I run a pre-launch waitlist for AfterPack, a JavaScript obfuscator I'm building. For months, the screen right after the email signup asked one open question: what matters most to you? Fewer than 1 in 10 signups ever answered it.
Then I changed that one screen, and the answer rate went to 100%. Every signup, 100+ in a row. This post is the breakdown of what changed and why it works - including the caveats.
The setup: an open question nobody answers
The original post-signup screen was a single free-form box. Polite, low-pressure, completely standard:
The original screen. 2 of 30 early signups typed anything at all - usually a single word, like "React".
Think about how strange that is. These aren't cold visitors - they just handed over their email for a product that doesn't exist yet. They are the highest-intent audience the project will ever have, and over 90% of them skipped the only question on the page.
The fix: make answering a reflex
I rebuilt the screen around one-tap chips - what are you building, what's your main concern, what describes you - with the free-form box demoted to the bottom:
The redesign. 100% of signups since have tapped at least one chip.
The mechanism is the oldest one in interaction design. An open question forces a busy person into System 2 thinking - stop, compose, type - and the rational move is to skip it. A chip is a recognition-over-recall task: read, tap, done. The fewer and more concrete the options, the faster the choice - Hick's law working for you instead of against you. Keeping the open box underneath the chips is textbook progressive disclosure: the cheap action first, the expensive one optional.
There's also a timing component. Someone who just typed their email has already said "yes" - the same instinct that makes you reach for your wallet after a meal. You agreed to the deal when you sat down. The post-signup screen catches users at the exact moment their commitment is highest and the cost of one more tap is trivial.
Interestingly, the free-form box itself now converts better too - 10 - 20%, up from under 10%. Tapping a chip seems to warm people up for typing.
The honest caveats
I didn't run a clean A/B test; the cohorts are sequential, so treat the exact percentages as directional. The sample is 100+ signups, not thousands. And chips constrain the answer space - you only learn about the dimensions you thought to ask. That's a real trade: closed data you actually collect versus open data you never get. (The free-form field underneath is the escape hatch for unknown-unknowns.)
The gap, though - sub-10% versus literally everyone - is not the kind of difference that needs a significance test to act on.
Why a waitlist at all
The less obvious half of this story: the waitlist isn't a vanity counter, it's how the distribution half of a product launch runs in parallel with the build. Building the right thing is one half; knowing where the traffic will come from is the other, and it's the half that quietly kills products.
While AfterPack is still being built, the landing page has been answering that question for months: four blog posts, a few link drops, about $500 of Google Ads (a generous overpayment for some eyeballs and learnings), and signups now arrive organically - a few per day. I ran the same play years ago with DataUnlocker, my web analytics SaaS - capturing intent early compounds while you sleep. It's the laziest version of Paul Graham's do things that don't scale.
With chips on the post-signup screen, that pile of anonymous emails turns into a live read on who is waiting: which frameworks, which concerns, which roles. That's customer development running unattended.
The takeaway
When a user gives you their email, that's already a "yes" - spend the moment after it on real questions, and make answering so dead simple that people would feel stupid not to. If your waitlist collects only emails, you're sitting on your one free shot and not taking it.
Originally published on nikitaeverywhere.com, with some fun interactive diagrams, check it out.
Top comments (0)