What Is a Value Proposition?

The term appears in an internal McKinsey staff paper from June 1988, "A business is a value delivery system", by Michael J. Lanning and Edward G. Michaels. McKinsey published an adapted version in the McKinsey Quarterly in 2000. There, the value proposition is a clear, simple statement of the benefits the company will provide to each customer segment, along with the approximate price it will charge for them. The authors add a condition: all of the company's customers should see significantly more benefit than they are asked to pay.

That definition yields four parts a good value proposition names. First, who it is for: a specific segment, not "all businesses". Second, what problem or job that customer has. Third, what benefit they get. Fourth, against which alternative: the product they use today, a competitor, or doing nothing. A 1979 McKinsey staff paper by John L. Forbis and Nitin T. Mehta puts it this way: to get customers to switch to your product, you have to give them at least as much value beyond the price as they receive from the product they currently use.

A value proposition is not a slogan either. The Cambridge Dictionary defines a slogan as "a short easily remembered phrase, especially one used to advertise an idea or a product", and a tagline as a short, easily remembered phrase that a company uses in its advertisements so that people recognise it or its products. Neither definition requires saying who the product is for or what it competes against. An invented example: "Invoicing has never been easier" is a slogan. "For freelancers who bill from their phone: send a complete invoice without opening a spreadsheet or waiting for your accountant" is a value proposition. The slogan can come out of the proposition, but it does not replace it.

James C. Anderson, James A. Narus and Wouter Van Rossum described a specific failure in the Harvard Business Review in March 2006, based on their research in business markets. They found no agreement on what constitutes a customer value proposition. And, they found, most value propositions claim savings and benefits without backing them up. If the supplier does not demonstrate and document the claim, they write, a customer manager will likely dismiss it as marketing puffery.

One tool for working on a value proposition is the Value Proposition Canvas. Alexander Osterwalder, Yves Pigneur, Gregory Bernarda and Alan Smith published it in the book Value Proposition Design (Wiley, October 2014), and Strategyzer offers it as the official tool "from the creators". The canvas has two halves. The customer profile describes one segment: the jobs they are trying to get done, the pains they hit along the way and the gains they are after. The value map lists what is offered and, for each item, which pain it relieves and which gain it creates.

The canvas organises assumptions; it proves nothing. Strategyzer says so itself on the tool's page: fit is a claim until customers confirm it, and the matches you drew have to be tested against evidence. A neatly completed canvas from a meeting room only shows what the team believes about the customer.

Testing takes two steps, and each measures something different. The first is customer interviews. According to the Nielsen Norman Group, they show what users need and value, but they collect reported rather than observed behaviour, with gaps in memory and the wish to look good. Jakob Nielsen summed it up in 2001: pay attention to what users do, not what they say, and their speculation about future behaviour is unreliable. So it is more useful to ask when someone last had the problem and what they did about it than to ask "would you buy this?". That gives you the segment, the problem and the real alternative, plus the words the customer uses to describe it.

The second step measures behaviour. An A/B test of the headline sets two wordings of the value proposition against each other on the same page, with traffic split at random and one primary metric fixed in advance, for example demo requests. The test measures which wording gets more visitors to act; whether the promise is true is settled by the product. For the result to mean anything you need enough traffic; with few visits, the interviews remain the main evidence.

The value proposition is the sentence visitors need to understand first. On a landing page its place is the headline, above the fold, and it has to match the promise of the ad or search result the visitor came from.

According to Google's documentation, the snippet shown in search results is determined primarily from the content of the page; the meta description is used sometimes, when it describes the page better than the text does. A value proposition that only exists in the meta description may be left out of the snippet. It has to be in the visible text of the page.

Lanning and Michaels pointed to an internal consequence: the chosen value proposition has to echo through the whole business, including advertising, which should communicate it instead of pursuing a potentially award-winning but irrelevant idea. For CRO, that means a new headline does not fix a value proposition the product does not deliver.

• Name a specific segment, built on data rather than assumptions, for example from a buyer persona.
• Write down the alternative the offer competes with, even if it is "keep using the spreadsheet".
• Use in the headline the words customers used in the interviews.
• Test with interviews first and an A/B test second, and record which part of the proposition each step tested.

• Confusing the value proposition with a slogan or tagline.
• Writing it for "everyone", without a segment.
• Promising benefits you cannot prove.
• Treating a completed Value Proposition Canvas as proof of fit.
• Asking in interviews whether someone would buy the product.