A facilities director at a mid-sized European museum forwarded me a one-page letter last autumn. It was a last-buy notice from their handset vendor. The model they'd been running for six years was being discontinued. They had until the end of the following fiscal year to place a final order, after which spare parts would be available for 24 months and then that was it. No firmware updates beyond the next minor release. No replacements beyond the last-buy window.
She had a budget request to file in six weeks. The question in her email was short: do we renew, or do we stop?
This is the article I wish I'd had to send her. It's for directors of facilities, sustainability officers, and procurement teams staring at the same letter, or about to. Not a broad sustainability pitch, and not a migration how-to. Just a clear read on what end-of-life actually means, what the regulations require, and how to think about the decision that's sitting on your desk.
What "end of life" actually means
Vendors use the phrase loosely, and it matters which version you're getting.
Last-buy notice. You have a fixed window to place a final order at current pricing. After the window closes, you can't buy new units of this model at all. This is almost always the first signal.
Spare parts cutoff. Batteries, screens, lanyards, replacement casings — available for a stated period after last-buy, typically 18 to 36 months. After that, repair is your problem.
Firmware support cutoff. Security patches and OS updates stop. The devices keep working, but they stop receiving fixes. If your content delivery depends on a vendor-hosted CMS, the cutoff can functionally brick the fleet the day the vendor sunsets the backend.
Service contract expiry. The managed-service agreement — break-fix, battery swaps, seasonal logistics — either rolls over or doesn't. Some vendors will keep servicing discontinued models; others tie service continuation to signing for the replacement product.
Read the letter carefully. The last-buy date is what procurement latches onto. The firmware and CMS cutoffs are usually more dangerous, because they can strand perfectly functional hardware in a year where you weren't planning a capital project.
The realistic horizon
Five to seven years. That's the working life of a handset fleet in daily use.
The batteries set the ceiling. Lithium-ion cells degrade after roughly 500 full charge cycles, and a device that gets charged every night hits that inside three years. Fleets compensate by cycling units and replacing batteries, but after the second battery swap the economics stop making sense. Screens crack. Headphone jacks fail. Lanyard clips snap. By year five, the attrition rate on a fleet of 80 units is usually 15 to 25% annually — you're running a miniature logistics operation just to keep the fleet sized for peak season.
If your fleet was installed in 2019 or 2020, you're inside the window now. If you renewed during the post-pandemic catch-up in 2022, you have maybe three years before the same conversation lands on your desk. Plan accordingly.
The compounding risk of renewing
Here's the part most facilities teams don't see until they're a year into a renewal.
You sign the new contract. New hardware arrives. New firmware, new CMS, new staff training. The vendor quietly raises the service fee on year-three renewal. Visitor numbers flatten or drop — which they have, in real terms, at a lot of institutions post-2020 — and you're locked into a capex depreciation schedule that assumed a busier museum than the one you have.
Meanwhile, a fraction of your visitors are still using the devices. The rest are on their phones, squinting at QR codes you put up during a pilot three years ago that nobody finished rolling out. Your staff spends time managing a fleet that's serving a shrinking share of visits. The cost per actual use climbs every year.
Then year six arrives. The model gets a last-buy notice. You're here again.
A renewal at end-of-life isn't a neutral decision. It's a five- to seven-year commitment to the same operating model, at a moment when the vendor has already told you the model is changing. If visitor behavior keeps shifting toward phones — and it will — the renewal locks in a sunk cost that will make the next transition harder, not easier.
Responsible disposal — what the regulations actually require
The handsets don't just vanish when you're done. Disposal is regulated, and the rules have teeth.
EU: WEEE Directive (2012/19/EU). Audio guide handsets fall squarely under Category 6 (small IT and telecommunications equipment). Producers are responsible for takeback and recycling, but as the end user you're required to hand the fleet over to a registered WEEE processor, not a general waste contractor. Most EU member states also require you to keep documentation — transfer notes, weights, processor registration numbers — for audit. If your institution has public funding, expect these records to show up in sustainability reporting.
UK: WEEE Regulations 2013. Functionally similar to the EU directive post-Brexit. Use an Approved Authorised Treatment Facility (AATF). The Environment Agency publishes the register. Your vendor's takeback scheme usually counts, but confirm they're actually processing, not warehousing.
US: state-level patchwork. No federal equivalent. 25+ states have e-waste laws, with California, New York, Washington, and Illinois among the strictest. If your museum operates across state lines, the rules of the most restrictive state are usually the safe baseline. The EPA's SMM program is guidance, not enforcement — the enforcement is at the state level and varies wildly.
Lithium batteries — everywhere. This is the part that trips institutions up. Devices with embedded lithium-ion batteries can't go through the general electronics waste stream in most jurisdictions. They have to be removed by the processor (or handled as a lithium-battery waste stream from the start) because damaged cells are a fire hazard. If you're storing the retired fleet in a back-of-house room while you figure out disposal, store it away from combustibles and document the location. Several museum fires in the last decade have started in storage rooms full of ageing electronics.
Get the certificates of destruction. File them with your fixed-asset register. If a trustee asks three years from now where the 2020 handset fleet went, you want a one-line answer.
The circular option
Before the processor gets them, consider whether the fleet has a second life.
Smaller museums, historical societies, and regional heritage sites often operate on budgets that can't touch new handset purchases. A donated fleet with two or three years of service left is a genuine gift — especially if it comes with the CMS access to keep running the content. I've seen fleets move from a national museum to a cluster of regional sites and run for another four years. The receiving institution handles eventual disposal.
Refurbishment programmes exist too. Some vendors run them directly; independent refurbishers service the major brands. Replace the batteries, swap the lanyards, reset the firmware, resell or redeploy. It's not free — refurb costs 30-50% of new — but it's genuinely circular, and for institutions that want to keep hardware for operational reasons (language coverage, accessibility requirements) it's the cleanest path.
The one thing not to do: a quiet skip run. It's illegal under WEEE and EPA rules, it shows up in sustainability audits, and it wastes the residual value of hardware that someone else could use.
For a broader view of the environmental tradeoffs across the whole fleet lifecycle, see our writeup on the sustainability of museum audio guides.
The TCO comparison that actually matters
Most procurement decks at end-of-life frame the choice as "new fleet price vs. current fleet price." That's the wrong comparison.
The real comparison is between a five-year renewal TCO and a five-year BYOD TCO. On the renewal side, you're adding up:
- Capex for the new fleet (80-150 units at $300-600 per unit, depending on spec)
- Annual service and CMS fees (typically $15-40k)
- Battery and parts replacement through years three to five
- Staff time on fleet management (one FTE-equivalent at most mid-sized institutions, whether you see it on an org chart or not)
- Charging infrastructure energy and replacement
- Disposal cost at year six
On the BYOD side, you're looking at software licensing, WiFi infrastructure (usually already needed for other reasons), and content migration work. The full worked example lives in our audio guide total cost of ownership breakdown, but the short version: across a five-year horizon, BYOD typically lands at 40-70% of a hardware renewal's TCO, and the delta widens every year you extend the comparison.
There's a deeper issue underneath the numbers. Hardware renewal is a bet on a specific visitor volume. You buy 100 units because you expect roughly that many concurrent users at peak. If the peak doesn't materialise, the unused capacity is paid-for anyway. BYOD scales with actual use. You're not guessing.
This is the same calculation that's driven deferred maintenance across museum operations — capex commitments sized for yesterday's assumptions, still draining budgets today. End-of-life is the one moment in the cycle where you get to revise the assumptions without a penalty.
My position
If you're inside 18 months of end-of-life, don't reinvest. Budget the transition now.
That's blunt, and it doesn't cover every case. If you have a specific regulatory requirement for dedicated devices (some national museums do, under accessibility law), or you're in a venue with genuine connectivity problems that won't be solved in the transition window, a tactical refurb to extend the current fleet by two or three years is reasonable. But "we've always used handsets" isn't a reason, and neither is "the new model from the same vendor looks fine." The renewal letter is the universe telling you to reconsider the operating model, not just the inventory.
For institutions already thinking about the transition, the dedicated device vs. BYOD comparison walks through the functional tradeoffs, and replacing rented audio guide hardware covers the operational migration path in detail.
Why end-of-life is the moment the model breaks
Capex audio guide programmes only make sense when you can amortise the fleet against a predictable visitor volume over five to seven years. That's the implicit contract. You commit the capital; the fleet pays it back in visits.
Two things have broken that contract. Visitor volumes are less predictable than they were a decade ago, and the share of visitors who want to use their own phone climbs every year. Both trends make the amortisation math fragile. You buy for the peak that might not come, and you carry the capex whether the fleet gets used or not.
End-of-life is the exact moment where the capex model stops making sense. You've extracted the value out of the current fleet. The sunk cost is spent. The next decision is genuinely fresh — and the choice is between signing another five-year hardware commitment or moving to an operating model that doesn't require guessing future visitor numbers.
An opex model — revenue-share, per-interaction, or simple subscription that scales with use — removes the bet. You pay for the visits you actually deliver. If visitor numbers drop, your costs drop. If they grow, you add capacity without a procurement cycle. The sunk-cost risk that pushed you into the last renewal isn't there, because there's nothing to sink.
This is the framing Musa was built around. The end-of-life window is where it lands most cleanly — the fleet is retiring, the capex is finished, and the next commitment hasn't been made yet. That's the opening.
Whatever path you take, the letter on your desk is a real deadline. Don't let it become an automatic renewal by inertia. If you want to talk through the numbers for your specific fleet size and visitor profile, we're happy to work through it with you.