Skyfliq
← All posts

How to Write a Case Study That Actually Sells

Most case studies read like press releases and never get used. Here is how to pick the right customer, run the interview, structure the story, and make sure sales actually sends it.

Ask ten companies for a case study and nine will send you a press release. Logo at the top, a paragraph about how thrilled everyone is, a quote that could have been written about any vendor, and a number with no context. Nobody reads them. Sales does not send them. They sit in a resources folder, quietly costing the money they took to produce.

Knowing how to write a case study that sells means doing something narrower and harder: letting a prospect recognise their own situation in someone else's, then showing exactly what changed. That is the whole job. This guide covers how to choose the right customer, run the interview, structure the story, and get it used — with the parts most teams skip.

Why most case studies fail

Three reasons, in order of how often they turn up.

They are about you, not the customer. The subject of a good case study is the customer's problem. Your product is a supporting character that shows up in act two. When the first paragraph is about your platform's capabilities, the reader stops being interested — they are not shopping for a hero.

The numbers are unbelievable. "412% increase in productivity." Nobody believes it, and worse, the exaggeration makes the credible numbers next to it look invented too. A modest, specific, verifiable figure outperforms a spectacular vague one every time.

There is no before. Plenty of published case studies describe an outcome with no starting point. "Now processes 3,000 orders a week" means nothing until you know they were doing 900 and drowning. The gap is the story. Without the before, there is no gap.

The reader is not asking "is this product good?" They are asking "is this person me?" Everything in the case study should help them answer.

Picking the right customer

The instinct is to ask your happiest customer. Better instinct: ask the customer whose starting position most resembles your best prospects.

Look for four things.

  • Recognisable situation. Same size, same industry, same constraint as the buyers you want. A 400-person enterprise story does not help a ten-person team see themselves.
  • A real before. They had a problem with a cost attached. Customers who were "already doing fine and just wanted to try something new" make weak stories.
  • Numbers they will let you publish. Not necessarily revenue — hours saved, response time, conversion rate, headcount avoided. Anything measurable and specific.
  • Someone who will actually talk. A genuinely enthusiastic champion beats a bigger logo with a nervous comms team. A famous name with three rounds of legal review often produces something so sanded down it says nothing.

If you are early and have no customers of the right shape yet, write the story you do have and be honest about scale. A five-person agency story told properly is more persuasive to five-person agencies than a Fortune 500 story with the details removed.

Ask at the right moment

Right after a win is when people say yes. A successful launch, a renewal, a support ticket that ended with "honestly this has been brilliant" — that is the moment. Six months later they have forgotten what it was like before, which is exactly the part you need them to remember. Our guide to getting testimonials covers the timing and the ask in more detail; the same windows apply here.

Skyfliq Inbox

Every comment and DM in one inbox.

Reply across all your platforms from a single unified inbox. Free for 15 days, no card required.

Unify your inbox →

The interview is the whole thing

Everything good in a case study comes out of the interview. Everything dull comes from writing it from your own CRM notes.

Thirty to forty-five minutes, recorded, with permission. Do not send questions in advance — you will get prepared answers, and prepared answers are where the specificity disappears.

Questions that produce usable material:

  • Walk me through what a typical week looked like before. What was the most annoying part?
  • What made you start looking? Was there a specific moment?
  • What else did you consider? What almost made you pick something else?
  • What was the first thing that changed after you switched?
  • What surprised you — good or bad?
  • What would you tell someone in your position a year ago?
  • Is there anything you would still like to be better?

That last one matters more than it looks. A case study that admits one limitation is dramatically more credible than one that does not, and prospects who see an honest caveat trust the rest of the numbers. It also flushes out product feedback you would not otherwise hear.

What to listen for: the moment they stop giving you the official version and start telling you what actually happened. It usually starts with "honestly" or "to be fair." Those are the quotes you use.

Get the numbers before you write

Ask for the specific measurements during the interview, and ask what they were before. If they do not have the before figure to hand, ask what they remember — "roughly two days to respond, sometimes three" is publishable and honest. An estimate labelled as an estimate is fine; a precise-sounding number you invented is not.

A case study structure that works (use it as your template)

Six sections, in this order. It maps to how a skeptical reader actually evaluates a story.

1. The one-line result, at the top

Before anything else: who, what changed, by how much. "A twelve-person recruitment agency cut time-to-first-response from two days to under four hours — and doubled interview bookings without hiring." Most people read this and nothing else. Make it carry the whole story.

2. Who they are

Two or three sentences. Size, sector, what they sell, why their situation is distinctive. Enough for the reader to place themselves next to it.

3. The before, with a cost attached

The most important section and the one most often rushed. What was happening, why it was a problem, and what it was costing — in hours, money, missed deals or staff frustration. Use their words. This is where the interview earns its keep.

4. What they tried and why they chose you

Include the alternatives. A case study that pretends there was no competition reads like a brochure. "They trialled two other tools; one was too complex for a team without an ops person, the other could not handle their approval flow" tells a prospect more about fit than any feature list, and pre-answers an objection they were already forming.

5. What actually happened

The change, in sequence. What went live first, what took time, where it was harder than expected. Include the friction — "the first two weeks were messy while we migrated old records" makes everything after it believable.

6. The result, with numbers and a caveat

Specific figures, the time period they cover, and an honest note on what has not changed. Then one forward-looking line about what they are doing next.

Length, format and the version nobody makes

Around 800–1,200 words is the sweet spot for a written case study. Long enough for the story, short enough to finish.

But the written page is the least-used format. Make three versions from the same interview:

  • The one-pager PDF. What sales actually sends. Result at the top, three paragraphs, one quote, one chart.
  • The web page. Indexed, linkable, with the full story and a clear next step.
  • The 60-second cut. Three or four sentences of the customer talking, with captions. This is what travels on social. If the interview was recorded on video, you already have it.

Most teams make only the web page, which is why case studies end up unused. Sales does not send links to a blog; they send a PDF. Build the one-pager in the same sitting or it will never get made.

Skyfliq Analytics

Finally know what’s actually working.

Track reach, engagement, and ROI across every channel in one dashboard. 15-day free trial, no card required.

See your analytics →

Getting approval without losing the story

The approval round is where good case studies get ruined. Two things reduce the damage.

Send the draft with the quotes highlighted. Reviewers focus on what they said, which is what they care about, rather than rewriting your narrative.

Ask for factual corrections specifically. "Please flag anything inaccurate or anything you cannot publish" gets a different, much lighter edit than "let us know your thoughts," which invites a rewrite by committee.

If legal strips the numbers out, negotiate for ranges or relative figures — "cut response time by more than half" survives review more often than the exact hours, and still does the job.

Making sure it gets used

A published case study with no distribution plan is a document, not an asset. Build the rollout into the project:

  • Tell sales it exists, with one line on which deal stage and which objection it answers. A case study nobody can find in the moment they need it might as well not exist.
  • Link it from the pages where doubt happens — pricing, the relevant feature page, the comparison page. Doubt is where proof belongs.
  • Cut it into social posts over the following weeks. One number, one quote, one before-and-after. Each stands alone — see social media storytelling for how to frame them.
  • Put it in the nurture sequence for prospects who match that customer's profile. Relevance beats volume here — a matching story sent to twenty right people outperforms a generic one sent to two thousand.
  • Refresh the numbers annually. A case study with 2024 figures on it in 2026 reads as abandoned. A one-line update — "as of this year, they are at 6,000 orders a week" — buys another year of credibility for ten minutes of work.

Four openings, and which one to use

The first sentence decides whether the rest gets read. Four that work, roughly in order of how often they fit:

Open on the frustration. "Every Monday, Priya's team spent the first three hours of the week copying order details between two systems." Concrete, immediately recognisable, and the reader knows within a line whether this is their story.

Open on the number. "Four hours to two hours to eleven minutes." Best when the result is genuinely striking and the metric needs no explanation.

Open on the decision. "They had two weeks to choose, and the obvious option was the one they did not pick." Good when the buying process itself is the interesting part — useful in categories where prospects are comparing several vendors.

Open on the quote. Only if the quote is genuinely good. "I was ready to go back to spreadsheets" earns the spot. "We're delighted with the partnership" does not.

What to avoid: opening with the customer's founding year, their headcount, or a sentence that begins "In today's competitive landscape." Those are throat-clearing, and readers scroll past throat-clearing.

Skyfliq for Teams

Plan, approve, publish — together.

A shared calendar, approvals, and roles so your whole team ships on time. 15-day free trial, no card required.

Try Skyfliq free →

How many do you need?

Fewer than you think, but spread across the situations your buyers are in. Three case studies covering your three main buyer types beats twelve covering one. The prospect only needs to find themselves once.

If you can only do one this quarter, pick the customer whose before-state matches your most common objection. That is the one that will get sent the most.

The short version

Choose a customer who looks like your prospects, not the one with the biggest logo. Interview them properly and listen for the moment they stop being diplomatic. Lead with the result, spend real time on the before, admit one thing that is still imperfect, and make the one-pager the same day.

Do that three times and you will have something more persuasive than anything you can write about yourself.

Related guides: How to Write a Blog Post That Ranks on Google

Frequently asked questions

How long should a marketing case study be?

Around 800 to 1,200 words for the written version. Long enough to tell the before-and-after properly, short enough that people finish it. Make a one-page PDF and a sixty-second video cut from the same interview — the PDF is what sales actually sends.

What should a case study include?

Six parts: a one-line result at the top, two or three sentences on who the customer is, the before state with a cost attached, what else they considered and why they chose you, what actually happened including the friction, and the result with numbers plus one honest caveat.

How do I get customers to agree to a case study?

Ask right after a win — a successful launch, a renewal, or a support conversation that ended well. Six months later they have forgotten what the before state felt like, which is the part you most need them to remember.

Should a case study mention competitors?

Yes, generally. Including what else they trialled and why it did not fit tells a prospect more about whether you suit their situation than any feature list, and it pre-answers an objection they were already forming.

What makes a case study believable?

Specific modest numbers beat spectacular vague ones, a clearly stated before state, and one honest limitation. A case study that admits something is still imperfect makes every other claim in it more credible.

Ready to grow faster?

Start your free trial of Skyfliq — no card required, cancel anytime.