Writing a case study a buyer will actually finish

A case study is the only sales document that works while you are asleep. It gets forwarded, pasted into an internal paper, read on a phone between two meetings. Most of the ones I am asked to review cannot survive that trip. They open with the vendor's name, run to twelve hundred words, describe a partnership, and put the result in the last paragraph where almost nobody reaches it.

The fix is not better writing. It is accepting that your reader is skimming, and building the document so a skim still lands.

You are competing with a skim, not with another vendor

In 2008 Jakob Nielsen analysed 45,237 page views captured from instrumented browsers and concluded that readers have time to read at most 28 percent of the words on an average page, and that about 20 percent is the realistic figure. Readers got through half the words only on pages of 111 words or fewer.

That work is eighteen years old and it measured general web reading rather than buyers reading proof. I use it as direction, not as measurement. The direction is enough: assume four words in five go unread, then ask what the one in five has to contain.

Lead with their number, not your name

The first thirty words decide whether the rest gets read, so put the outcome in them. Not the sector, not the background, not your product name. "A Hamilton logistics firm cut quote turnaround from three days to four hours" is a headline. "Helping ACME transform their customer journey" is a label on a folder.

Your name belongs at the bottom. The buyer already knows who sent it.

The one number that has to be in it

Every case study needs one figure the customer said out loud, with the measure named and the period stated. Before and after, over a stated timeframe, in a unit they recognise: hours, dollars, days, or percentage points off a base you also give.

"Sales grew 40 percent" is not that figure. Forty percent of what, over how long, against what trend. A buyer who has read three of these can tell a measured result from a flattering one, and so can the finance person they forward it to.

If your customer will not say a number, you do not have a case study yet. You have a testimonial. Publish it as one, call it a quote, and keep looking for a customer who will.

What does your buyer do with it after the meeting?

Forwards it. Forrester reported in December 2024 that an average of 13 people inside an organisation are involved in a buying decision, and that 89 percent of purchases involve two or more departments. Most of those people will never be in a room with you.

So the document has to work with nobody presenting it. Name the problem in the customer's language rather than yours. Describe what you did plainly enough that an operations manager recognises their own week in it. Give the number with its measure. Anything that needs you standing beside it to make sense should come out.

This is also why a case study belongs to whoever owns revenue rather than to whoever owns the website. The question it answers is a deal question: what does this buyer need to believe, and which proof moves them? A marketing team can write it beautifully and still answer a different question well.

A shape that survives two minutes

Six parts, 400 to 600 words, one page:

• A headline carrying the result.
• One line on the customer, specific enough that a similar buyer recognises themselves.
• The problem in their words, two sentences, including what it was costing them.
• What you did, three short points, no feature names.
• The number, with its measure and its period.
• One quote that says something only a customer would say.

If a section is there because the customer asked for it rather than because a buyer needs it, cut it and tell them why.

In New Zealand the name does half the work

Here a named customer is worth three anonymous ones. The market is small enough that a buyer either knows the company or knows somebody who does, and that recognition carries more weight than any adjective you could add to it.

Which makes consent part of the work rather than an afterthought. Ask when the result lands and the customer is pleased, agree in writing what may be named, and let them read it before it goes anywhere. Keep their pricing and contract terms out of it entirely. The easiest moment to secure reference rights is before you have discounted anything, by making them part of what the discount buys.

Open your best case study and read only the first thirty words and the last line. If a buyer read nothing else, would they know what changed, by how much, and over what period?

Sources. Reading behaviour: Jakob Nielsen, "How Little Do Users Read?", Nielsen Norman Group, 6 May 2008, which analyses 45,237 page views (reduced from 59,573 after removing visits under four seconds, visits over ten minutes and pages under 20 words) collected in 2005 from the instrumented browsers of 25 users by Harald Weinreich, Hartmut Obendorf, Eelco Herder and Matthias Mayer, "Not Quite the Average: An Empirical Study of Web Use," ACM Transactions on the Web, volume 2, number 1, February 2008, article 5. Nielsen writes that users "have time to read at most 28% of the words during an average visit; 20% is more likely," and that readers read half the information only on pages of 111 words or fewer. That study is eighteen years old, rests on 25 above-average readers, and measures general web use rather than buyers reading sales material, so I use it as direction rather than as a current measurement. Buying group size: Forrester, "To Master B2B Buying Mayhem, Providers Must Prioritize Buyers' Needs," press release, 4 December 2024, which reports that an average of 13 people in an organisation are involved in the buying decision and that 89 percent of purchases involve two or more departments. Forrester published no sample size or fielding dates with that release, so treat both figures as the firm's own reported findings rather than as something independently checkable. One figure was considered and cut: the widely repeated claim that B2B buyers spend only 17 percent of their buying time with potential suppliers is attributed to Gartner across dozens of secondary articles, but I could not find it stated on a page Gartner publishes, so it is not in this article. The six-part structure, the thirty-word test, the advice on trading reference rights against a discount, and the observation about named customers in a market this size come from my own work with New Zealand tech founders rather than from published data.

Nick Burns is a fractional CRO for New Zealand B2B tech companies, at home and expanding internationally. He co-founded Emendo and sold it to McKesson, then the 14th-largest company on the Fortune 500, and has since helped 60+ tech companies grow sales.