3 Weeks After Launch: What I've Fixed and What's Next

Launching is the easy part

The actual launch went fine. Press the button, watch the downloads trickle in, refresh the analytics page more than any healthy person should. That part has a shape to it — a beginning and an end. What comes after is less defined. Three weeks of watching real people use the thing you built, and the gap between what you imagined and what actually happens starts to sharpen.

Here's where Ribbit stands right now, honestly.

The onboarding problem

The original onboarding handed new users directly to Lily, the AI assistant. The idea was simple: tell Lily where you want to go and she'd build your first plan. In theory, that's the product's strongest pitch — conversational trip planning that actually works.

In practice, it was a terrible first impression.

AI is unpredictable by nature. That's a feature when you already trust it, when you've seen it rearrange your itinerary or surface an experience you wouldn't have found. But when you've just downloaded an app and you're trying to figure out what it even does, unpredictability isn't charming — it's confusing. Some users got a great first plan in two messages. Others got stuck in a loop of clarifying questions that went nowhere. The variance was too wide for something that's supposed to feel reliable on first use.

So I replaced it with a structured onboarding wizard. It asks a few clear questions — where are you going, when, who's coming, what kind of trip — and builds your first plan from those answers. No ambiguity, no conversational dead ends. Once the plan exists and the user has something concrete to look at, then Lily shows up with destination-specific suggestions and context about what she can do.

It's less flashy. It's dramatically more effective. The first experience is now predictable, and Lily enters the picture at the moment where her strengths actually land.

Zero paid users

Three weeks in and nobody has converted to a paid subscription. I'm going to say that plainly because dancing around it would be worse.

The context matters, though. The user base is tiny — we're talking about a handful of organic downloads with zero marketing spend behind them. A significant chunk of those early users hit the broken onboarding and probably never made it to the point where the app's real value shows up. And I deliberately paused all marketing while I fixed the foundation, which means the people who would convert haven't seen Ribbit yet.

Zero paid users on a small sample with known onboarding issues and no active marketing isn't a failure signal. It's just noise. The failure signal would be fixing onboarding, driving real traffic, and still seeing zero. That test hasn't happened yet.

Pausing marketing

This was the hardest call. Every instinct says push — get eyes on the app, run some ads, see what happens. And when you're solo, there's nobody to tell you whether that instinct is right or whether you're about to pour money into a leaky bucket.

The math wasn't complicated. If new users land on broken onboarding and bounce, I've paid to acquire someone who now has a negative first impression of the product. That's worse than not acquiring them at all. At least a stranger has no opinion — a churned user has a bad one.

So I stopped spending and focused on the rebuild. It cost me three weeks of potential growth. It saved me from compounding a problem that would have been harder to fix with a larger pool of disappointed users.

I still don't know if it was the right call. But it was the defensible one.

New App Store screenshots

The original screenshots were rushed and didn't represent what Ribbit actually does well. New ones are going live this week — cleaner, more focused on the core experience. Small thing, but the App Store listing is the front door and it should look like the product behind it.

What's next

The foundation feels solid now. Onboarding works. The core planning experience is strong. Lily is genuinely useful once users get past the first few minutes.

The next phase is straightforward: turn marketing back on and see what real acquisition numbers look like with a product that actually retains. I'm not expecting a hockey stick. I'm expecting data — real signal on what converts, what retains, and where people drop off.

Ribbit is free to download on iOS and Android. If you try it, I'd genuinely like to hear what works and what doesn't. And if you're building something solo and any of this sounds familiar — reach out..