I put a founder-answered chat button on every logged-in page. Small ship, might be my highest-leverage move this month.
Quick build-in-public update from BuildBase (buildbase.app - a Node/React SDK that bundles auth, billing, workspaces, notifications, etc. into one npm install).
Where we are: ~100+ signups, still working on landing our first external paying customers. Five of our own products are running on it in production, so the dogfood proof is real, but the paid-conversion story is a work in progress.
The thing I've been quietly obsessing over is activation. Not traffic. Not SEO. Not new features. Just: people sign up, land in the console, and some non-trivial number of them bounce before they hit the moment where BuildBase actually saves them time. That drop-off is the whole ballgame right now.
So today I shipped what looks like a boring update:
-
Fixed a handful of dashboard bugs on the console that were quietly making the first-run experience feel scrappier than it is.
-
Added a Crisp chat button on every logged-in page. Not just on marketing pages. Every screen inside console.buildbase.app.
The second one is the actual bet. Here's the reasoning:
At pre-revenue stage, my worst enemy isn't a competitor. It's silent bounce. Someone signs up, gets confused on step 3 of setup, and closes the tab. I never know what step 3 was. I never get to fix it. They never come back.
Chat on every logged-in page collapses that. When someone gets stuck, they can yell at me in one click. I'm the one replying. Not a bot. Not tier-1 outsourced support. Just me, usually within an hour or two.
Two things I get from this:
- The user gets unblocked and might actually finish setup instead of bouncing.
- I get a live stream of exactly where people trip. That's product research I'd otherwise pay a lot for.
The honest tradeoff: this doesn't scale. If we hit 10,000 users I can't personally answer every chat. But at 100 signups, the "doesn't scale" version is the correct version. Paul Graham was right about that one.
The other honest thing: I could have shipped this on day one. I didn't, because "founders should focus on writing code" is a comfortable story and answering chat feels like a distraction from the roadmap. It isn't. It IS the roadmap, at this stage.
Curious how others handled support at the pre-revenue-to-first-customers stage. Did you gate chat behind a paid tier, or leave it wide open? Anyone regret going too high-touch too early? Would love war stories.
The chat button catches your loud stuck users, but the silent bounce you're actually worried about is the person who closes the tab without clicking anything. The ones who yell at you were probably going to finish anyway, they're already your most engaged. So chat gives you a great qualitative stream, but it self-selects for the vocal minority, and the activation loss usually lives in the quiet majority who never raise a hand. What paired well for me was a passive stuck-detector alongside it: if someone sits on setup step 3 past some dwell time with no progress event, proactively surface a nudge (or the chat itself). That turns a silent bounce into a signal you didn't have to wait for. I'd also reframe the doesn't-scale worry: at this stage chat isn't support, it's research, and the win condition is that chats about step 3 stop recurring because you fixed the wall into the product. If the same confusion keeps coming, you're bandaging, not fixing. Do you tag incoming chats by which setup step triggered them? That's the line between a one-off and a systematic wall worth shipping a fix for.