Which Customers to Ask for a Case Study: A 6-Point Scoring Checklist

Ask the customer who looks most like the buyer you want next, who can describe a before and after in their own words and who is willing to be named. Score each candidate with the six-point checklist below and ask only those who reach 8 out of 12.

Why is "ask your happiest customer" the wrong rule?

Because happy is not the same as useful. A story works when a prospect reads it and thinks "that is us". Your happiest customer is often your oldest account, your smallest or your most unusual, and nobody shopping for the next deal sees themselves in them. The same goes for your biggest logo if it sits in a segment you are not selling into.

The question is not who likes you. It is who is most likely to produce a story that a specific buyer will believe. That is a selection decision, and it deserves a scorecard instead of a gut feel.

If you are deciding how to get the yes once you have picked, getting customers to agree covers that. This post is about who goes on the list.

What does a good case study candidate look like?

Look for six things. Each one is scored 0, 1 or 2.

Add the six scores. 8 or more: ask this month, unless signal 1 scores 0. A 0 on signal 1 is a veto, whatever the total. 5 to 7: wait for the next win, then ask. Under 5: do not ask yet.

A few notes on the table.

Signal 1 outweighs the rest. If your pipeline is mid-market finance teams, a story from an agency with twelve staff is not proof for that deal, however enthusiastic they are. The guide to how vertical SaaS gets first case studies takes this to its conclusion: in a single trade, matching the buyer's business is nearly the whole job.

Signal 2 does not require a number. A customer who can say "we used to rebuild the schedule by hand every Sunday night and now we don't" has a story. The number helps, but a case study without metrics can still carry a deal. What you cannot use is a customer who has no before.

Signal 3 is a gate, not a bonus. Check early whether the customer's company allows being named. Regulated firms (banks, hospitals, public bodies) often route this through legal, so ask the champion who signs off before you spend an interview on it. If it does not, the question becomes whether a described-but-unnamed story is still worth the effort, and for the right lookalike it often is. The post on proof without a name covers how to write it.

Signal 4 is easier when the story helps the champion too. A champion who wants visibility, with their own leadership or on LinkedIn, will chase their own approvals for you.

Signal 5 is the one teams ignore. A customer who just hit a goal is eager to talk about it. A customer who just filed a ticket is not. Ask within a few days of a win.

What does scoring look like on real candidates?

Take three invented customers of a project-management tool whose pipeline is mid-size construction firms.

• ●A 15-person design studio, delighted, replies in minutes, happy to be named. Lookalike 0, result 1, named 2, champion 2, timing 1, tenure 2: 8. It reaches 8 only because the studio is easy to work with. The zero on signal 1 is the real answer: it will not help win a construction deal, so ask it for a quote, not a story.
• ●A 120-person contractor, the exact buyer, with a site manager who can say "we cut rework tickets from about 40 a month to about 15". Lookalike 2, result 2, named 1 (logo only), champion 1, timing 2, tenure 1: 9. Ask first.
• ●A 300-person builder, same segment, but in the middle of a renewal negotiation. Lookalike 2, result 1, named 1, champion 1, timing 0, tenure 2: 7. Wait until the renewal signs.

The checklist does not replace judgement. It stops the easiest customer outranking the most useful one, and it gives you a reason to show the colleague who keeps suggesting the friendliest account.

What does the bar look like in published stories?

Customers who clear the checklist can give you what published case studies usually contain. In StoryVoice's Customer Proof Index 2026, a crawl of 1,000 B2B software company websites, we sampled stories from the 335 companies that had at least one, and found that 84.8% of them named the customer, 83.6% included a direct quote, and 89.3% had a quantified result somewhere. Only 10.7% contained no number at all. The full Customer Proof Index has the method, and it is our own study of one directory's listings, so read it as directional.

Read that as your minimum. Each of those three is present in more than 8 in 10 of the companies with readable stories, so treat a name, a quote and one clear result as your minimum. If a candidate cannot give you those three, either through the interview or through their own words, they score low on signals 2 and 3 and should wait.

Which customers should you skip for now?

Some customers are good candidates at the wrong moment. Others are not candidates at all.

None of these is a permanent no. The point is to stop spending asks on customers who are likely to be distracted, defensive or unable to describe a result.

How do you build the shortlist?

Treat it as a funnel, not a wish. Ask for twice as many stories as you want, because replies thin out at every step. Teams that scale case study production keep a standing list instead of choosing a name each time someone asks for proof.

Here is a shortlist template. Copy it into a spreadsheet and fill one row per customer.

Then do three things.

1. ●Sort by total, then by lookalike. Ties go to the customer closest to your pipeline.
2. ●Ask the top six. As a rule of thumb, six candidates for three stories allows for the replies that never come.
3. ●Re-score monthly. Timing moves. A 6 today becomes a 9 the week after a big win.

Ask the champion directly, say who will see the story, and keep the request short. If you want a template to adapt, the case study request email tool writes one from your details.

What happens after they say yes?

Once someone on your shortlist agrees, the work shifts from deciding to producing. With StoryVoice you enter the customer's details and the outcomes you want covered, and the voice AI interviews them in about five minutes on their phone. It writes the case study, sends it to them to approve by link, and makes four assets from the same interview: a LinkedIn post, an X post, a pull quote and a sales email. You publish, and a single story costs $39. The details of interviewing without a call are in a separate post.

Because each story is $39 (or $99 for three, $149 for five, credits that never expire), you can afford to run the whole shortlist instead of picking a single hero customer. The scoring decides the order, not whether to try.

How do you decide, in one line?

• ●Pick the customer who looks like the deal you want next, if you can only ask one.
• ●Pick the customer with the cleanest before and after, if two score the same.
• ●Pick the one who can be named, if you are still tied.
• ●Wait for a win if the score is 5 to 7 and the timing is poor.
• ●Skip for now if the score is under 5, or the customer is mid-dispute.

Send your first interview link. Interview yourself free to see the finished story, then $39 for each real customer story.

---

Customer Proof Index figures are StoryVoice's own research (1,000 B2B software company websites, crawled 2026-08-21, 954 usable, 335 with at least one story), published at storyvoice.io/research/customer-proof-index-2026, which we opened on 2026-10-16. The 6-point scoring checklist is our editorial method, not a study. Pricing is from storyvoice.io/pricing. Last updated 2026-10-16.