How to Conduct Customer Reference Calls During Due Diligence

Most investors treat customer reference calls as a formality. You call the three names the founder gave you, they say nice things, you check the box and move on. That's how you get burned.

Done right, customer references are the highest-signal part of your entire due diligence process. A 30-minute call with the right person will tell you more than 50 hours of market research.

Here's how to actually run them.

Who to Call (and How to Get the Real Names)

Start by asking the founder for 5-8 customer references — not 3. You want enough to spot patterns. Tell them you need a mix: power users, occasional users, and at least 2 churned customers or prospects who didn't close.

That last request is the tell. A confident founder who believes in their product will give you churned customers without flinching. A defensive founder will push back, offer excuses, or conveniently "not have those contacts anymore." Take note.

Don't stop with the list they give you. Do your own sourcing:

• Search LinkedIn for people with titles like "VP Operations" or "Head of Finance" at companies the startup counts as customers
• Check the company's case studies, press releases, and testimonials — those people are fair game
• Ask each reference: "Who else at your company uses this? Can I speak with them?"
• Ask your network if anyone knows customers in this space

The references a founder doesn't give you are often more valuable than the ones they do.

Setting Up the Call

Keep it 30 minutes. Tell them it's a quick reference call — most people will say yes to 30 minutes when they'd say no to an hour.

Do not send questions in advance. You want natural, unconsidered responses. Pre-sent questions give them time to craft polished answers that sound like marketing copy.

Open with: "I'll be honest — I'm trying to understand the real experience of using [product]. Not just what's on the website, but what it's actually like day-to-day. I'd love your candid take."

That framing matters. It signals you're not there to validate a press release.

20 Questions That Get Real Answers

Product fit and value:

1. What problem were you trying to solve when you found them?
2. What else did you evaluate? Why did you choose this over alternatives?
3. How long did it take to go live? What did implementation actually look like?
4. What's the one thing this does better than anything else?
5. What do you use it for most? What do you barely use?

Retention signals:

1. How embedded is this in your workflow? What would break if you turned it off tomorrow?
2. Have you expanded usage since you started? What drove that decision?
3. If your budget got cut 20%, would this survive the cut? Why?
4. Have you ever seriously considered switching? What triggered that?
5. What would have to change for you to cancel?

Team and support:

1. How's the support experience? Give me a specific example.
2. Have you dealt with a bug or outage? How did they handle it?
3. What's your relationship with the founding team like?
4. Have they delivered on the roadmap they promised you?

Competitive and strategic:

1. Do you use any competing products? How does this compare?
2. What do your peers at other companies think of this product?
3. Would you recommend it to a colleague? Have you?
4. What would make this a must-have vs. nice-to-have?

The calibration questions:

1. On a scale of 1-10, how would you rate them? (Then ask: what would make it a 10?)
2. Is there anything I should know that I haven't asked about?

Question 19 is a calibration tool. If someone says "8" and then lists a few minor issues, that's a strong reference. If someone says "8" and then spends 10 minutes on problems, that "8" is really a "5."

Question 20 almost always surfaces something. Leave the silence. Let them think.

Reading Between the Lines

References lie by omission more than commission. They won't say "this product is terrible." They'll say things that, if you're listening, mean the same thing.

Soft red flags:

• "It works for what we need it for" — translation: we've settled for it
• "We're kind of locked in at this point" — translation: switching cost is the only thing keeping us
• "The team is really responsive to feedback" — translation: the product has had issues that needed fixing
• "We're still figuring out the best way to use it" — translation: the product hasn't found a clear use case even with existing customers
• Long pauses before answering "would you recommend it?"

Hard red flags:

• Specific unresolved complaints about bugs, reliability, or support response times
• "We were supposed to get [feature] 6 months ago and it still isn't there"
• Any mention of security incidents or data issues
• "We're actually evaluating alternatives right now"
• Churned customers who cite pricing as the reason — often a cover story for product disappointment

Spotting Coached References

When you hear three references all use the same language ("game-changer," "saves us X hours per week," "the team is incredible"), that's a coaching signal.

Founders sometimes brief their best customers before reference calls. That's not inherently dishonest — it's smart PR — but it means you're getting a rehearsed version of the truth.

Break through it with specificity. Ask for a specific story: "Tell me about the last time something went wrong and how they handled it." A coached reference will give you a generic answer. A real one will tell you a real story.

Also notice: do they volunteer any negatives unprompted? A real, satisfied customer will say something like "the UI could be cleaner, but the core functionality is rock solid." Unbroken praise with zero texture is a yellow flag.

How Many Calls to Do

For a seed deal: 5-7 calls minimum. At least one churned customer or lost prospect.

For a Series A or B: 10-15 calls. You should be talking to customers across different segments — enterprise vs. SMB, early adopters vs. recent acquisitions, power users vs. occasional users.

If the company has fewer than 10 customers, talk to all of them. Yes, all of them. That's possible and that's what you should do.

Pattern recognition is the point. One happy customer means nothing. Five happy customers with consistent themes means something. Three customers who use different words to describe the same frustration? That's signal.

When References Should Kill a Deal

References should kill a deal when you see any of the following:

The "can't get references" problem: If the founder can't produce 5 customer contacts after you ask, that tells you either the customer base is too small to fund, or the relationships are too weak for anyone to vouch for them. Either is a problem.

Consistent churn language: When 3 of your 7 references mention they've "thought about switching" or are "evaluating other options," the retention story in the pitch deck is fiction.

The NPS disaster: If you can't get a single unprompted customer who says "I'd recommend this to everyone," you don't have product-market fit — you have product-market tolerance.

Security or compliance issues: One reference mentioning a data issue is a yellow flag. Two is a pattern. Any mention of a breach or unreported incident in a regulated industry is a deal-stopper.

Mismatch with the pitch: The founder says "we're mission-critical infrastructure for enterprise finance teams." The references say "we use it for a few reports each quarter." That gap in how the product is actually used vs. how it's being sold will compound post-investment.

What to Do After the Calls

Document every call in a structured format: what they said, what you inferred, red/yellow/green flags. Share with your partners before the investment committee meeting — don't summarize, let them read the raw notes.

If references surface material issues, go back to the founder with specific questions. Don't say "one of your customers raised concerns about reliability." Say: "I heard from multiple customers that the uptime in Q3 was a problem. Can you walk me through what happened?"

How a founder responds to being confronted with customer feedback tells you as much as the feedback itself. Defensiveness and excuse-making are red flags. Ownership and a clear explanation of what changed are green flags.

Customer references are a competitive advantage for early-stage investors. Most people don't run them well. The ones who do catch the deals that blow up before they blow up.