Lessons Learned Launching A Side Project in 48 Hours
Launching a side project can drag on forever, at least if you have a full time job. My friend Andreas and I know this all too well, as we’ve discussing ideas forever, but never really launched anything together.
So in order to break out of the procrastination, we decided to set off an entire weekend in our calendars in order to build and launch a product idea Andreas had been thinking about a year.
The result was BugRex.com — an on-demand chat for people who need technical help. It’s operated almost 24 hours a day by professional developers around the world, which’ll help you for up to 20 minutes for 10 USD.
So far, we’ve had a few hundred chats and five completed sales. In this post I want to share the lessons we learned.
The Idea
Our thesis is the following: when building products or learning new technical stuff, you’re often stuck with bugs and concepts you don’t understand. While there are many good alternatives out there to help you, like Stack Overflow, subreddits and public chat rooms, those aren’t always sufficient.
We had often talked about how sweet it would be to get one-to-one help from someone who knew their shit; someone who could guarantee that they’d fix the respective issue. You know, the kind of help you’d get at school if you raised your hand.
So what about a service where struggling developers can get instant one-to-one help from other developers?
We knew that there were two solutions out there already, HackHands and Codementor.io. However, they’re both quite pricey, starting on $1 per minute, which rules out a lot of students and young people.
Secondly, we didn’t want any signup, which both of these requires. So, naïve as we are, we set out to launch a competitor in a weekend.
Minimum Trashable Brand
We met up on Wednesday afternoon the same week, as we wanted to do a little pre-weekend test.
In the days before, we hacked together a site called ReactHelp, using Unbounce as the landing page builder and Olark as the chat solution.
It was a free chat line for React newbies in the need for help, with me as the operator — basically a simpler version of our product, targeting very narrow niche.
Lesson learned: The benefit of choosing a trashable brand in the beginning is that you can treat it with less caution. We knew we were going to scrap it eventually, so we weren’t afraid to launch it even though it looked quite crappy.
We posted it to a few relevant subreddits, and pretty quickly got a few chats going. A lot of them where spammy messages, but a few were developers with actual problems.
This proved that people at least were willing to unload their problems onto a stranger in a chat they had just discovered. Not a huge insight, but still.
The Product
I continued to answer a few requests on ReactHelp the following days and by the time we met up again on Friday afternoon, ReactHelp.com had been picked up by Javascript Weekly, and suddenly got an all time high in terms of visitors. This gave us a nice motivational boost to kickstart the weekend.
However, we didn’t want to continue with ReactHelp.com, as we saw more potential in a slightly broader approach.
So we went back to the drawing table. This time we coded the landing page ourselves, using GitHub Pages as hosting, and added customized the Olark chat box — a super nice feature which makes it possible to style it exactly how you want it.
We also decided to start charging for help, $5 per bug fix, as we wouldn’t really learn anything from giving out more free advice. It’s when you start charging for your product you learn if it’s actually worth anything to your users.
In order to seem a bit more professional, we also a button to the top right of the site saying Become an Expert, which linked to a TypeForm application form. This was actually the very last feature we added, which later turned out to be one of our smartest moves.
The Launch
Saturday at around dinner time we were finally ready to launch. To prepare for viral traffic and hockey stick growth, we’d hired a developer, Kirill, through Upwork to help us with the incoming chats.
We submitted the site to a bunch of subreddits and relevant Facebook groups, and just as with ReactHelp, we started seeing some traffic, along with chat requests.
At some point, we started getting traffic from Hacker News as well. Our plan to wait with a Show HN post until Sunday, but that obviously failed, as someone had posted the site there already. Luckily though, it actually reached the front page.
This resulted in a good amount of traffic, totaling at around 1.4K visitors throughout the evening.
It also got a lot of traffic from the /programming subreddit, where it performed far better than any other subreddit. This is likely to do with how we presented it there:
Lesson learned: It’s better to be personal an ask for feedback than pitching your product on Reddit. Just look at the different engagement in the /programming and /coding subreddits in the image above: 64 votes and 81 comment versus nothing at all.
At this time we were three people operating the chat: me, Andreas and Kirill from Upwork. We were all quite busy answering people.
However, the signal to noise ratio was horrible.
Almost all of the chats went in the lines of something like:
Which was kind of depressing.
But there were some highlights as well: a tiny minority of the requests were people who had coding related problems, and we actually made 5 USD fixing a CSS bug for a guy in Turkey, which was kind of fun, as he was thrilled we could fix it for him. We also managed to help a few others, but none which were willing to pay for the help.
What struck me though, was the amount of time it took to help someone. It took rarely less than 15 minutes, and sometimes we didn’t even manage to solve their problem at all.
Thoughout the evening the site got quite a lot of good feedback on Reddit, but also it’s fair amount of criticism of course, which helped us to do a lot of small improvements.
Lesson learned: The Reddit /programming community is a lot harsher than people on Hacker News. When sharing your stuff with these kinds of communities, expect to get criticism, and don’t take it personally.
When we left the office late at night, our feeling towards BugRex was kind of mixed, to be honest.
To a degree, we felt that we had misunderstood how the internet works: of course we’d get a lot of spam when we open up a chat for everybody. Duh! So we wondered if this open chat could work at all. Though we hoped and believed that the signal/noise ratio would improve when the HN/Reddit rush calmed down.
We also questioned wether there are actually enough bugs out there which are in the sweet spot. By that, I mean that there’s an overlap between experts’ willingness to solve the problems and the customers’ willingness to pay to get them solved.
Fixing Supply
Fast forward to Sunday morning. When we met up at the office, we logged into TypeForm (remember that Become and Expert button) and saw that we had gotten over 80 applications. Wow.
Which meant developers were at least interested in selling their advice. So even though the demand side of the marketplace was slow and polluted by spam, the supply seemed easier to fill.
This was great news, as we had expected that the only way we could get experts operating our chats was through hiring people on Upwork, which would be far to expensive for us, at least with the current revenue numbers.
Lesson learned: If you need help, partners or collaborators on a project, asking passively for it (e.g. through a form) can work just as well —or much better — than actively looking for people (e.g. finding people on Upwork).
So I spend the Sunday going through applications and sending out interview invites, while Andreas worked on improvements on the site.
We also decided to the raise the price and change the focus slightly. Rather than focusing solely on bug fixing, we geared it to towards talking with a developer in general.
So the value proposition was now: 20 minutes of help from a professional developer for 10 USD.
The experts we talked to were fine with this solution. Actually, very few of them mentioned money as their motivation to be a BugRex expert, but rather that they enjoyed helping people and saw it as a way of learning more stuff themselves.
Lesson learned: There is hope in this world; people aren’t necessarily the money optimizing consumers you learned about in micro economics at university.
By the evening we had hired operators in three different continents: Europe, Africa and North America. This would make the service available almost 24 hours a day.
What now?
On Monday we finally had our first proper sale, without me or Andreas being involved. An american developer earned 10 USD helping a guy in Hong Kong with a JavaScript issue.
All in all, the customers have been thrilled by the service they’ve gotten, which fills us with excitement, so we’re going to continue working on this project.
However, there are a lot of challenges ahead, and hypothesis’ that need to be proven. Here are a few questions we ask ourselves:
- Do we have the right business model? Perhaps we’re doing a mistake by charging for the chats?
- Will our customers return to BugRex next time they have problems? In other words, is the service good enough to create a habit?
- Which marketing channels will work best? Content, Social media, SEO, paid marketing etc.
- How will we continue to attract awesome experts?
We’d love to get your input, so if you have any comments or suggestions, feel free to leave a comment.
And don’t forget the heart if you liked the article :)
