KPIs for Museums: The Performance Indicators That Actually Matter

In brief. A museum doesn't need forty indicators: five are enough, chosen with one test, would this number ever change a decision? They are actual visit duration against planned duration, how much people understood, how many return or leave a contact, how much they spend beyond the ticket, and how many come from the region versus from further away. All five cost nothing to collect, even without any software.

The number everyone counts, and what it doesn't tell you

Ask a director how the season went, and the answer is almost always a single number: admissions.

That's understandable. Admissions get counted because government agencies count them, because they land in annual reports, funding applications, the local paper. It's the one number the system actually asks for, which is exactly why it's the one almost everyone produces.

The trouble is, it only tells you one thing: how many people crossed the threshold. Not whether they stayed. Not whether they understood anything. Not whether they'll come back.

Economists know this well as Goodhart's law, and in the phrasing that made it famous: when a measure becomes a target, it stops being a good measure. Charles Goodhart coined the idea in 1975, writing about monetary policy, and the anthropologist Marilyn Strathern condensed it into that exact sentence two decades later, studying how universities get evaluated.

In cultural venues you can see it with the naked eye. When the only number that counts is how many people walk in, you end up designing the entrance rather than the visit: the blockbuster show, the coach group crossing three galleries in twenty minutes. Admissions go up. The experience, often, goes down. And the number never notices.

What KPIs are, for a museum

KPIs, short for key performance indicators, are the small set of measures an organisation watches to know whether it's heading where it meant to go. In a museum they don't just track tickets and costs: above all, they track the quality of the experience, the whole reason the museum exists.

This isn't just a management-textbook exercise, either. The UK Museum Accreditation Scheme, run jointly by Arts Council England, the Welsh Government, Museums Galleries Scotland and the Northern Ireland Museums Council, sets out what a museum has to do to hold Accreditation status. Under standard 8.2, on assessing user needs, a museum is expected to use feedback from users to inform how it and its programmes develop.

The opposite risk is just as real: the forty-indicator dashboard nobody looks at, because whoever should is also covering the front desk. So there's only one selection test: if this number changed, would I change a decision? Anything that fails it is a number for a CV, not for running a museum.

Five indicators come out of that question.

The five KPIs that are enough for a museum

1. How long people stay: actual dwell time against planned dwell time

What it measures. The gap between the visit you designed and the visit that actually happens. It's the first attention indicator, and the most studied one in museum literature.

How to calculate it. Median actual visit duration, divided by the duration the route was designed for. If the route is meant to take sixty minutes and the real median is twenty-five, the ratio is 0.4.

How to collect it for free. A notebook and a watch. Pick twenty random visitors on an average day, note the time they go in and out. Twenty observations aren't a statistic, but they're enough to spot a big gap, and a big gap is almost always what you'll find.

What decision it changes. If people stay far less than planned, the problem isn't a public in a hurry: it's a route asking for more than it gives back. Shorten it, split it into two itineraries, put the best piece first instead of last. Watch for the opposite signal too: a very long stay with low ratings isn't enthusiasm, it's often people wandering in circles looking for the exit.

2. How much stuck: one question on the way out

What it measures. What's left behind. Average satisfaction tells you whether the visit was pleasant, not whether it left anything with the visitor.

How to calculate it. Percentage of visitors who can name at least one thing they learned, or that stayed with them, out of everyone asked.

How to collect it for free. One question at the exit, asked aloud or on a card: "what are you taking home with you?" Not a ten-item survey. One open question, gathered for a week, tells you more than a year of star ratings.

What decision it changes. If twenty out of twenty people name the same object, the rest of the collection isn't communicating: that room needs rewriting, not refurbishing. If nobody names the piece you invested the most in, the problem is how you tell its story, not what you own.

3. How many come back, or at least leave a contact

What it measures. The museum's ability to turn a visit into a relationship. At most cultural venues, the visitor disappears for good at the door: an online shop knows everything about whoever buys a t-shirt, while a museum often doesn't know who just walked through.

How to calculate it. Two numbers, in ascending order of difficulty: the percentage of visitors who leave a contact with explicit consent, and the percentage who say they've visited before (or who show up with a membership).

How to collect it for free. An exit form with a reason worth filling in, not "sign up for our newsletter" but something of real value: bonus content, a discount on the next show, the digital catalogue. Plus one question at the ticket desk: first visit, or back again?

What decision it changes. With a contact list, the next show gets announced to people who already loved the last one, the cheapest audience there is to reach. Without one, every show starts from zero and has to pay to get noticed.

4. How much people spend beyond the ticket

What it measures. The value visitors place on the experience after living it, and the museum's financial health in the parts where the ticket price alone doesn't cover costs.

How to calculate it. Shop, café, donation and add-on service revenue, divided by the number of visitors in the period. Reads as revenue per visitor.

How to collect it for free. You already have it: it's the receipts and donations you record anyway, divided by the admissions you count anyway. All that's missing is doing the division, every quarter.

What decision it changes. It's the number a trustee or a city funding officer grasps instantly, which makes it the one that gets the next round of decisions financed. It's also an indirect read on how the visit ends: people buy the memory of something that left a mark, not of a visit that ended in tiredness.

5. Who's local, and who's just passing through

What it measures. Who actually walks in, which often looks quite different from the audience described in planning documents.

How to calculate it. Share of visitors who live in the city or region, against the share from further afield or abroad. Add language, and you have the other half of the picture.

How to collect it for free. One question at the ticket desk, home town, for two weeks in high season and two in low. A digital audio guide gives you the language for free, no asking required.

What decision it changes. Everything to do with communication and languages. If the audience is largely passing through, foreign-language content, a presence on tourist channels and deals with tour operators make sense. If it's mostly local, you need reasons to come back: a new show, a membership, an evening event. It's also the only way to notice something uncomfortable and useful, like fewer visitors coming from the museum's own city than from a foreign one three hundred kilometres away.

The two indicators to stop watching

Followers and likes. Neither tells you whether anyone walked in, or understood anything. It's an audience that looks like yours, but isn't: part of it doesn't live within a couple of hundred kilometres, and part isn't even a person.

That doesn't mean abandon social media, it means stop using it as a measure of results. If you want one digital number that actually counts, keep the percentage of visitors who, once on your website, find opening hours, prices and directions in under a minute. That figure moves admissions. Follower counts almost never do.

Who looks at the numbers, how often, and who to compare with

Three organisational details that matter more than which indicators you pick.

Who looks at them. The problem is rarely producing the data, it's that nobody's job is to read it. Without a name attached to the number, the number is useless. In a small museum that name is almost always one person, and it should be written down, not assumed.

How often. Quarterly suits nearly everyone. Monthly produces noise, since weather and school trips move the numbers more than your decisions do. Yearly produces archaeology: by the time you notice something, the season is over.

Who to compare with. There's no single right value for everyone: a small civic museum and a historic house museum in a heritage city don't have the same numbers, and shouldn't. The useful comparison is with yourself three months ago, against the same season a year earlier. Comparing yourself with the museum down the road is only good for an argument.

KPIs for large and small museums: what changes

Size doesn't change which indicators matter, it changes how many you can afford to follow.

A large museum with a dedicated team adds cost per visitor, gallery-by-gallery efficiency, individual show performance, event participation and finer audience segmentation on top of the five. That makes sense: it has people to produce those numbers and people to read them.

A small museum trying to do the same ends up measuring nothing at all. Five numbers collected roughly but actually looked at beat forty perfect ones sitting in a file nobody opens. If you can only start with one, start with the second on this list: one question at the exit, for a week. Best ratio of effort to surprise of anything here.

How these numbers arrive on their own, when the visit is digital

Three of the five indicators collect themselves, without asking anyone anything, when visitors use a digital audio guide: actual visit duration, the points where people stop or drop off, and the language that tells you who you're dealing with. A fourth, satisfaction, comes from the questionnaire offered at the end of the route, while the visitor still has the phone in hand. That's the moment people actually answer: paper questionnaires left at the exit typically get one to three responses per hundred visitors, and almost only from those who are either delighted or furious.

This piece doesn't come from a textbook. On our platform we've watched over a million visits across more than two hundred cultural venues and collected over 37,000 visitor questionnaires, and the thing that struck us most is that the museums that improve aren't the ones measuring everything: they're the ones watching a few things, regularly, and changing something after they've looked.

If you want to see what these numbers look like when they arrive automatically, our page on data analysis for cultural venues shows the dashboard every amuseapp museum has in its console, with month-on-month comparisons and visitor questions grouped by topic. And if the question is what it costs, the answer is on the museum audio guide cost page.

Frequently asked questions about museum KPIs

How many KPIs should a museum track?

Five are enough for most museums, and more than enough for a small one. The test isn't completeness, it's actionability: if an indicator can't change a decision, don't bother tracking it.

Is visitor numbers a bad KPI?

No, it's necessary but not sufficient. Track it, because funders and government reports ask for it, but pair it with at least one experience-quality indicator, or you'll end up optimising for the entrance instead of the visit.

How do you measure visitor satisfaction on no budget?

With one open question at the exit, gathered by voice or on a card for a week, plus a read-through of the free-text comments already sitting on review sites. A digital audio guide's end-of-visit questionnaire collects far more responses than any paper form.

How often should a museum review its KPIs?

Every quarter, compared against the same quarter the year before, to cancel out seasonality. Monthly monitoring produces noise, yearly monitoring arrives too late to fix anything.

Are there benchmark values that apply to every museum?

No, and generic benchmarks mislead more than they help, because they depend on type, size, location and audience mix. The comparison worth trusting is your own museum's track record over time.

Does museum accreditation require measuring visitor satisfaction?

The UK Museum Accreditation Scheme doesn't spell out a specific satisfaction score, but standard 8.2 does require museums to gather feedback from users and use it to shape how the museum and its programmes develop. Already having a feedback tool in place makes that self-assessment considerably easier.

The bottom line

A museum that measures five things and changes one every quarter improves faster than one that measures forty and changes none. The question worth starting from isn't which data we could collect, it's which decision we're currently making blind.

*Want to find out which of these numbers your museum already has without realising it? Get in touch and let's talk it through, starting from your own data.*