Same shortlist. Different answers.

The question a shortlist cannot answer

A talent shortlist arrives as a list of names. Sometimes it arrives ranked. Either way it carries an implicit claim: these are the strongest candidates, in roughly this order. The claim is almost never examined, because examining it requires a comparison nobody has the data to make.

We built a shortlist of the kind that lands on a marketing director's desk. Thirty women who publicly fronted a paid fashion or beauty campaign running between roughly May 2025 and May 2026, drawn from a trade press scan of WWD, Vogue Business, Business of Fashion, Glossy, Cosmetics Business and brand newsrooms. The list spans established figures (Nicole Kidman for Chanel, Gillian Anderson for L'Oréal Paris), founder-faces of their own brands (Beyoncé, Rihanna, Selena Gomez), and newer names in their first major ambassador roles.

Then we scored all thirty through Pulse, our audience demand platform, in ten markets: the UK, the US, Australia, Canada, Germany, France, Italy, Spain, Mexico and Brazil.

The output is a ranking. The ranking is the least useful part of it.

What a demand score actually is

Before any of the findings mean anything, it is worth being precise about what a 0 to 100 demand score is, because the construction determines what you are allowed to conclude from it.

Up to eight signals feed each score, drawn from six platforms. Three are search signals: Google search volume over the trailing twelve months, year-on-year growth or decline in that search volume, and YouTube search volume. Four are following signals: Instagram, X, TikTok and YouTube subscribers. The eighth is organic web traffic to the entity's own site, arriving via Google search.

Each of those raw measures is converted into a percentile rank within the peer group, per market, before anything is combined. Margot Robbie's Google search figure of roughly 181,000 average monthly searches does not enter the score as 181,000. It enters as 93, meaning her search volume sits above 93 per cent of this shortlist. The eight percentiles are then combined into a score for each market, the ten markets are combined with equal weight into a headline figure, and the result is rescaled across the peer group so the strongest name reads 100 and the weakest reads 0.

Equal market weighting is a default, not a law. It is the right starting point for a shortlist exercise where the campaign footprint is not yet fixed, because it lets every market speak at the same volume. It is the wrong setting the moment a brand knows that 60 per cent of the media budget is going to one territory.

That last step is the one that gets misread, so it is worth stating plainly rather than burying in a footnote. The scores are relative to these 30 names and to nothing else. The name at the bottom of the weighted view scores 0. She is a working actor who fronted a major campaign inside the window, and she scores 41 in the UK, which puts her mid-table in the one market a British retailer would be buying. Her 0 means "lowest of these 30 women in fashion and beauty", not "no audience demand".

This is a real constraint, and a deliberate one. An absolute demand measure sounds more useful than it is. Search volumes are not comparable across languages or market sizes without normalisation, follower counts inflate over time and almost never fall, and a raw number gives you no way to judge whether it is good. The question a brand is actually asking is comparative: of the people I could realistically book, who has the most audience attention in the market I care about? A peer-group percentile answers exactly that question and refuses to answer the ones it cannot. The cost is that the peer group has to be constructed honestly, because it defines the scale. Ours is deliberately narrow: one sector, one gender, one twelve-month window of live campaigns. Market concentration would show up in any sector; which names turn out to be the specialists is particular to this one.

The ranking, and why it is a starting point

On the weighted all-country view the top of the list runs: Margot Robbie 100, Jenna Ortega 97, Dua Lipa 93, Selena Gomez 90, Rosalía 86, Sydney Sweeney 83, Rihanna 79, Beyoncé 76, Kendall Jenner 72.

Nothing about that list is surprising, which is roughly the point. It broadly matches what an experienced person in the room would have guessed. If the exercise ended here it would have cost money to confirm an intuition.

The value starts when you notice that this single ordering is the average of ten quite different orderings, and that no campaign in the world runs against the average.

It depends on where you are running the campaign

Three of the names at the top hold their position almost everywhere. Jenna Ortega's scores across the ten markets span just ten points, from 90 in Australia to 100 in five of the ten. Sydney Sweeney spans fourteen. Margot Robbie spans seventeen, from 83 in France to 100 in the UK, Australia and Italy. These are genuinely global faces, and a multi-market campaign can use any one of them as a single global partner without losing meaningful demand in any territory. That is a real and expensive property, and the data supports paying for it.

Now look at the other end of that distribution.

Bruna Marquezine scores 90 in Brazil and 3 in the UK, a gap of 87 points. Across all ten markets she ranges from 90 in Brazil down to 0 in Australia, a 90-point spread that is the widest on the list alongside Rosalía. In Brazil she sits fourth, behind Jenna Ortega (100), Dua Lipa (97) and Beyoncé (93), and ahead of Margot Robbie, who scores 86 there. Aya Nakamura scores 79 in France and 0 in both the US and the UK. Eiza González scores 66 in Mexico and 7 in the UK.

The pattern is broader than those three names, and the additional cases are more interesting because they are less predictable from biography. Gillian Anderson scores 66 in the UK and 10 in the US, a 56-point spread, despite being American-born and having spent much of her career in American productions. Nicole Kidman scores 79 in Australia against 31 in France. Coco Gauff scores 55 in the US and 10 in Brazil. Emma Chamberlain scores 45 in Canada and 3 in France.

Nationality is a weak guide to any of this. It co-occurs with market strength often enough to be tempting and fails often enough to be dangerous. Gillian Anderson is the clean counter-example: American-born, and materially stronger in Britain than in the country she comes from.

The commercial consequence is straightforward. For a single-market campaign, a concentrated name can offer demand comparable to the global tier in the one territory that matters, at a cost that almost certainly does not reflect global star pricing. That is not a small saving; it is frequently the difference between affording talent and not. It is only visible if you are looking at the market-level scores rather than the headline.

Dua Lipa is the case that sits between the two patterns and is worth its own paragraph. She is third on the weighted view. In the UK, her home market, she scores 79, which places her seventh, behind Margot Robbie, Jenna Ortega, Sydney Sweeney, Olivia Rodrigo, Rihanna and Beyoncé. Seventh of thirty is a high score, not a weak one. What distinguishes her is what happens outside the UK: first in Mexico at 100, second in Brazil at 97, third in Italy at 93, fourth in Spain at 86. For a British brand that needs a face with genuine international range, she is among the strongest candidates on the list. For a brand leading with Latin or Mediterranean markets that also needs solid English-speaking reach, she may be the strongest available, and her UK ranking would have talked you out of her.

It depends on what you are asking them to do

The second thing a headline score hides is the shape of the demand underneath it.

Margot Robbie tops this list. She also has no social media accounts on any platform Pulse tracks: no Instagram, no X, no TikTok, no YouTube channel, and no personal site generating measurable organic traffic. She deactivated her Instagram in 2021. Her score of 100 is built entirely from search behaviour and press attention, which is to say from people going and looking for her rather than from an audience she owns and can address.

Whether that is a strength or a disqualification is entirely a function of the brief. For a campaign built on discovery, search visibility, PR and earned attention, it is not a weakness at all. For a brand that needs the talent to activate an existing following directly, through their own channels, on a schedule the brand controls, she cannot do it at any price. No amount of demand score changes that.

Jenna Ortega sits second and operates on Instagram only, with roughly 38.7 million followers globally. She scores 97th percentile on Google search and 93rd on YouTube search. She is a search phenomenon with a substantial but deliberately single-platform social presence.

Shay Mitchell is close to the inverse. She scores 7th percentile on Google search and 3rd on YouTube search, but 74th on Instagram, 57th on X and 54th on TikTok. Nobody is searching for her in volume. A large, engaged community follows her directly. If you are booking her, you are buying access to that community, and the brief needs to be built around content she publishes rather than attention she attracts.

Selena Gomez has the largest Instagram following on the list at 100th percentile, along with 100th on TikTok and 95th on X, against 72nd on Google search and a search trend that has been falling steadily across all three years of monthly data, down roughly 39 per cent on a trailing twelve-month basis. A social mass asset, with fading search momentum underneath it.

Four names, four completely different assets, arriving in the same shortlist under the same heading.

Which means the brief has to specify things that briefs habitually leave implicit. Before demand data can answer who is right, someone has to have decided: is the campaign built on the talent's owned channels or on earned and paid attention? If owned, which platforms specifically, because the four names above are strong on four different ones? Is the objective discovery, where search behaviour is the relevant signal, or is it reach into an existing community, where following is? And which markets is the campaign actually running in, at what weight? Those four answers narrow thirty names much harder than any ranking does. Without them, the data can produce a list, but the list is answering a question the brand has not yet asked.

It depends on whether you are buying the moment or the trajectory

The third thing a score cannot show you is direction, and direction is what you are actually buying, because a campaign runs forwards.

We looked at monthly global Google search volume for the full window the data covers, April 2023 to March 2026, and took a trailing twelve-month average to strip out seasonality and single-month spikes.

Most of the names on this list are on the downward slope of their current cycle. Beyoncé's trailing average peaked in August 2025 and had fallen to about 71 per cent of that peak by March 2026. Demi Moore's peaked in the same month and stood at about 67 per cent by March. Chappell Roan's peaked earlier, in April 2025, and had fallen to about 56 per cent.

Rosalía is the exception in this group. Her trailing average sits at its highest point in the entire window at the final month of data, March 2026, following the release of Lux in November 2025. She is the one name here measurably still rising.

Search is the fastest-moving of the eight signals, and this is structural rather than incidental. Follower counts are cumulative stocks. They ratchet upwards and almost never fall, so a following signal records that someone was once popular far more reliably than it records that they are popular now. Search is a flow. It measures attention in the month it happened. That is why the demand cycle is visible in search and effectively invisible in the follower signals, and it is why the score includes a year-on-year search growth component.

That weighting is itself a judgement worth surfacing. Beyoncé illustrates the tension. She sits eighth on the weighted view at 76, on the strength of very large follower counts and solid search volume. On the search growth signal she scores 0, the lowest of the thirty, with search down around 95,000 average monthly searches year on year. A score that weighted momentum more heavily would rank her materially lower. A score that ignored it would rank her higher. Ours does something in between, and a client with a strong view on this can have the weighting changed, which is preferable to pretending there is a neutral answer.

What the trend chart is showing, then, is not a growth ranking. It is where each name sits in their demand cycle at the point the data was cut. A partnership activated with someone whose wave has already broken buys declining demand for the duration of the campaign. The right question is which names are still on the upward slope within the activation window, and which have a release, a tour, an award season or a franchise moment coming that would generate the next peak.

What this changes about how a shortlist gets used

None of this replaces the shortlist. The thirty names were chosen by people who know the sector, and the data did not reveal that any of them was a mistake.

What it does is answer the question the shortlist itself cannot: not who is biggest, but right for what. Right for which market, when the same name can score 90 and 3 in the same week. Right for which platform, when the top name has no social accounts and a name in the middle of the pack has a community of millions. Right for this point in their cycle, when most of these names are past their peak and one is not.

Same thirty names. A different answer every time the brief changes. Which is an argument for writing the brief more precisely before anyone starts arguing about names.

The ranking was the least interesting thing in the data. Every genuinely useful finding came from breaking it apart. The moment you specify a market and a platform, the order changes completely, and the name you would have picked from the headline is frequently not the name the brief actually needs. Simon Jacobs, Audience Strategies