Mastermind · May 10, 2026
Public Likes, Private Bookmarks
A startup with twelve hearts and seventy bookmarks.
That was the line that stopped me on Thursday night, scrolling through Demo Day Connect's leaderboard with the room half empty and the cleanup crew folding chairs. The likes were lower than I expected. The bookmarks were three times higher than the likes. And the two numbers were not picking the same companies.
We built Demo Day Connect for Stanford Founders Demo Day this past week. It's a web app: investors and founders both sign in, every investor gets $100,000 of play money, every founder gets a pitch card with a like button, a bookmark, an invest button, a private chat, and a deck-request flow. We shipped it in three days. By the end of the day on Thursday, more than a thousand people had used it, $1.3M of play money had been deployed across 57 active commitments, 355 likes had been hit, and 62 founders had claimed 490 deck requests. It is the most viral piece of software I have ever shipped. It ran for six hours.
The number I keep coming back to isn't the thousand users. It's the bookmark column.
Likes are public. The number on the card updates live for everyone. The investor knows other investors see it, the founder knows their friends are watching, the founder's roommate is screenshotting the leaderboard. A like on Demo Day Connect is not a private signal of interest. It's a social act, a small public bet on what the room thinks is going to win.
Bookmarks are private. The only person who sees that you bookmarked a startup is you. The founder doesn't get notified. The leaderboard doesn't move. There is no social cost or social return.
By Session 3, those two columns had decoupled. There was a category of startups with low likes and high bookmarks: companies the room was not loudly cheering for but that individual investors were quietly flagging to revisit. Halo and Tales was the canonical example. Twelve likes. Seventy bookmarks. If you only looked at likes, you'd have written it off as a session-2 disappointment. If you looked at bookmarks, you'd have noticed that a meaningful slice of investors thought this was the most underpriced opportunity on the floor.
The divergence between public and private signals isn't new. We've known about it since A/B tests existed. What's new is how clean the signal is at a live event. In async product analytics, your "likes" and your "bookmarks" are not actually comparable. The like came from someone scrolling a feed; the bookmark came from someone deep in a research session. Different cohorts, different intents, different times of day, no shared denominator. The data is dirty by design.
At a demo day, there's no async. Everyone is in the same room, on the same schedule, with the same context, evaluating the same set of companies in the same hour. The denominator is identical. So when a private signal contradicts a public one, that contradiction is almost pure, stripped of every confound you'd see in a normal product surface.
The other thing live events do is collapse the time window. If you spend a quarter watching dashboards for a SaaS product, the signal you get is laggy, smeared across a thousand sessions, contaminated by retention dynamics. If you spend an afternoon watching a leaderboard for a demo day, the signal is dense, contemporaneous, behavioral. By Session 3 you have a saturating sample of what 400 actual investors think about 50 actual companies. No SaaS dashboard gives you that resolution.
The thesis I'd defend if pressed is this. Synchronized live events are a software leverage point that gets undervalued because most people building tools optimize for the asynchronous web. The web rewards persistent attention spread thinly across millions. A live event flips it: you get six hours of dense, synchronized attention from a few hundred high-context people. If you build software for those six hours, the engagement per square inch of code you ship is absurd compared to anything you'd build for the open web. Demo Day Connect is the loudest example I've personally hit, but it's not the only one. Every hackathon I've run with a custom platform has had this property. People show up because they're in the room. They sign in because they're committed to the afternoon. They use the thing because using the thing is the event.
The leverage compounds when you build features that are only legible at scale, in real time, in one room. Public likes versus private bookmarks is one. Another is the cross-table on which startups got the most decks requested versus the most chats started; those two cohorts overlap less than you'd think. A third is investment concentration: $1.3M of play money distributed across 50 startups is not uniform. The fat tail tells you who has unambiguous demand. None of those signals exist if the event isn't synchronized.
The next move is making the asymmetry actionable in the room, not after. Right now the founder of Halo and Tales doesn't know they're getting bookmarked. The investors who bookmarked them don't know they're the only ones who saw it that way. There's an obvious feature where you surface the asymmetry quietly, on the founder's side: "you have unusually high private interest." Then watch what they do with it. Do they follow up more aggressively? Do they re-pitch on the floor differently? Does the asymmetry close?
That's the part I want to play with next. The like is for the room. The bookmark is for the investor's own brain. The space between the two is where actual interest lives. We just had a thousand-person, six-hour window where I could see that space clean. The next product to build is the one that closes it.
Check in next week. If Halo and Tales raises faster than the heart count predicted, the leaderboard was lying.