the founder is two weeks into the partner meeting cycle for a series a. four firms are circling. on a friday afternoon, the partner who's been most leaned-in goes quiet for six days. monday after, the email arrives. "we've decided to pass. the team here didn't get to a strong enough conviction. we'd love to stay close."
the founder reads that email twelve times trying to find a number, a reason, a thing they could have done differently. there isn't one in the email. but there was one, and it happened on a thursday morning, on a thirty-minute call with someone who used to work at the company two years ago.
investors back-channel every term sheet. they call ex-customers, ex-employees, prior investors, and sometimes ex-cofounders. founders rarely hear the bad readouts. but the language those calls produce is remarkably consistent, and learning to recognize it changes who you list as a reference and how you brief them.
what a bad reference actually sounds like
the bad reference does not say "they're terrible." nobody says that. the bad reference says one of about six things that sound supportive but read as a pass to a partner who has heard the pattern before.
read that exchange in the voice of someone who likes the founder but wouldn't bet on them. "i'd take their call" is the social-circle floor. it tells the partner the relationship is intact. it tells the partner nothing about whether the founder ships, decides under pressure, or runs an honest forecast. the partner hangs up and writes one line in the deal memo: "reference soft." that line is enough to lose the round.
the other tells in the same family: "they're a really good person." "they're working really hard on it." "i think there's something there." every one of these is true. none of them say what the partner needs to hear, which is a specific story attached to a specific decision the founder made under pressure that turned out to be the right call.
what a good reference actually sounds like
the good reference has a story. that's the whole difference.
a partner doing diligence on a $12m series a in 2026 is not trying to learn the founder's personality. the personality has already been visible across four partner meetings. the partner is trying to learn whether, when the company hit a hard moment that the founder hasn't told them about, the founder did the thing the partner is now betting on them to do.
the language of a good reference: "when their biggest customer churned in q3 of 2024 — about 30% of arr at the time — they cut burn by $80k a month in two weeks and won the customer back with a new sku in march. i remember the week they decided. i'd back them." that's a thirty-second story. it has a named event, a dollar number, a decision, an outcome. the partner writes it down word for word.
why most founders pick the wrong references
most founders give investors the names of people who like them. that's the wrong filter. the right filter is people who have a story about a specific decision the founder made, ideally under stress, that turned out to matter. the second filter is whether the person enjoys telling that story.
a founder running this consciously goes back through the last three years of company history, identifies the five hardest weeks, and names the person on the outside who saw the decision get made. they're often not an obvious choice — not a board member, not a happy customer, not an ex-vp. more often a vendor who got paid late and worked something out, a customer who almost left and didn't, an early employee who left on good terms after the company changed direction.
the reference call isn't a popularity test — it's the partner asking whether you've already done the thing they're paying you to do next.
the forward reference move
the founders who run this best don't just pick good references. they brief them. not a script — the partner can hear a script in twelve seconds — but a frame. "the partner is going to ask about the q3 churn event. tell them what we did, what worked, what didn't."
then the parallel move: forward references. before the partner gets to diligence, the founder introduces them to people who have the right story to tell — "you should talk to maria, she watched us through the worst month we've had." the partner walks into diligence with the bad weeks already in the bloodstream, told by people who lived them. that partner is harder to spook with whatever comes up on the random ex-employee call.
what to listen for in your own outbound
every founder running a process should be doing their own diligence on the partners. ask the partners' references — other founders the firm has backed, ideally ones whose companies didn't break out — "what's [partner name] like when things are going badly?" the answer either has a story or doesn't. a partner whose founders can't tell a hard-week story will be quiet at your board meeting in q3 of 2027.
how zift handles this
zift tracks the events in your company history that produce reference-worthy stories — major churn, major saves, hiring decisions, pricing changes, runway moments — and surfaces them as a timeline on the monday briefing. when you're heading into a process, the briefing flags the five strongest moments and who, on the outside, saw each one happen.
if you're a finance lead at a series a team running a reference-and-diligence playbook across multiple stakeholders, zift handles that too.
the email that passes on you is the second event. the call that produced it happened a week earlier, and the names on that call were ones you put on the list yourself.