Unique selling proposition (USP)

Origin

Reeves, chairman of Ted Bates & Company, published Reality in Advertising with Alfred A. Knopf in 1961. He wrote that the theory “originated at Ted Bates & Company in the early 1940’s” [1, ch. 13]. The problem was waste. Bates had been splitting the population into people who remembered a campaign and people who did not, then measuring how many in each group used the product [1, ch. 3]. Penetration: the share of the population who remember your current advertising. Usage pull: the difference in product usage between those who remember it and those who do not [1, ch. 3].

The definition comes in chapter 13, “The Copy Leverage”, where he complains that U.S.P. had become “perhaps the most misused series of letters in advertising” and gives it three parts [1, ch. 13]. First: “Each advertisement must make a proposition to the consumer.” Second, the proposition must be one “that the competition either cannot, or does not, offer”. Third, it “must be so strong that it can move the mass millions, i.e., pull over new customers to your product” [1, ch. 13]. He shows condition two failing alone with a toothpaste that advertised “it comes out like a ribbon and lies flat on your brush”: unique, and useless [1, ch. 13].

The USP is not a slogan, and Reeves says so. It is a product claim chosen because the Bates data said such claims produced the most usage pull. So that data is the first question.

What the evidence shows

Reeves’ own evidence. The Bates studies covered “seventy-eight of the biggest package-goods advertisers in the United States”, with budgets from $400,000 to $17,500,000 a year [1, chs. 3 and 4]. Penetration ran from 1% to 78% of the population [1, ch. 4]. His summary statistic: companies applying his principles achieved penetration 15.4% above the competing-brand average and usage pull 16.2% above, which he multiplies into a 34% advantage [1, ch. 6].

Treat these numbers as what they are: agency data on the agency’s own clients, never published in a journal, with no sampling detail, no confidence intervals, anonymous brands and no replication. One more problem, our reading rather than a published one. People who already use a product are more likely to remember its advertising, so part of the gap between rememberers and non-rememberers is caused by usage rather than causing it. Reeves reports a correlation and calls it pull.

The peer-reviewed record supports something narrower. Ehrenberg’s 1974 paper in the Journal of Advertising Research reviewed the buying-behaviour evidence for frequently bought goods and found no evidence that advertising “actually works by any strong form of persuasion” [2]. His sequence is awareness, trial, reinforcement: “for frequently bought products, repeat buying is the main determinant of sales volume and here advertising must be reinforcing rather than persuasive” [2]. That answers Reeves’ third condition directly. Reeves needs a claim strong enough to move new buyers. Ehrenberg’s data say most volume comes from existing buyers being reminded.

The one part of Reeves that later evidence supports is staying with a campaign. He called too-frequent change “the multi-million dollar error” and reported a dental cream campaign running twenty-two years while holding first place [1, ch. 8]. Ehrenberg’s model predicts that outcome, for a different reason.

The language has moved on without him. Romaniuk, Sharp and Ehrenberg summarise Kotler’s Marketing Management as talking “of a meaningful perceived difference that provides buyers with their reason to purchase” [4, p. 43]. Kotler places the USP inside positioning, as one answer to how many benefits to promote, under a wider value proposition. Aaker’s Building Strong Brands (1996) treats the functional benefit as one layer of a brand identity that also carries emotional and self-expressive benefits, and warns against fixing on product attributes [6]. Ries and Trout, in their 1972 Advertising Age series and 1981 book, moved the battleground from the product to the prospect’s mind [7]. Their debt to Reeves is the single message; their departure is that it is built from perception, not a product claim.

The challenge

Two published bodies of evidence cut against the USP, at different conditions.

Romaniuk, Sharp and Ehrenberg (2007) went after the second condition, uniqueness. They asked current users whether the brand they buy is “different” or “unique”, across seven Australian categories collected by telephone and ten UK categories from Young & Rubicam’s 1999 Brand Asset Valuator survey, 17 markets in all [4, pp. 45-46]. Averaged across those markets, 11% of a brand’s own users said it was different, 10% said unique and 17% said either [4, Table 2]. Brands with an obvious functional difference scored higher: Aldi, 67% for different [4, p. 46]. Apple, “often presented as a poster child for differentiation”, still had 77% of its users not perceiving it as different or unique [4, p. 48]. The authors’ conclusion: “less trying to find unique selling propositions and more trying to find unique identifying characteristics” [4, p. 52].

It measures stated perception, not what drives choice, mostly for established brands [4, p. 49]. What it removes is Reeves’ second-condition assumption: that buyers of a successful brand carry a perceived unique reason to buy it. Mostly they do not.

Kantar reads the same question from the other side. Its BrandZ work with the Saïd Business School at Oxford modelled 872 brands from 2006 to 2022 and found perceived difference to be the leading brand factor in share-price outperformance, 35% of the brand effect against 0.6% for salience, and that Meaningful and Different together explain 94% of pricing power, salience 6% [10]. Read with Romaniuk: few buyers perceive a difference, but where they do, it is what lets the brand charge more. Neither result says a unique product claim is how that difference gets built. Kantar’s own factor analysis puts 30% of difference on distinctive assets and the rest on positioning, product experience, service and advertising associations [10].

Pricing power: Meaningful and Different94%

Pricing power: Salient6%

Share-price outperformance, brand effect: Different35%

Share-price outperformance, brand effect: Salient0.6%

Buyers rarely call a brand different, but the difference they do perceive is what lets it charge more; salience fills the shelf, not the margin.

Source: Kantar, Think Different: The DNA of breakthrough brand value growth, 2023, Kantar BrandZ with Saïd Business School Oxford, n=872 brands 2006 to 2022, figures in text. Redrawn by Brave New Media.

Unit: % of effect explained

Show the data

Item	% of effect explained
Pricing power: Meaningful and Different	94
Pricing power: Salient	6
Share-price outperformance, brand effect: Different	35
Share-price outperformance, brand effect: Salient	0.6

Binet and Field went after the mechanism. Their 2013 IPA report The Long and the Short of It analysed 996 campaigns from 700 brands across 83 categories in the IPA Databank [5]. Campaigns coded as rational (a product or price message) reported very large effects on sales in 46% of cases, against 55% for emotional campaigns; on market share 28% against 31%; on loyalty 5% against 11%; on price sensitivity 2% against 5%; on new customers the two tied at 27% [5, Figures 44, 53, 54]. Very large profit growth was reported by 16% of rational campaigns, 25% of combined and 29% of emotional [5]. The gap grows with time: at one year 13% of emotional campaigns against 10% of rational; at two years 30% against 20%; at three years or more 43% against 23% [5].

Rational campaignsEmotional campaigns

Sales46%55%

Market share28%31%

Price sensitivity2%5%

Loyalty5%11%

New customers27%27%

The rational claim ties on new customers, the one metric the USP was designed for, and loses on the four that compound.

Source: Binet and Field, The Long and the Short of It, IPA, 2013, n=996 campaigns, Figures 44, 53, 54. Redrawn by Brave New Media.

Unit: % of cases

Show the data

Metric	Rational campaigns	Emotional campaigns
Sales	46	55
Market share	28	31
Price sensitivity	2	5
Loyalty	5	11
New customers	27	27

Two cautions. The IPA Databank is a set of award entries, so it over-represents campaigns that worked. And the rational versus emotional split is a coding of each campaign’s dominant approach, not a controlled test. Binet and Field’s own reading: rational messaging gives a big direct effect that decays quickly, emotional priming a smaller one that builds [5, slides 13 and 14].

The industry’s own attribution work puts the weight in the same place. Nielsen Catalina Solutions decomposed the sales lift of nearly 500 FMCG campaigns run in 2016 and the first quarter of 2017 and attributed 47% of it to the creative, 22% to reach, 15% to brand factors such as price and penetration, 9% to targeting, 5% to recency and 2% to context [8]. Cannes Lions and WARC found the same from the other end: across 3,786 effectiveness award cases from 2011 to 2019, campaigns that had also won creative awards averaged 4.5 WARC effectiveness points against 2.6 for those that had not [13]. What both measure is the quality of the execution, not the presence of a claim.

Creative47%

Reach22%

Brand15%

Targeting9%

Recency5%

Context2%

Half of what advertising sells comes from the execution; the claim lives inside that half and is not the whole of it.

Source: Nielsen, When it comes to advertising effectiveness, what is key?, 2017, Nielsen Catalina Solutions, n=nearly 500 FMCG campaigns 2016 to Q1 2017, figure Percent sales contribution by advertising element. Redrawn by Brave New Media.

Unit: % of sales contribution

Show the data

Item	% of sales contribution
Creative	47
Reach	22
Brand	15
Targeting	9
Recency	5
Context	2

Where the literature is split. Reeves and Ehrenberg agree that campaigns should not change often and that memory matters. They disagree on what memory does. Reeves: the remembered claim pulls new users. Ehrenberg: the remembered brand reassures existing ones. The Bates data cannot separate the two; it never distinguished new buyers from existing buyers who remembered the ad. Sharp and Romaniuk side with Ehrenberg, in chapter 8 of How Brands Grow [9]. Binet and Field sit between: rational claims work, for a while. Kantar’s BrandZ data side with Reeves on whether difference pays and against him on how it is built [10]. No one has replicated Reeves’ 34%.

Cialdini did not address the USP. His work matters because it names what a single rational claim leaves out. Social proof (chapter 4), authority (chapter 5) and scarcity (chapter 6) in the 2021 edition of Influence move buyers without any claim that the product is unique [3]. A queue, a credential and “only three left” are propositions of a different kind, each with an experimental literature behind it.

Edelman’s 2019 brand-trust survey of 16,000 people in eight countries puts a number on what the claim leaves out. 81% said being able to trust the brand to do what is right was a deal breaker or a deciding factor in a purchase, in the same band as quality at 85%, convenience and value at 84% and ingredients at 82% [12]. The same survey found 13% had strong trust in a brand message after seeing it once and 87% after seeing it across six channels [12]. A proposition, however unique, is one message seen once.

Quality85%

Convenience84%

Value84%

Ingredients82%

I can trust the brand to do what is right81%

Trust sits four points behind quality as a condition of purchase; the same survey found strong trust in a brand message rose from 13% after one viewing to 87% after six channels.

Source: Edelman, 2019 Edelman Trust Barometer Special Report: In Brands We Trust?, 2019, n=16,000 in eight countries, press release and executive summary. Redrawn by Brave New Media.

Unit: % of respondents

Show the data

Item	% of respondents
Quality	85
Convenience	84
Value	84
Ingredients	82
I can trust the brand to do what is right	81

Where practitioners get it wrong

The slogan mistake. Reeves complained in 1961 that U.S.P. was being applied “to slogans, slick phrases, strange pictures, mere headlines” [1, ch. 13]. Sixty-five years later the output of a USP workshop is still a tagline, and a tagline the product cannot prove or disprove is not a proposition.

The uniqueness hunt. Teams spend months on a feature no competitor has, then find buyers do not value it (Reeves’ toothpaste ribbon [1, ch. 13]) or cannot perceive it (the 17% in Romaniuk, Sharp and Ehrenberg [4]). The search eats the money that distinctive assets, consistent creative and reach would have used.

The forever-rational plan. Firms that take the USP literally build every campaign as a product claim and judge it on short-term response, where rational campaigns look best (very large direct effects: rational above combined above emotional [5, slide 32]) and where the three-year profit gap is invisible. The B2B Institute at LinkedIn and System1 rated 1,700 B2B ads on System1’s five-star emotional-response scale: 76.7% scored one star, the level System1 associates with no market-share growth, and 0.5% scored four or five [11].

1 star76.7%

2 stars16.8%

3 stars6.0%

4 stars0.4%

5 stars0.1%

Three in four B2B ads produced the lowest emotional response System1 measures, the level it associates with no long-term share growth, and a B2B ad was 154 times more likely to score one star than four or five.

Source: LinkedIn B2B Institute and System1, Cashing In On Creativity, 2021, n=1,700 B2B ads in six categories, System1 star rating, p. 11. Redrawn by Brave New Media.

Unit: % of B2B ads

Show the data

System1 star rating	% of B2B ads	Share growth mapped to rating, points a year
1 star	76.7	0.0
2 stars	16.8	0.5
3 stars	6.0	1.0
4 stars	0.4	2.0
5 stars	0.1	3.0

Our own observation. In the fourth month of a residential developer engagement we stopped buying attention and went to earn it: a full editorial interview with the chief executive in a sector trade publication, secured in the annual editorial plan without paying for placement, on the seven things the market was anxious about, pricing, cost pressure, regulation, and whether developers deliver what they promise. No unique product claim anywhere in it. In a market asking whether a developer will still exist at handover, third-party endorsement did work that no proposition about the product could.

What to do with it

1. Write the claim in Reeves’ form first: “Buy this, and you will get this specific benefit.” If no testable benefit finishes the sentence, there is no proposition, only a slogan [1, ch. 13].
2. Test the benefit for importance, not uniqueness. Ask current buyers, unprompted, why they buy. If fewer than one in five name the benefit, uniqueness will not rescue it [4].
3. Expect low differentiation scores. Fewer than 20% of users calling a brand “different” is the norm across 17 markets [4, Table 2]. Put the budget into distinctive assets: name, colour, shape, sound, character, used identically everywhere [4, p. 52].
4. Keep the rational claim in activation and hold it to activation metrics. Judge brand advertising on three-year profit, where emotional campaigns lead 43% to 23% [5], and test the execution before the claim: creative carried 47% of sales lift in Nielsen’s decomposition [8].
5. Change the claim rarely and the execution often. Reeves’ one finding that has held is the cost of switching campaigns [1, ch. 8]; Ehrenberg explains why [2].
6. When the product has no unique benefit, use Cialdini’s levers instead of inventing one: visible customer numbers, credentials, real limits on supply [3, chs. 4 to 6]. Never fake the scarcity.
7. If anyone quotes Reeves’ 34%, ask for the sampling method; the book has none [1, ch. 6].