's about a decade out of date. The assumption is that replacing an audio guide system is a six-figure capex with another six figures of content work behind it. That used to be roughly right.
It isn't anymore. A modern BYOD platform — visitors use their own phones, you manage content through a web interface, the vendor handles infrastructure — sits in a different economic category entirely. Setup is typically €10,000–€35,000 depending on content scope. Annual cost runs €4,000–€18,000 for most mid-sized museums. There's no fleet to buy, no charging rack to maintain, no per-device repair line.
For a fuller picture of how the numbers actually shake out, our piece on audio guide pricing models walks through the common structures.
The other thing that's changed: content production. Re-recording a tour used to mean booking a studio, paying voice talent per language, and waiting six weeks per round. AI-generated audio collapses that. New language? Hours, not months. Updated stop because a work moved? Edit the script, regenerate, publish. The content workflow on an AI audio guide is closer to updating a website than producing a record. That changes what "keeping content current" costs over a five-year window — the line item that quietly bled the most money on legacy systems.
When you compare the all-in cost of a modern replacement against the all-in cost of one more cycle on the old system — repair, staffing, content rewrites, lost adoption — the gap usually isn't close. We've watched museums save six figures over a three-year window and end up with a guide that more visitors actually use.
The decision, said clearly
The question isn't whether your current system still works. It's whether your current program is working, and whether the cost of one more cycle on it is higher than the cost of starting over.
For most museums sitting on a five-plus-year-old hardware deployment with sub-15% adoption and a renewal quote in front of them: it is. Sign one more contract and you've committed to three more years of the same numbers. Two of those years could have been on a system more visitors actually used.
The systems that are worth keeping have a number behind them. Real adoption. Active content. Languages that match your visitor mix. A vendor relationship that feels like a partnership rather than a meter running. If your current setup has those, keep it. If it doesn't, the renewal quote is the moment to do the work — not the moment to defer it for another cycle.
One practical note on what "starting over" now means. Newer platforms like Musa run on per-interaction or revenue-share pricing, which means the replacement isn't another capex gamble — there's no sunk cost to amortize against, no floor-price minimum, and the guide only costs the museum when visitors actually engage. On a legacy hardware system, every under-adopted year was still fully paid for. On the new model, the opposite is true: if adoption climbs, that's margin; if it doesn't, the museum hasn't overspent. That's what makes switching economically rational now in a way it wasn't five years ago.
If you want a second pair of eyes on whether your current system is worth keeping, we're happy to look at the numbers. Send us your adoption data, your renewal quote, and your last twelve months of repair invoices. We'll tell you honestly whether replacement makes sense for your museum or whether one more cycle is the right call. Get in touch, or read more on what a modern AI museum guide actually does differently.