Museum attendance benchmarks: how does your venue compare?
Every museum eventually asks whether its numbers are good. The honest answer starts with a harder question: compared to what, counted how? This is a guide to benchmarking attendance properly, the KPIs that compare fairly, the baselines that mean something, and the definition traps that quietly break most comparisons.
At some point in every museum’s annual cycle, the question arrives: from a board member, a funder, or the director’s own doubt: are these numbers good? Two hundred thousand visits. Good compared to what?
Benchmarking is the discipline of giving that question a fair answer. And most attempts at it fail before they begin, not for lack of data, but because the comparison was never comparable. A museum measures itself against a neighbour that counts differently, a national average that blends incomparable institutions, or its own past measured by a different method. The result looks like insight and behaves like noise.
This guide covers how to do it properly: which KPIs actually travel between venues and years, what to benchmark against, and the definition pitfalls that decide the outcome before anyone looks at a chart. What it deliberately does not contain is a table of reference figures, and the reason why is part of the method.
The KPIs worth benchmarking
Total visits is where every benchmark starts, and where too many stop. The total is the number funders ask for and the press quotes, so it has to be right, but as a comparison tool it is blunt. Raw totals reward big buildings, long seasons and generous opening hours, none of which say anything about how well an institution is performing.
Visits per open hour is the total made comparable. Dividing attendance by the hours the venue was actually open removes the differences that make raw totals misleading: a museum closed on Mondays against one open seven days, a venue that extended evening hours against its own shorter-season past. It is the difference between asking “how many came?” and “how hard is each open hour working?”, and it is the first number to check when a year-on-year total moves after an opening-hours change, because it tells you whether the institution grew or just stayed open longer.
Dwell time measures engagement rather than throughput: how long visitors actually spend in the building and in each gallery, rather than how many passed the door. Two venues with identical attendance can be having entirely different relationships with their visitors: one holds them for an afternoon, the other processes them in forty minutes. Dwell time explained covers how to read it, and what dwell reveals about which exhibits hold visitors shows what it looks like in a museum specifically.
Repeat-visit rate answers the question the total hides: is attendance the same audience returning, or one-off reach? A rising total driven by loyal repeat visitors is a different story (for programming, for membership, for funders) than the same total driven by tourists who will never come back. Anonymous, aggregate measurement, with visitor opt-in, can separate the two without identifying anyone.
Together, these four turn “we had 212,000 visits” into something benchmarkable: how many, at what rate, how engaged, and how loyal.
What to benchmark against: three baselines
Your own history is the gold standard, and it is not close. Comparing this year’s autumn with last year’s autumn holds constant everything that makes external comparison treacherous: the building, the city, the entry model, the counting definition. The only requirement is consistency: the same measurement method, covering the same entrances, applying the same definition, year after year. A venue that switched counting methods mid-series has broken its own baseline, which is why the method’s stability matters as much as its accuracy.
The season and the calendar come next. Museum attendance breathes with school holidays, tourist flows and weather, so a fair comparison is like-for-like: the same weeks, not adjacent ones. A temporary exhibition judged against the month before it opened may just be measuring the start of the school holidays. Benchmarking against the same period in previous years, or against a multi-year average for that season, separates what the institution did from what the calendar did.
Official statistics provide the outer frame. In Sweden, the Agency for Cultural Policy Analysis (Kulturanalys) publishes the annual Museums statistics; in England, Arts Council England’s analysis of around 1,200 accredited museums tracks the sector’s recovery. These figures cannot tell you whether your venue is doing well (they blend institutions of every size and model), but they answer a question nothing else can: did your curve move with the sector’s, or against it? A flat year during a sector-wide decline is an achievement; the same flat year during a sector-wide recovery is a warning. Our annual review of museum attendance in numbers reads the latest official figures and what they mean; this guide is about how to use them.
The definition pitfalls that break comparisons
Every benchmark rests on a definition of “a visit”, and the definitions differ more than most comparisons acknowledge.
What counts as a visit. Kulturanalys distinguishes facility visits (everyone entering any part of the building, café and shop included) from activity visits, participation in the core offer, a subset. Two museums both reporting “visits” can be measuring different things, and an institution switching between the bases can gain or lose double-digit percentages without a single extra visitor. The first question about any figure you benchmark against is: counted how, and by which definition?
School classes and groups. One framework counts a school visit as thirty visits; another counts the booking; a third counts participants only if the visit included a guided programme. For venues where school visits are a large share of attendance, this single choice can move the comparison more than any real change in performance.
Free entry versus paid. A paid venue can lean on ticketing data, but tickets miss re-entries, secondary entrances and non-ticketed events. A free-entry venue has no tickets to lean on, and a door count includes everyone who came for the café. The two produce numbers that look alike and mean different things, which is why benchmarking a paid venue against a free one on raw visits is close to meaningless.
Re-entries and secondary doors. A count that covers one entrance during staffed hours is a sample, not a measurement. Comparing it against a continuous count of every entrance is not a benchmark; it is two different quantities sharing a name.
The pattern is the same in every case: compared to what? turns out to mean counted how?
Why this page gives you no benchmark table
It would be easy to publish one (a tidy row of averages per venue type) and it would be worse than useless. Any published benchmark figure blends institutions that the previous section just showed to be incomparable: national museums with village museums, free with paid, facility counts with activity counts, continuous measurement with clicker samples. A number produced that way tells you about the mix of the sample, not about your venue. In sectors with standardised definitions, published ranges can guide a first read, which is why our retail conversion benchmarks exist; museums are not such a sector.
The benchmark worth having is built, not looked up:
- Measure consistently: continuously, at every entrance, by one definition, with a method stable enough to trust across years.
- Normalise: visits per open hour, so schedule changes don’t masquerade as performance.
- Compare within: your own history, like season against like season.
- Contextualise: read your curve against the official statistics, as direction rather than verdict.
An institution that does these four things can answer “are these numbers good?” with evidence. An institution reaching for someone else’s average is answering a different question and hoping nobody notices.
The measurement underneath
None of this works on top of estimates. A benchmark series is only as good as its weakest year, which is why the foundation is continuous, auditable counting: every entrance, every open hour, calibrated against manual control counts, with the series available for inspection. Museum visitor analytics provides exactly that layer, anonymously, with the only footfall method in Europe approved by a data protection authority, and the same data that builds the internal benchmark feeds the external report. How to take these numbers to the people who fund you is its own craft, and our guide to visitor figures for funders covers it: the benchmark tells you where you stand; the funder report is where standing somewhere starts to pay.
Frequently asked questions
What is a good attendance benchmark for a museum?
There is no universal figure, and any source offering one is blending institutions that cannot be compared: national museums with village museums, free entry with paid, building counts with exhibition counts. A workable benchmark is built, not looked up, your own history, measured consistently, normalised per open hour, and read against the season and the national trend.
How do you compare attendance between museums fairly?
Align the definitions first: what counts as a visit, how school groups are treated, whether the count covers the whole building or only the core offer. Then compare rates rather than totals, visits per open hour removes differences in opening days and hours, and compare like periods, season against season. A comparison that skips these steps measures the definitions, not the museums.
Which KPIs should a museum benchmark beyond total visits?
Three earn their place: dwell time, which shows engagement rather than throughput; repeat-visit rate (measured with visitor opt-in), which shows whether attendance is loyalty or one-off reach; and visits per open hour, which makes numbers comparable across seasons, opening schedules and years. Together they turn a headline count into a picture of how the institution is actually performing.