Every Sunday, you'll get a new lesson about product, design & startups to your inbox. Researched, heavily user focused & without fluff.
|
Vibe coding has been great for starting new ideas and getting some initial feedback. But there's always a point where you actually want to go out and get real customers. That's where you realize it needs to look professional and trustworthy to actually convince paying customers. Because no one is going to send you thousands of euros for software that looks like you made it on the weekend. I've fixed 25 software tools this year. Here are the most common mistakes I saw across those 25. Mistake 1: There's no onboardingWhen you're vibe coding, you spend a lot of time inside the software you're building. You know how it should work. But especially with vibe-coded software, people land on a home screen or dashboard that's just cluttered with features and information. They have no idea where to start. Some of my design friends' vibe-coded software, 10 out of 10, starts with a long and intensive onboarding. Because most vibe-coded software runs on information about the user, and then tries to prepare the best solution. Like a copywriting tool that learns how you write. How do you get started? By uploading all your content, your messages, and so on. For the Curio e-learning platform, we fixed the onboarding, and the user activation rate rose from 18% to 38%. That means instead of every seventh user actually starting to use the software, every third one does. That's a big jump. Mistake 2: The dashboard is overwhelmingFixing the onboarding leads straight into the second issue. Most of this software has a home screen that's completely overwhelming. The best software home screens mix the most important KPIs with the onboarding itself, so there are always onboarding steps that still need to be completed. Some of them need you to connect your own GPT, for instance. Until you do that, the dashboard is basically useless. So why not just use that space for the onboarding instead? I always see the same 3 metrics on any new dashboard. Maybe yours has them too. These are just assumptions that they're important, most of the time. Just some big numbers that don't matter at all. To fix that, we talk to the customers or users of the software and ask which metrics they actually use and care about. Nine out of ten times, we have to change those metrics and we have to change the dashboard to make it useful. Mistake 3: You only tested the happy pathThe happy path is a design concept where you interact with the AI or the software and everything goes right, and you see the perfect outcome. As you can imagine, that's usually not the case. So you need to design for all the other cases too. This is what's missing a lot:
For instance, I was working on a tool that writes product descriptions. Every time you started writing a description for a product, you couldn't stop it, so you had to wait it out. In theory, you could start the copywriting process for a thousand products at once and never be able to stop them. That's the kind of thing you need to work on. Otherwise people spend credits they don't want to spend, and you end up paying a lot of AI credits for it. Mistake 4: No self-cancellationIn the EU, canceling a plan should be as easy as signing up for one. I've seen live, paid vibe-coded software where the only way to cancel was emailing the founder directly. That's illegal, and it's also just a bad user experience. We have to fix that. Mistake 5: Nobody set up the emailsGood software isn't just the software itself. It's also all the emails that come around it: welcome email, onboarding email, confirmation link, maybe a short campaign in the first few days that shows you how the software works. You could ask them to sign up for a guided demo in case they haven't yet. And emails that get sent as invoices. All of this needs to be thought about. Let me fix your software, I have experienceOut of those 25 startups, we fixed the ones that were missing this. If you vibe-coded your software and you want to go out there, look professional, and get ready for real customers, or profit more from the ones you have, book a call with me below.
I've seen it all. I can help you, and we can fix this in a two-week design sprint. PS.: This post was recorded using WisprFlow, and manually edited. This is how I talk. Let me know if you like this over highly-edited blog posts :) |
Every Sunday, you'll get a new lesson about product, design & startups to your inbox. Researched, heavily user focused & without fluff.