A director I spoke with last month had 180,000 euros to spend on visitor experience this year. She knew she needed two things: an audio guide and a signage refresh. The signage had been installed in 1998, in a typeface no one would choose today, with a numbering system that referenced a gallery layout that hadn't existed since 2011. The museum had no audio guide at all. Visitors got a paper map that was photocopied so many times the floor plan looked like a Rorschach test.
She asked me which one to do first. I told her the honest answer is more interesting than either of the obvious ones.
Why this is actually a hard sequencing question
The intuitive answer goes: you have to be able to find things before you can hear about them. Wayfinding is foundational. Audio is decoration. Fix the foundation first.
That logic is clean, and it's wrong about half the time. It assumes the two investments are independent — that signage solves wayfinding and audio solves interpretation, and they don't overlap. In 2026, they overlap a lot. A modern AI audio guide can route a visitor turn-by-turn through a building, answer "what's near me," and surface the next stop on a tour without the visitor ever looking up at a sign. Sometimes the audio guide IS the wayfinding system, just delivered through earbuds instead of vinyl on a wall.
The intuitive answer also ignores cost asymmetry. Signage hasn't gotten cheaper in twenty years. Vinyl is still vinyl, fabrication labour is still labour, ADA-compliant heights and contrast ratios still apply. AI audio guides have collapsed in price over the last five years. A project that would have cost 200,000 euros in 2020 can be deployed for under 30,000 today, and the marginal cost of adding a new language or tour is effectively zero. So the trade-off isn't symmetrical anymore. You're not choosing between two equally expensive options. You're choosing between something that costs what it always cost and something that costs a tenth of what it used to.
That changes the sequencing math in a way most boards haven't caught up to.
The case for fixing wayfinding first
There are real situations where signage genuinely belongs first, and I want to be specific about them rather than hand-wave.
Architecturally complex buildings where visitors are visibly distressed. If your floor staff spend more than 20% of their time giving directions, if you regularly find lost visitors crying near the medieval wing, if your front-of-house team has a running joke about "the people who never find the second floor" — fix the signage. This isn't a comfort issue. It's a safety and dignity issue. An audio guide can route the people who download it. Signage routes everyone, including the 70-year-old visitor who isn't going to install anything on her phone.
Signage that's actively misleading. A wrong sign is worse than no sign. If you have arrows pointing to galleries that closed in 2015, room numbers that don't match your current map, or directional language in a language no one on staff speaks anymore — the existing signage is doing damage. An audio guide layered on top of broken signage creates a confused two-channel experience. Strip the bad signage even if you don't replace it with a full system.
Buildings where most visitors won't use any guide at all. Some museums — small regional collections, free municipal museums, family-oriented sites — have audio guide adoption rates under 10%. If 90% of your visitors are navigating purely on physical cues, you can't route them through audio. They need real signs.
In all three of these cases, signage isn't competing with audio. It's solving a problem audio can't reach.
The case for audio first — including the "guide is the wayfinding" angle
For most mid-sized museums in 2026, audio belongs first. Here's why.
The biggest shift in the last few years is that a modern AI guide doesn't just narrate. It routes. A visitor opens the app, picks a 90-minute tour, and the guide tells them where to stand, when to move, and what to skip if they're running short on time. They never need to consult a wall sign for "this way to the impressionists." The guide handles it. We've covered the mechanics of this in detail in how audio guides work as navigation tools at large museums, but the operational consequence is the part directors miss: one investment now produces two outputs. You're funding interpretation AND wayfinding in the same line item.
The cost story compounds this. A signage refresh is one-and-done capital expenditure. You spend 200,000 euros, you get signs, the signs sit there until they're dated again in fifteen years. An AI audio guide is a much smaller upfront spend with ongoing content flexibility. New exhibition opens? Update the guide in an afternoon. Want to add Mandarin? It's a configuration change, not a fabrication job. Want to test a kids' tour? Ship it next week, kill it if it doesn't work.
There's a softer argument too, which is that audio guides generate data. Signage doesn't. When you put up a directional sign, you have no idea whether visitors followed it, ignored it, or got confused by it. When you ship an audio tour, you can see exactly which stops people complete, where they drop off, which questions they ask. That data is a flywheel — it tells you what to fix next, including what signage problems are real and which ones you imagined. We get into this further in our piece on using audio guide data to run museum operations.
The honest version of the audio-first case is this: in 2026, the dollar you'd spend on signage probably has a better near-term return as audio. Not because signage doesn't matter, but because audio has gotten so much cheaper that the comparison isn't what it was when your last facilities consultant gave you a quote.
The hybrid sequencing most museums should actually do
The cleanest answer isn't "audio first" or "signage first." It's "do a basic signage triage AND an audio guide together, in the same budget cycle."
Signage triage is not a full signage refresh. It's spending maybe 15,000 to 25,000 euros to remove actively wrong signs, fix the three or four worst directional problems, and install clear identifiers at gallery entrances. No design firm, no master plan, no consultant. Just: stop the bleeding. Then put the rest of the budget into an audio guide that handles the rest of the wayfinding load through turn-by-turn audio.
This works because it acknowledges what signage is actually for in a museum with an audio guide. Signage becomes the floor — the fallback layer that catches people who don't have the app, who turned it off, or who need orientation in a moment when they don't want a voice in their ear. The audio guide becomes the ceiling — the rich layer that handles routing, context, and interpretation for the people who opt in.
Two birds. One budget cycle.
The trap to avoid is letting a signage consultant scope a full refresh before you've considered whether you need one. Signage firms quote what they always quote — the comprehensive system, the master plan, the family of typefaces. That deliverable made sense when audio guides cost as much as signage. It often doesn't now. If you're going to commission a signage project, scope it after you've decided what your audio guide will carry, not before.
What other investments belong ahead of either
Here's the part most boards don't want to hear. Sometimes neither audio nor signage should be your next dollar.
Visitor data infrastructure. If you can't tell me your average dwell time, your repeat visit rate, or which galleries get the most foot traffic, you can't measure the return on any experience investment. You'll spend 100,000 euros on something and have no defensible answer to "did it work?" eighteen months later. A basic ticketing system upgrade, a visitor counting setup, and a feedback mechanism — even a paper survey — should come before either signage or audio at most museums that don't already have them. Without that, you're flying blind on every experience decision for the next decade.
Statutory accessibility work. This isn't a strategic choice. If your building has compliance gaps — accessible routes, hearing loops, large-print materials, captions for video content — those have to come first. Funders increasingly require it as a condition of grant approval. Boards and trustees carry liability for it. And from a values standpoint, building a beautiful new audio guide while disabled visitors can't access half your galleries is a hard look.
The basics of revenue and dwell time. A working café and a decent shop generate more revenue per visitor than most experience improvements, and they extend dwell time, which improves every other metric. If your café closes at 2pm and your shop sells postcards from 2008, fix that before you fund a guide. We get into the underlying logic of this kind of board-level prioritization in our business case template for an audio guide, but the principle is the same: don't fund the thing that sounds exciting if a more boring investment has bigger returns.
The honest hierarchy at most museums looks like: statutory work first, basic data infrastructure second, revenue-generating amenities third, then experience investments like audio and signage. Skipping the first three to get to the fun ones is how museums end up with beautiful guides that no one can prove are worth the money.
How to actually decide for your museum
Walk through this in order, not in parallel.
First, ask whether you have a wayfinding emergency. Not a wayfinding annoyance, an emergency. Are visitors getting lost in a way that creates real distress, real complaints, real safety concerns? If yes, you need at minimum a signage triage before anything else. If no, keep going.
Second, ask whether your existing signage is actively wrong. Not just dated, wrong. If yes, strip and replace the worst offenders. Budget 15,000 to 25,000. Don't commission a master plan.
Third, ask whether you have visitor data. If you can't measure outcomes, fix that first. The most expensive investment is the one you can't evaluate.
Fourth, ask whether your statutory accessibility is solid. If not, that's the next dollar.
Fifth — and only fifth — start the audio guide conversation. Scope it as both an interpretation tool and a wayfinding tool. Look at modern AI museum guide options that handle turn-by-turn routing, not just narration. Treat the cost expectations from five years ago as outdated; get fresh quotes.
Sixth, after you've shipped the audio guide and have six months of usage data, decide whether you still need a full signage refresh. You may discover you don't. Or you may discover you need a much smaller one than you originally scoped. Either way, you'll be deciding with evidence instead of intuition.
The director I started this piece with ended up doing a 22,000-euro signage triage and putting the rest into an AI audio guide. Eight months later her data showed that 60% of her visitors were using the guide for navigation, her front-of-house team was spending half as much time giving directions, and the full signage refresh that had been on her three-year capital plan got pushed to year five. The money she didn't spend on signs is now funding a curator position. That's the real argument for getting the sequencing right — it's not about which line item wins. It's about freeing capital for the work only humans can do.
The sequencing works in large part because the economics of audio have shifted. Platforms like Musa price on per-interaction or revenue share rather than as a capital project, which means the "audio first" side of the comparison isn't a €150K capex gamble any more — it's an opex line that only costs the museum when visitors actually use it. That's why the trade-off against a signage refresh now favours audio in ways it didn't five years ago: every visitor who opens the guide delivers margin on interpretation and wayfinding at the same time, rather than amortising against a fixed bill.
If you're sitting on a similar decision and want to pressure-test which order makes sense for your specific building and audience, we'd be glad to walk through it with you.