Most SaaS case studies read like a press release the customer was too polite to edit. "Acme transformed its workflow and saw incredible results." A technical buyer reads that and learns one thing: you couldn't get a number.
If you have two or three reference customers and no written proof anywhere on your site, this is how to write a SaaS case study that survives a sceptical reader. It's a how-to about your customers, so the examples below are templates, not anyone's real story.
The structure that survives a sceptical reader
Four parts, in this order. Think of it as a bug report with a happy ending: reproduction steps first, the fix second, verification last.
Before. The customer's situation in their own terms: team size, the tool or process they used, what it cost them in hours or errors. Specific enough that a reader in the same spot recognises themselves.
What changed. What they actually did with your product. Which feature, which workflow, how long setup took. Not "implemented our platform".
The number, with its period. One or two measurable results, each with the timeframe attached and how it was measured. "Month-end close went from six days to two, measured over the first two quarters after rollout" is a claim. "Faster closes" is an adjective.
Their words. A named person, with their role, saying it in their own voice. One good sentence beats a paragraph of praise.
A template for the headline: [customer type] cut [specific task] from [before] to [after] in [period]. If you can't fill in every bracket honestly, you don't have a case study yet. You have a testimonial, which is useful, but different.
Running the customer conversation
The interview is where most case studies are won or lost. Book thirty minutes, ask permission to record, and use questions that produce stories rather than compliments:
What were you using before, and what finally made you look for something else?
Walk me through the first week. What was harder than you expected?
What do you do now that you didn't do before? What stopped happening?
Is there a number your team tracks that moved, and over what period?
If a peer asked whether they should buy this, what would you actually tell them?
That last question is usually where your quote comes from. Follow up on anything vague: "a lot faster" becomes a number when you ask "faster than what, and measured how?"
Then write it up and send the customer the full draft, quotes included, for approval in writing before it goes anywhere. That isn't only good manners. Under section 74.02 of the Competition Act, publishing a testimonial is reviewable conduct unless it was already published by the person who gave it, or that person approved it and gave written permission before publication, and what you publish matches what they approved (sourced from the Justice Laws Website).
When the customer won't be named
Plenty of B2B customers can't be named, especially in regulated industries. An anonymised case study can still prove a lot if you're disciplined about it.
Describe the customer by segment, business model and stage only. "A Series A logistics software company" works.
Leave out anything searchable. A city, an investor, a distinctive integration, a named partner. Withholding the name doesn't anonymise a story only one company could have.
Keep the number and its period. Anonymity changes who, not how much. The figure still has to be real and measured.
Get written approval for the anonymised version too. Permission to share the numbers isn't the same as permission for the story around them.
The line every performance claim has to clear
In Canada, a claim about how well your product performs has to be based on an adequate and proper test, and the proof lies with whoever makes the claim (sourced from the Competition Act, section 74.01). The Competition Bureau adds that the testing has to be done before the claim is made, reflect real-world use, and support the general impression the claim creates (sourced from the Competition Bureau).
In plain terms, for a case study:
Keep the customer's data or report behind every figure on file before you publish.
Don't turn one customer's result into a promise to everyone ("cut your close time by 60 percent"). Say what happened to them.
Watch the general impression, not only the literal words. A true number under a misleading headline is still a problem.
This is marketing guidance, not legal advice. If you're unsure about a specific claim, talk to counsel before you publish.
Where the case study goes once it's written
A case study buried on a customers page nobody visits does a fraction of its job. Put the number and the quote where buying decisions happen: next to the plan it relates to, on the signup page, in the follow-up after a demo. Proof matters most on the pricing page before a demo, and it's often the missing piece when a blog gets traffic but no signups.
If that follow-up goes out by email, it's a commercial electronic message under Canada's Anti-Spam Legislation (CASL), so it needs consent, clear identification of who sent it, and a working unsubscribe.
And if writing the case study shows you that you can't say in one line what problem you solved for that customer, the gap isn't the case study. That's positioning and messaging work, and it's worth fixing first, because every future case study will hang off it.
Which of your customers could you call this week, and what number would they say moved?