These lessons are listed in chronological order from oldest to newest.

2018

Just do it

I wanted to start a project that I could eventually make a business. There were a lot of things I did wrong at the beginning. First of all, Sebastian Ruder pointed out that I need to build something instead of thinking so much. Previously I had ugly code and messy documents because I never stopped to clean anything. I talked Sebastian how I had this idea and this time I wanted to spend some time thinking and making sure I can structure the project correctly. He instead advised me to do something and just prove the idea before focusing too much on that. I think the mistake I made was that I focused too much scaling when I haven’t even validated the idea.

2019

Don’t lock into solution but think what’s the root problem

I got idea for TN from own problem. I liked reading Google News (or actually the news in Chrome) but didn’t liked the fact that I often pressed some news because the title was interesting but then I was either interested of understanding what it was about in high level or some small thing. Never interested of reading the whole story. I thought that I need to build Google News like app and then add summarization. The biggest mistake I did here was that I locked into the most obvious solution. To make the app work I should create as good news recommendation as Google has which was of course impossible but I never thought about it. I don’t know how this could have been solved because there is no plugins in apps but I should have build something top of Google News not create also the news app when the only feature I was interested was summarization. In some sense I started with a solution instead of digging deeper to understand the core problem summarization not news recommendations etc.

Simple MVP

I started building TrimmedNews thinking that it takes a month to do MVP. Then it just grew and took more time (I never added that many new features but I wasn’t available to predict how long some steps could take) and in the end it took three months to create MVP. It was pretty good MVP but the problem was that I spent way too much time building it without any validation. I should have started with something much simpler like just email news letter where I have summarized some small niche of news like YC news. That way I could have learn something before building anything. And of course as pointed above it was a mistake to build such a big thing instead of something top of existing things. I also think I have could started with showcase website to see how people respond. They actually liked the idea at least on Product Hunt so maybe the execution was poor. It’s also important to remember rule “fold early”. The longer you spend doing the product the harder it is to see the truth if it doesn’t work and you need to do something else.

Complaints about look are useless

There was never a good retention alought the app got pretty good amount (~80) of upvotes on Product Hunt. Victor Bercaru contacted me on Reddit and that is where we started talking. I wasn’t good at visual part and he was so it was one of our biggest benefits working with him. Now actually, 6 months later, I realized that we spent a lot of time doing good on boarding and making sure it looked pretty (we always talked how complains about design wasn’t good because it wasn’t the main point but still we spent some time doing it) when we both knew that it shouldn’t matter. I guess Victor wanted this because he is growth person and it’s something that helps grow but not get product market fit. If the idea is good enough it will take of without great design and if it doesn't get traction that way it’s going to be slow grow the rest of the app’s life. The mistake was that we spent too much time designing. It might actually be easier if the app looks great because if people say that they don’t use it because the look it means that the feature(s) isn’t good enough. It’s kind of like good threshold features should beat to achieve success in level of Facebook.

If there is an option to do something later, it might be better to wait if the cost is small

First I was a little bit worried about taking co-founder in because I was afraid that he didn't have enough knowledge. I was so close to say no to him because of these fears but then I got advice from one YC company founder who said that it’s always possible to fire them later. That’s pretty good advice and I was happy to follow it because my co-founder ended up helping me a lot and it was really useful to have him. I couldn't hope any better. The problem although in this is that you might waste someone else's time and that's something need to be thought also.

Iterate quickly

When the app didn’t took off after the modification in design we started randomly test different feature ideas. There wasn’t any rules how we created those we just looked some reports and tried to understand how people acted and then created something. Private groups was one example. It was pretty good idea because we had the app and it was pretty quick to build the features but it still took something like two weeks per feature. At this point we had already made the mistake to build the app but we should have put it to bookshelf and just create simple websites in days to see how people react. So the iteration process was overly complicated and took too much time.

Talk with customers (and read Mom Test)

We (or mostly I) always talked with random people and met some great news readers. The biggest problem was that I hadn’t read “Mom Test” which is why I showed the app first and then asked opinion about it. We never did customer interviews that might have shown us the problems but we just asked some random questions without any particular goal.

Promoting

One thing I did well was that I promoted the product on forums, Slack channels, and everywhere people just wanted to listen. That way I was available to get someone write a story about the app to makeuseof without even asking.

When quitting is right