The short version
A London attraction is not a British audience. Madame Tussauds London draws visitors from Germany, France, the United States and well beyond, and the mix is different again in Amsterdam or New York. Commissioning figures who are popular in the host country answers a question nobody asked. So does commissioning globally famous names, which flattens the differences that matter.
The method here is to weight celebrity demand by the share of visitors each country actually sends to a given site. It changes which names rise: some figures with no global profile are dominant in a market that supplies a large slice of the door, and some globally famous names turn out to be redundant for a particular location. This case study describes the approach rather than the answers. Which figures Madame Tussauds is considering is commercially confidential, so no scores appear here.
The question a location asks that a market does not
Most audience questions have a country attached. Who resonates in the UK, who resonates in Brazil, who should front a campaign running in Mexico. The geography is given, and the job is to find the people.
A visitor attraction inverts that. The place is fixed, and the geography walks in through the door in proportions the operator does not control. Madame Tussauds London is not serving a British audience that happens to be in London. It is serving a blend, and that blend is a property of the building rather than of the country it sits in.
That sounds like a distinction without a difference until you try to make a decision with it. Every new wax figure is a commission: a real cost, a long lead time, and a fixed amount of floor space that something else could have occupied. Get it right and visitors photograph it, share it, and remember the visit. Get it wrong and it occupies prime space while people walk past.
Why the obvious approaches both fail
There are two intuitive ways to answer this, and they fail in opposite directions.
The first is to use the host country. Figures popular in the UK for London, figures popular in the US for New York. This is the approach most likely to be adopted by default, because the data is easy to get and the logic sounds right. It fails because it silently assumes the audience matches the postcode. A figure who is enormous in Britain may mean very little to a German visitor standing in the same room, and if German visitors are a meaningful share of the door, that commission has under-served them.
The second is to use global popularity. Pick names that are famous everywhere and you cannot go far wrong. This fails more quietly. Global rankings are an average, and an average deliberately erases the variation that a specific location needs. A name that is moderately known in twenty countries can outrank a name that is dominant in the three countries that actually supply your visitors. The global list is not wrong so much as it is answering a question about the world rather than about the building.
Both approaches share the same flaw: they treat the audience as a single undifferentiated thing, either national or global. The audience at a given attraction is neither.
Think of it as a weighted average rather than a simple one. If a quarter of your visitors come from Germany, German demand should count for a quarter of the decision, not a tenth and not a half. A global popularity ranking weights every country by its population or its search volume, which is to say by something that has nothing to do with who visits you. A host-country ranking weights one country at one hundred per cent and every other country at zero. Neither of those is the actual audience. The actual audience is the visitor mix, and that is a number the operator already knows.
What changes when you weight by visitor mix
Three things surface that neither of the other approaches shows.
Cross-market names separate from locally concentrated ones. Some figures score well almost everywhere. Others score exceptionally in two or three markets and unremarkably elsewhere. Neither profile is better in the abstract. Which one you want depends entirely on the visitor mix of the site you are commissioning for. A location drawing heavily from two countries should probably prefer the name that is dominant in exactly those two over the name that is respectable in thirty.
Names the English-language press does not cover become visible. This is the finding that tends to surprise people. A figure can be genuinely major in one country and close to unknown in the anglophone media that most commissioning conversations are conducted in. If that country sends a large share of visitors, the weighted view puts them near the top of the list while a global ranking leaves them off it entirely. The risk this guards against is real and structural: decisions get made from the coverage the decision-makers happen to read.
Portfolio balance becomes checkable. Individual commissions are usually judged one at a time, which makes it hard to notice an accumulated skew. With a weighted view you can ask a portfolio question instead: given who actually visits this site, is the collection over-weighted towards some markets and thin on others? That is a question with an evidence-based answer rather than an opinion-based one.
How the weighting works
The mechanism is deliberately unexciting, which is a point in its favour.
Pulse measures demand independently within each country, rather than producing one global figure. That is the part that makes the rest possible: if demand has already been averaged into a single worldwide number, the market-level variation has been destroyed and cannot be recovered.
Each location then contributes its own visitor composition, which the operator knows from its own admissions data. That composition becomes the weighting. A country supplying a fifth of the visitors carries a fifth of the weight in that location's score, and a country supplying two per cent carries two per cent.
Combining the two gives a demand score specific to a location rather than to a country. The same figure can score differently at two sites in different cities, which is the correct behaviour and the entire point.
One consequence worth stating plainly: the output is only as good as the visitor composition that goes into it. This is not a method that discovers who your audience is. It takes the audience you already know about and tells you who that audience cares about. Those are different questions, and conflating them would overstate what the approach does.
Pulse is not counting the visitors. The attraction already knows where its visitors come from, from ticketing and admissions data. What Pulse supplies is the other half: who has genuine audience demand inside each of those countries. The weighting is the arithmetic that joins the two together. If the visitor data is stale or wrong, the weighted demand score inherits that problem, which is why the composition figures are worth checking before anyone acts on the output.
Where else the same shape appears
Madame Tussauds is a clear case because the audience mix is so visible, but the structure is not unusual. Any operation that serves an international audience from a fixed location has the same problem, usually without naming it.
A flagship store in a tourist district is stocked for a catchment that is substantially not local. A festival books a lineup that will be judged by attendees who travelled to be there. A museum programmes exhibitions for a visitor base drawn from far beyond its own city. In each case, the decision is routinely made against the host country's tastes or against a global ranking, when the relevant audience is neither.
The common thread is that the composition of the audience should shape the decision, and that composition is usually already known to the organisation, sitting in a system nobody has connected to the demand question.
What this is and is not
This case study describes an approach, not a result. Which figures Madame Tussauds is evaluating, and how they scored, is commercially sensitive and is not disclosed here or anywhere else. No demand scores appear in this piece, and the absence is deliberate rather than an oversight.
What can be said is the shape of the finding: weighting by visitor mix reorders the candidates relative to both a host-country view and a global view, and the reordering is large enough to change which figures make the cut.
A London attraction is not a British audience. It is a blend that belongs to that building, and it is different again in Amsterdam or New York. Once you see the problem that way, reading demand one country at a time looks obviously wrong, and so does reading it globally. Simon Jacobs, Audience Strategies
Three things to take away
- The relevant geography is the audience, not the address. A fixed location with international visitors is not served by the host country's demand data, and a global ranking is an average that erases exactly the variation the location depends on. Weight each market by the share of visitors it actually sends.
- Weighting surfaces names that coverage misses. A figure can be dominant in one country and invisible in the English-language press where commissioning conversations happen. If that country supplies a large share of visitors, a weighted view ranks them highly while a global view omits them entirely.
- The method needs the audience data you already hold. This does not discover who visits you. It joins your existing visitor composition to independent country-level demand. The output inherits the quality of the visitor figures, so those are worth verifying before acting on the result.
About Pulse. Pulse is Audience Strategies' audience demand platform. It measures demand for public figures independently within each market, rather than producing a single global figure, which is what makes location-level weighting possible. It is built for decisions where the question is not who is biggest overall, but who matters to a particular audience in a particular place.
This case study is published with Madame Tussauds' agreement. It describes the analytical approach only. No demand scores, candidate figures or visitor composition figures are disclosed.
About this analysis. Demand is measured independently within each market rather than aggregated into a single global score, so that market-level variation is preserved and available for weighting. Each location's visitor composition, supplied by the operator from its own admissions data, is applied as the weighting: a market contributing a given share of visitors carries that share of the location's demand score. The result is a demand score specific to a site rather than to a country, and the same figure can score differently at two locations with different visitor mixes. The method takes visitor composition as an input and does not estimate it; the output is therefore sensitive to the accuracy and the vintage of the operator's own visitor data. Candidate figures, demand scores and visitor composition for Madame Tussauds are commercially confidential and are not disclosed in this case study, which describes the approach rather than the results. Published July 2026.