How do I write a customer case study that sales will actually use?

1. Blog
2. Discoverability & Content

By Saroj Jha · November 2, 2026 · 14 min read

Sales uses a case study when it answers the question a stalled buyer is asking: has this worked for a company like mine, against the option I am comparing you with? So I choose the customer to fill a gap in the current pipeline, not because they are the happiest, and I get permission for each element in writing before drafting. The story starts with the problem in the buyer’s own words, names the alternative they considered, and states one measured result with its baseline, period and definition. Then I cut that one interview into formats for each stage of the deal and put it where sellers already work.

The rest of this post follows one fictional company from choosing the customer to the sales call.

Running case: Example Co. Fictional

Example Co. is a fictional seed-stage B2B SaaS company that automates month-end close for finance teams at companies with 200 to 2,000 employees. It has 18 customers and one published case study, about a manufacturer. Its 40 open opportunities split into logistics (14), manufacturing (12), retail (9) and other (5), and its sellers keep hearing the same objection from logistics buyers: “Your one story is about a factory. We are not a factory.” Every Example Co. number here is fictional and its arithmetic is shown.

Which customer and which result should the story be about?

Most founders pick the happiest customer or the best-known logo. I pick the customer whose story answers the question sales cannot answer today. That starts with the pipeline, not with the customer list: which segments, sizes or use cases are open deals in, and which of them have no matching proof?

Similarity matters more than fame. A buyer at a 400-person logistics firm learns little from a 10,000-person bank, however impressive the logo. The Customer Advocacy, Reviews and NPS post covers how to build the pool of references and case-study candidates; this post starts once you have a few willing customers to choose from.

The result has to be one the customer can show. I prefer a single operational number the buyer already tracks, such as days to close the books or hours of manual work, over a revenue claim the customer cannot isolate. If you did not record the “before” number when the customer started, the result will be a recollection, and recollections do not survive a finance buyer’s questions.

Example Co., step 1 Fictional

The one manufacturing story covers 12 of 40 open opportunities, 30%. A logistics story would add 14, taking coverage to 26 of 40, 65%; a retail story after that would reach 35 of 40, 87.5%. Example Co. chooses a 400-employee logistics customer that recorded its close time before go-live.

Do this

Tag every open opportunity by segment and main use case, mark which ones already have a matching case study, and choose your next customer to cover the largest uncovered group.

How do you get permission and stay within the FTC’s rules?

Ask for permission in pieces, not as one yes. A customer may happily approve the result and the quote but not the company name, the logo or the named person. I send the draft with a list of each element (name, logo, person’s name and title, photo, quote, each number) and record who approved which item and when. The Case Study Kit has the permission email and the consent record.

In the US, a customer quote in your marketing is an endorsement under the FTC’s Endorsement Guides, 16 CFR Part 255, last revised in 2023. Four rules matter most for case studies. Endorsements must reflect the honest opinions, findings, beliefs or experience of the endorser, and while you need not use the customer’s exact words, an endorsement may not be presented out of context or reworded so as to distort the endorser’s opinion or experience^[3]. If you lack substantiation that the customer’s result is representative, the ad should clearly and conspicuously disclose the generally expected performance, and the Guides say a “results not typical” disclaimer is insufficient on its own^[3].

The third rule is about what you gave the customer. A connection the audience would not expect, which can include a business relationship, must be disclosed clearly and conspicuously^[3]. The FTC’s own question-and-answer page says that if customers were told in advance they might be featured in an ad, or were promised something for it, you should disclose that fact^[4]. A discount offered in exchange for appearing in a case study is exactly such a connection, so I either do not offer one or disclose it on the page.

The fourth rule is newer. The FTC’s Rule on the Use of Consumer Reviews and Testimonials, 16 CFR Part 465, makes it a violation for a business to create a testimonial that misrepresents that the testimonialist exists or used the product, and requires a clear disclosure when an officer, manager, employee or agent gives one, if that relationship is not otherwise clear to the audience and the business knew or should have known about it^[5]. The FTC announced the rule in August 2024 and said it lets the agency seek civil penalties against knowing violators^[6]. These are US rules; if you sell in the EU or the UK, check the local rules on endorsements and reviews as well.

The current maximum under Section 5(m)(1)(A) of the FTC Act is $53,088 per violation^[2], and the FTC left its penalty amounts unchanged for 2026^[7]. A composite customer presented as one real company falls on the wrong side of this rule. This is orientation, not legal advice.

Example Co., step 2 Fictional

The logistics customer approves its name, the controller’s quote and the close-time numbers, but not its logo or the controller’s photo. Example Co. offered no discount for participating, so it has no material connection to disclose; it does disclose, beside the result, the typical result across customers (see step 4).

Do this

Send the draft with a numbered list of every element you want to use, get a written yes or no on each, and keep that record with the published file.

What should the story actually say?

A case study that sells reads like the buyer’s own situation, then shows the way out. I write it in four parts, and each one answers a question a buyer asks on a sales call.

The problem, in the buyer’s words

Open with what the customer said before they bought, in their language rather than yours. “Our close took twelve working days and the controller lived in spreadsheets for half of every month” lands with a peer; “struggled with financial operations inefficiencies” does not. The words come from the interview, so record it and use the quote as said, within the rewording limit above.

The alternative they considered

Every buyer compares you with something: a competitor, an internal build, a cheaper tool, or doing nothing. Say which one this customer weighed and why they chose you. This is the part sellers use most, because it answers the objection they are hearing now. The competitive analysis you already have tells you which alternatives matter.

The change and the result

Describe what changed in the customer’s work, in a few concrete steps, then give the result. Keep the product’s features in service of the change; a feature list belongs on the product page.

The person and the proof

End with who the buyer can believe: a named role, a quote and, where approved, an offer of a reference call. A named controller at a named company is worth more than three anonymous paragraphs.

Example Co., step 3 Fictional

Problem: “Close took twelve working days, and I spent half of every month reconciling spreadsheets,” in the controller’s words. Alternative: hiring a second accountant, budgeted and then cancelled. Change: bank and freight-invoice matching moved into the product, with exceptions routed to one reviewer. Proof: the controller’s name, title and quote, and an offer of one reference call a quarter.

Do this

Before you write a word, fill in four lines from the interview notes (problem quote, alternative, change, result) and do not draft until all four are filled.

How do you state a result a buyer will believe?

A result needs four things beside the number: the baseline, the period, the definition and what else changed. “Cut close time by 47%” invites three questions; “average close fell from 12 to 6.3 working days across the first three month-ends after go-live, measured from period end to final sign-off” answers them before they are asked.

I state means over several periods rather than the single best month, and I say what else happened at the same time, such as a new hire or a system change. Finance and operations buyers check numbers, and one inflated figure discredits the whole page. This is also where the 69% figure bites^[1]: if the website, the PDF and the seller’s slide give different numbers for the same customer, buyers notice.

Then decide whether the result is typical. One customer’s result is one data point. If you have results for other customers measured the same way, you can state the typical result next to the featured one; if you do not, the Endorsement Guides’ typicality rule^[3] means a single headline result needs care.

Example Co., step 4 Fictional

Before go-live, the last three closes took 13, 12 and 11 working days, a mean of 12. After, they took 7, 6 and 6, a mean of 6.33. The cut is 12 − 6.33 = 5.67 days, or 5.67 ÷ 12 = 47.2%. Across the 11 customers with a recorded baseline, the median cut is 3 days (range 0 to 5.67), so the page says: “Customers with a recorded baseline cut their close by a median of 3 working days.”

Do this

Record the baseline, period and definition for every new customer at kickoff, so the next case study starts with real numbers.

Which format works at which stage of a deal?

One good interview gives you six or more assets. The mistake is to publish only the two-page PDF and expect it to work everywhere. Buyers early in their research want a line of proof; a buyer writing a business case wants the numbers and the method.

Gartner’s 2024 survey found that 61% of B2B buyers prefer an overall rep-free buying experience^[1], so most of the early proof has to work without a seller present. That argues for proof on the pages buyers visit alone (the homepage, the product and pricing pages, review profiles) and for a version a champion can forward inside their company without editing. Review sites play a related role, covered in Review Sites for Startups.

My usual set, from one interview, is below. I write the full story first and cut the others from it, so every number matches.

Example Co., step 5 Fictional

From one 45-minute interview Example Co. publishes a one-line proof point (“A 400-person logistics firm cut its close from 12 to 6.3 days”) for the homepage and ads, a 150-word summary for the logistics landing page and outbound emails, the full web page, a two-page PDF for champions, one slide for the sales deck, and a reference-call offer for late-stage deals.

Do this

Write the full case study once, then cut the proof point, the summary and the slide from it the same day, and check that each uses the same numbers.

How does sales actually use a case study?

Sellers use what they can find in ten seconds during a deal. If the case study lives only on the website’s resources page, it will not be used. Put each version where sellers work, tagged by segment, use case and the alternative it beats, so “logistics, replacing a hire” returns the right story.

Teach the moment to use it. In my view, the best moment is right after a buyer names an objection or an alternative on a discovery call: the seller sends the matching story with one line on why it fits. The Discovery Call Framework post covers how to surface that objection, and the Sales and Marketing Alignment post covers the shared definitions that keep the numbers consistent.

Measure use, not views. Log which opportunities received a case study and compare outcomes, while remembering that sellers choose when to share, so a difference in win rates shows association, not cause.

Example Co., step 6 Fictional

Over the next quarter, sellers logged the logistics story as shared in 15 new opportunities across segments, 6 of which closed won (6 ÷ 15 = 40.0%), against 4 won of 17 where no case study was shared (4 ÷ 17 = 23.5%). Example Co. reads this as a reason to keep going and to record why each deal was won, not as proof that the story caused the wins.

Do this

Add a “case study shared” field to opportunities in your CRM and review it with sales once a quarter alongside win-loss notes.

What changes by stage?

• Pre-seed — With a handful of design partners, a named quote and an honest description of the pilot is enough; avoid percentage claims until you have baselines. Put the kickoff permission clause in every pilot agreement now.
• Seed — Write one full case study per main segment in the pipeline, with a recorded baseline, and cut the formats from each. State the typical result once you have several customers measured the same way.
• Series A–B — Build a tagged library by segment, use case and alternative, refresh stories older than about two years, and review case-study use against win-loss findings each quarter.

Free template

Run the permission and baseline steps from one kit

The Case Study Kit gives you the kickoff permission clause, the baseline record, the results interview, the evidence worksheet, the permission email and the consent record used in this post.

The program that feeds case studies, from reviews to references, is in Customer Advocacy, Reviews and NPS, and the step-by-step build of a proof library is in the Customer Proof and Advocacy Playbook. If you want to check each claim before it goes live, the Claims & Proof Checker is the place to start.

If you want me to look at your own funnel, positioning and channels, my Growth Audit ends with three prioritized fixes and a 30-day action plan.