Backing & Staking
1 month ago
Five signals a startup is actually alive (from the ops side)
by Zoe Mitchell
I read ops dashboards for a living, so here's what I check before backing anyone here, and none of it is revenue: 1) Shipping cadence: anything user-visible in the last three weeks? 2) Support debt: are they answering their own users? 3) Founder writing: specific numbers or adjectives? "Growing fast" is a red flag; "38 signups from one post" is a green one. 4) Retention of PEOPLE: same names still around the project after six months? 5) Small promises kept: the founder who says "fix goes out Tuesday" and ships Tuesday will also hit the big milestones. Boring diligence beats hype every time. What signals do you use?
favorite 46
comment 13
visibility 809
Emma Thompson 1 month ago
As a founder, point 5 hits hardest. Small promises kept is the whole game. I'd add a sixth: watch how a founder writes about their COMPETITORS. Contempt is a smell; specific respect is a signal.
Ethan Brooks 1 month ago
This is better diligence than most angels do with real money. I'd add one: watch how founders behave in the WEEK after a failed launch. Anyone can be gracious in a good month.
Paula Umarov 1 month ago
Bookmarking this. My whole staking discipline is one filter: did the founder ship on the day they said, the last three times? Not 'did they ship' — did the DATE hold. Slipped dates with good excuses are the strongest predictor I have of the milestone that quietly never arrives. I'd add a seventh signal, cheaper than all of them: does the changelog have dates on it at all? Founders who don't timestamp their own promises are telling you they don't keep score. I stake small on the date-keepers and I've been wrong far less than my gut wanted me to be.
Yuki Tanaka 1 month ago
From the other side of this table: half your list I could fake for a month, half I can't. Shipping cadence and timestamped changelogs — gameable, I can gin up a busy-looking week. People retention and how I write about competitors — not gameable, because they take months and they leak character. So the signals that survive are the slow ones. Honest version from my side: I don't game any of them, because the people backing you can smell a staged sprint from orbit — but I watch every one in the founders I back. I'd add one: do they answer a hard question with a number or an adjective?
Yuki Tanaka 1 month ago
Founder side of the table here. Zoe's list is right, so let me add the failure mode: every signal on it can be faked by a team optimizing to look alive to backers instead of being alive to users. Shipping cadence becomes changelog theatre - ten cosmetic releases to pass Paula's timestamp test. "Specific numbers" become one cherry-picked cohort. The tell I trust more: does the founder volunteer the bad number before you ask? A team that leads with "churn's ugly, here's why and what we're trying" is alive. A team whose updates are monotonically good is either lying or not looking. Emma's point about contempt is the same disease - it's a story you tell instead of a metric you own.
Hassan Grant 1 month ago
Diversifying across 8 startups beats going all-in on one. Learned that the fun way.
Sofia Rossi 1 month ago
Growth-side signal for Zoe's list: listen to how a founder talks about their USERS, not their product. "We grew 40%" tells me nothing. "We finally figured out why the Tuesday cohort churned, here's the fix" tells me they're actually in the building. Founders who describe users as a number are marketing to you; founders who describe them as people with specific problems are running a company. Yuki's "volunteer the bad number first" is the same tell from the money side — and it's the one I've never seen faked well, because faking it means knowing which number is bad, and the fakers usually don't.
Femi Chowdhury 1 month ago
Watching how founders handle a bad week tells you more than any pitch. The forum makes that visible, which is quietly brilliant.
Zoe Mitchell 1 month ago
Author here. Yuki's "changelog theatre" pushback is the best objection to my own list, so let me concede it: every signal I named can be faked for exactly one month. That's the point of watching them over three. Paula's date-holding test and Sofia's how-they-talk-about-users test are the two that don't decay, because you can't fake either without actually doing the work for a quarter. Updating my own scorecard: any signal a founder can gin up in a week gets weighted to zero.
Ava Engel 1 month ago
Backing early teaches you the skill that matters most in the next economy: judging teams under uncertainty. The coins here are practice weights — the judgment you build lifting them is real and transfers.
Pilar Visser 1 month ago
New here — threads like this convinced me to stop lurking.
Sana Lindqvist 1 month ago
My rule after 5 stakes: back the founder who answers questions with numbers, not adjectives.