We got the scroll speed wrong

The assumption we made

When we built Timed Scroll, the free feature that automatically scrolls the script in time with your speech, we made a choice about the default speed. We thought about typical speaking rates. We looked at YouTube tutorials. We looked at Ted Talk transcripts. Then we picked a number that felt 'middle ground' and shipped it.

The problem was that we tested it ourselves. We stood in the office, read our own scripts at our own pace, and thought, 'Yeah, that feels right.' We didn't test it with preachers doing a Sunday sermon. We didn't test it with YouTubers doing sponsored reads. We didn't test it with someone nervous and rushing through takes. We just picked a number.

What the real data told us

Once Promptr was live, we had something we didn't have before: thousands of creators actually using the app. Not in a lab. In their bedrooms, their studios, their pulpits. And they started telling us what didn't work.

The messages came in different words, but the pattern was clear. Creators were either babying the scroll (speaking slowly, waiting for the text to catch up) or getting frustrated and switching to Smart Scroll (which lets you swipe manually). The default we'd chosen worked for maybe twenty percent of the people using it. For the rest, it was either too fast or it created cognitive friction.

We also noticed something in the free plan limits. People weren't upgrading to Creator just for video recording. They were using Timed Scroll more than we expected. That meant we'd shipped an acquisition feature that was pushing people away rather than inviting them to stay.

Why it's harder than it sounds

You'd think fixing this would be simple: just slow down the default. But it wasn't that straightforward. Speaking pace varies massively depending on content type. A preacher delivering emotionally charged material moves differently than someone reading ad copy. A nervous YouTuber records fifteen takes; a podcaster records one. Someone with a speech impediment moves at a different cadence than someone without.

We could have added a dropdown menu with preset speeds. But that's friction. New users want to open the app, paste their script, and hit record. They don't want a settings page before they've even started.

So we did something different. We lowered the default speed. Not arbitrarily. We looked at the actual scroll behaviour in our logs and chose a number that wouldn't punish the majority. Then we made Smart Scroll even easier to access, because sometimes automatic scrolling isn't the right tool anyway. You can toggle between Timed Scroll and Smart Scroll with one tap.

The Pro creators told us something else

Once we started paying attention, our Pro subscribers (who get voice scroll, which watches your voice cadence in real time) told us something useful. They weren't using it as a replacement for Timed Scroll. They were using it for second takes, or for moments when their pacing changed mid-record. Voice scroll is Pro because it's resource intensive and requires processing power, but it wasn't just a 'premium feature' for the sake of it. It solved an actual problem.

That gave us confidence that we were building the right tools, just with the wrong defaults. The issue wasn't the feature set. The issue was we'd assumed too much about how people would use them.

What this taught us about shipping

Defaults matter more than options. A menu full of settings doesn't make a product better; it makes it more confusing. We spend a lot of time now thinking about what ninety percent of users will do the first time they open Promptr. The remaining ten percent can explore. But the first moment has to work without explanation.

The other thing: you can't test your own product at your own pace. We're still wrong about things. Probably. But now we're listening for the moments when actual creators tell us so. That message from Marcus wasn't a complaint. It was feedback we should have asked for before launch.