The Multiple Is the Narrative: How Acquirers Actually Price a Brand Story

The £1.3 billion Puig paid for Charlotte Tilbury Limited in 2020 looked, on the trailing financials, like a category error. The business had reported 2018 revenue of £101 million on EBITDA of £3.8 million. Divide one by the other and the deal multiple resolves to 342 times trailing earnings, a number so large it suggests Puig's investment committee had taken leave of its senses, or that the financial press had garbled the story, or that some basic principle of valuation had been suspended for the occasion.

None of these things had happened. What had happened was that Puig was not paying for 2018 EBITDA. Puig was paying for the future Charlotte Tilbury would have if its customers kept doing what the customer research said they were doing, which they were going to keep doing because they had, by 2020, formed habits around the brand.

The £1.3 billion was the present value of those habits, projected forward at a confidence level the trailing accounts could not justify on their own. Brand was the variable that justified it. This essay names the mechanism and shows the maths.

I. A multiple is the price of a forecast, not the price of a year

The first thing to understand about an acquisition multiple is that it is not what it appears to be. It is reported as a ratio, twelve times EBITDA, fourteen times earnings, a particular number of times the most recent twelve months of cash flow, and the form of the ratio implies that the multiple is somehow a function of what just happened. It is not. The multiple is the present value of what the buyer believes is going to happen, compressed into a shorthand that uses last year's number as a reference point. The trailing figure is the denominator. The numerator is the buyer's view of the future. The multiple is what falls out when one is divided into the other.

This is not a technicality. It is the entire substance of why two businesses with identical trailing financials can sell for materially different prices. The buyer who pays a 4× multiple has, in effect, looked at the business and decided to commit capital based on a forecast that runs about 4 years before the model loses confidence in its projections. The buyer who pays 12× is committing capital against a forecast that runs three times as long. The first buyer is paying for what the business is already doing. The second buyer is paying for what the business is going to keep doing, for years, beyond any horizon the trailing data alone could justify.

What this essay calls the forecast horizon — the period over which an acquirer is willing to project the visible cash flows forward before reaching for terminal-value assumptions and the model's confidence collapses — is the most consequential single number in any private transaction. It is set, in published practice, by reference to the discounted cash-flow framework Aswath Damodaran has spent four decades writing about, in which the explicit-projection period is bounded by how long the analyst can defensibly model.(1) But the analyst's defensible modelling period is not a property of the spreadsheet. It is a property of the business being modelled. Specifically, it is a property of how predictable the business's customers are.

A multiple is a discounted cash-flow forecast wrapped in a story about durability. The story is about people, doing what people do.

II. What buyers are actually betting on

The thing an acquirer is doing when they extend the forecast horizon is not, fundamentally, an act of valuation. It is an act of inference about human behaviour.

The buyer is reviewing the trailing financials and asking whether the customer behaviour that produced those numbers is habitual or incidental. Habituated behaviour, customers who buy a particular brand on routine, who have built it into their grocery list or their morning ritual or their replacement cycle, is among the most predictable things human beings do. The empirical work of the Ehrenberg-Bass Institute has spent decades documenting this in laws-of-marketing form: brands grow by being mentally available at the moment of decision, and once a brand has built a place in a consumer's mental architecture, the consumer continues to retrieve it under enormous variation in promotional pressure, competitive activity, and even quality slippage.(2) This is not a theory. It is the law-like pattern visible in panel data across categories and decades.

Incidental behaviour is the opposite. It looks the same in the trailing twelve months, the revenue prints, the customers convert, the metrics line up, but it has no underlying habit structure. The customer bought because the price was right this quarter, or the channel was running a promotion, or a particular acquisition campaign hit. The behaviour will not repeat unless the conditions that produced it repeat. The trailing year, in this case, is not a base from which to project. It is a coincidence that is priced as if it were a base.

The acquirer's diligence team is, fundamentally, in the business of telling these two things apart. The questions they ask are not abstractions about strategy. They are probes for habituation. Customer research is read for evidence that the brand is mentally available, that customers retrieve it without prompting, describe it consistently, and attribute their loyalty to the brand rather than to last quarter's promotion. Pricing analysis is conducted to assess pricing power, a behavioural marker indicating that customers have developed sufficient trust to absorb a price increase without defection. Distinctive-asset audits are run because consistent visual codes are the mechanism by which the brand is recognised across encounters — and recognition under variation is the empirical fingerprint of habituated demand.

When the diligence team finds these markers, they extend the forecast horizon, because they are betting on a pattern of behaviour that the entire weight of consumer-research literature says will persist. When they don't, they shorten the horizon, because they are looking at trailing numbers with no habit structure underneath them, which means the numbers might do anything.

The multiple that emerges is, in this sense, a behavioural judgment denominated in pounds. The judgment is about whether the customers will keep doing what they've been doing. Brand is the evidence that they will.

III. The Charlotte Tilbury maths, reverse-engineered

This is where the £1.3 billion stops looking like a category error and starts looking like a behavioural bet that has now been settled.

Reverse-engineer the deal at industry-standard prestige beauty assumptions: revenue growth in the 20–30% range over a five-to-seven-year explicit-forecast period, EBITDA margins expanding to the 25% mature-category benchmark as scale is achieved, a normal private-market discount rate. The £1.3 billion that looks like 342 times trailing earnings collapses cleanly to a perfectly conventional 5–8× forward EBITDA at the end of the forecast period. From that angle, the deal is not aggressive. It is approximately rational, provided the forecast itself is rational. Provided, in other words, that the customers were actually going to keep doing what the customer research said they were doing.

Reverse-engineered scenarios, starting from reported 2020 revenue of £258.5m (Companies House), at industry-standard prestige beauty margins. The £1.3 billion valuation is internally consistent with several of these scenarios.(3)

Three things were being underwritten in the forecast. First, that the brand had built genuine mental availability, that when consumers reached for prestige makeup, Charlotte Tilbury was retrieved, repeatedly and across markets, as a brand worth considering. Second, that the pricing structure was supported by pricing power, in the behavioural sense: that customers had formed enough trust to absorb premium pricing without churning to the cheaper alternatives the category contains. Third, that the apparatus generating these patterns — the distinctive packaging, the retail aesthetic, the consistent narrative voice across channels — was robust enough to keep generating them after Tilbury herself stepped back from any individual product launch.

Six years on, the bet has been tested. Puig has publicly reported that Charlotte Tilbury Limited's net revenue has more than tripled since the 2020 acquisition, contributing materially to Puig's makeup segment growth.(4) Tripling from a 2020 base of £258.5m implies revenue approaching £800m by 2024–25, which sits comfortably inside the 20–30% CAGR scenarios above. The customers kept doing what the research said they were going to do. The forecast horizon proved out. The £1.3 billion, looked at retrospectively, was the price of a behavioural bet that turned out to be correctly read.

342× trailing was never the multiple. The multiple was always 5–7× forward, plus a story about whether the customers were habituated. The customers were habituated. The price was right.

IV.

What this looks like from the founder's side of the table

The structural implication is uncomfortable and worth stating plainly.

Most founders preparing for exit reverse the arithmetic. They start with a target multiple, we want to sell at 10× EBITDA, and work backwards to the EBITDA the business needs to print. The plan that emerges is, accordingly, a plan to grow EBITDA. This is not unreasonable. It just doesn't address the variable that actually moves the multiple, which is the buyer's confidence in extending the forecast horizon. Two founders running businesses with identical EBITDA can exit at materially different valuations, because the buyer will read different behavioural patterns underneath the identical numbers. The founder, whose customers buy out of habit, recognise the brand without prompting, and absorb price increases without defecting, can credibly support a five-to-seven-year forecast. The founder whose customers buy because the channel was right or the promotion landed cannot. The first founder sells at the high end of the range. The second sells at the low end. The EBITDA is the same. The behaviour is different. The price reflects the behaviour, not the EBITDA.

This is why, as developed at length in [[The Five-Year Window]], the work of building genuine customer habituation has to begin years before the deal table. Habituation is not a marketing campaign. It is the cumulative effect of consistent encounters between a brand and its customers across enough time that the customer's brain stops processing the brand as a fresh decision and starts processing it as a familiar one. The diligence team will read the customer research for evidence that this has happened. The evidence cannot be manufactured in twelve months because consumer memory does not work in twelve-month timeframes.

Two founders, identical EBITDA, different brands. One sells for £24m. One sells for £64m. The £40m delta is the forecast horizon, and the forecast horizon is whether the customers will keep showing up.

V. Why this matters at exit

Underneath the entire architecture of a private transaction is a question about human behaviour. The buyer is asking whether the people who made the trailing financials happen are going to keep making them happen for long enough to justify the price being negotiated. The financials are evidence. They are not the answer. The answer lies in the behavioural patterns underneath them, and whether those patterns are habituated or contingent.

[[The Brand Premium at Exit]] argues this principle at the level of strategy: brand is enterprise value because brand is the legible, transferable record of customer habituation, and customer habituation is what acquirers pay forward-multiples for. [[The Goodwill Fallacy]] explains why the formal accounting language of the firm, designed for verification rather than valuation, refuses to recognise this asset until the moment of sale. [[The Founder Dependency Discount]] explains why some founder-led businesses fail to convert visible customer affection into the kind of habituated, transferable behaviour buyers will underwrite — and why this is a structural failure rather than a personal one. [[The Five-Year Window]] explains why the lead time on building the asset cannot be compressed, because consumer memory cannot be compressed.

The present essay completes the picture by showing how all of this is settled, technically, in a single number. The number is the present value of a story about whether customers will keep doing what they have been doing. The story that people believe drives the premium. The story that does not get believed produces the discount.

Acquirers do not pay for last year's earnings. They pay for the years they believe will follow, and brand is the durability of that belief because brand is the record of customers having already learned to behave in a way the buyer can underwrite.

The buyers are sophisticated. The diligence teams are trained, sometimes brilliantly, to distinguish habituated demand from incidental demand, and the investment committees do not approve premium multiples on assets whose forecast horizons cannot be defended. None of this is mysterious. None of it is hidden. The only question, on the founder's side of the table, is whether the years of work required to produce habituated customer behaviour have been done, and whether the evidence that they have been done is legible, in diligence, to a buyer who will read it carefully and price it precisely.

What changes, depending on whether the founder has done that work, is whether they get paid for the asset they actually built, or for the asset their balance sheet was prepared to admit they had built. These are not the same number. The difference is the central argument of this publication's writing on exits — and the difference is, at the moment of exit, the difference between the multiple they expected and the multiple they receive.

Sources & References

(1) Aswath Damodaran, Closure in Valuation: Estimating Terminal Value, Stern School of Business, NYU. The standard treatment of the explicit-forecast-period and terminal-value structure underlying all DCF-based valuations. Damodaran's valuation spreadsheets and reference materials are the canonical practitioner resource for the mechanics described in section I.

(2) Byron Sharp, How Brands Grow: What Marketers Don't Know (Oxford University Press, 2010); Jenni Romaniuk and Byron Sharp, How Brands Grow: Part 2 (Oxford University Press, 2016). The Ehrenberg-Bass Institute's empirical research on mental availability and the laws of buying behaviour provides the foundation for the habituation argument made in section II.

(3) Reported financials per BeautyMatter's industry analysis of the Puig–Charlotte Tilbury transaction: 2018 revenue £101 million, EBITDA £3.8 million. 2020 revenue of £258.5 million per Companies House filings cited in FashionNetwork. Reverse-engineered scenarios calculated at industry-standard prestige beauty assumptions, including 25% mature EBITDA margins consistent with peer companies in the prestige-beauty category. The scenarios are illustrative of the mechanism rather than a reconstruction of Puig's specific internal model, which is not in the public record.

(4) Puig press release on the Charlotte Tilbury partnership extension, reporting that Charlotte Tilbury Limited's net revenue has more than tripled since the 2020 acquisition. The reported tripling between 2020 and 2024 implies a five-year revenue CAGR of approximately 25% from the 2020 base, consistent with the upper-mid range of the reverse-engineered scenarios in section III.

(5) Hamilton Helmer, 7 Powers: The Foundations of Business Strategy (Deep Strategy LLC, 2016). Helmer's framing of Branding as a structural source of persistent differential returns provides the strategic-economics complement to the behavioural argument made here. The "Power" producing the differential returns is, in valuation terms, what the acquirer is paying the multiple for.