Self-taught, bootstrapped, 980 paying customers

In short: I started out live-streaming weddings and local events, saw a gap in how streams reached their audiences, and taught myself to code a product to fill it. Splitstream let anyone send one live stream to many platforms at once. With no funding and a team of five at its largest, it reached 21,000 registered users and 980 paying customers worldwide, including people at Everton FC, Kaspersky and the German Bundestag. This is the story of what worked, what didn’t, and what I’d do differently.

Where it started: streaming other people’s events

Splitstream didn’t begin as a software company. It began as a service: I live-streamed weddings and local events for clients.

Doing that job showed me the gap. Clients wanted to be live on Facebook and YouTube at the same time. But every extra platform meant sending another full-quality stream over the venue’s internet connection, and that connection often couldn’t cope.

It also showed me the limits of the business I was in. The market for streaming weddings and local events was small. The market for live streaming itself was huge, and growing. Every streamer, not just my clients, had the same problem: their audience was spread across platforms, and reaching all of them at once was expensive, complicated, or both.

So I stopped selling a service and decided to build the tool. In July 2018, I went full time on it.

v1: teaching myself to build it

I wasn’t a software engineer. I taught myself PHP and React and built the first version on the Wowza streaming engine. The idea was simple: send your stream to Splitstream once, and Splitstream sends it on to every platform you choose.

I registered the domain in November 2018. By January 2019 early users were testing a beta; one of them wrote a public review noting bugs but liking the product. That same month, my co-founder and I incorporated Clay Engine Pvt. Ltd. in Hyderabad. The first paying customers arrived in May 2019.

v1 proved people wanted the product. It was also rudimentary and full of bugs.

Decision: rebuild instead of patching

The bugs weren’t the real problem. The streaming engine was. Wowza wasn’t designed to handle many incoming streams at once while keeping performance steady, and handling many streams at once was the entire product. Patching the application wouldn’t fix that.

So I rebuilt Splitstream from scratch: Express and Angular for the application, with Flussonic at the streaming core. Flussonic was far more powerful for this workload and more cost-effective to run. For v2 I brought in a backend contractor, with two interns supporting. The archived versions of the site show the switch: a React build in 2019, Angular builds after.

v2 launched in May 2020.

The product and the business model

Splitstream did one job well. You set up your destinations in a dashboard, pointed your streaming software (OBS or Streamlabs) at Splitstream once, and went live everywhere at the same time: YouTube, Facebook, Twitch, Twitter, DLive, custom RTMP servers and more, over 40 platforms in all. Channel doubling let users stream to several accounts on the same platform.

The business model was freemium:

Plan	Price	For
Basic	Free	Up to 5 channels, trying the product
Standard	US$9.99/month	Up to 15 channels, no bitrate limit
Pro	US$29.99/month	Up to 30 channels, website streaming
Pro+	US$59.99/month	Up to 50 channels, high concurrency

Annual plans were discounted. The free tier brought people in; the paid tiers unlocked the channels, quality and destinations that serious streamers needed. About 4.7% of registered users became paying customers.

The clearest description of the value came from a customer, not from me. In September 2021, a German neighbourhood publication described streaming events from local pubs through Splitstream to Facebook and YouTube:

“…we saved 50 percent of our data volume, because the video only had to be sent once. […] The only cost we had was 10 euros for a monthly subscription to [Splitstream].”

— Kiez und Kneipe, 12 September 2021 (opens in new tab) (translated from German)

That was exactly the gap I’d found streaming weddings.

Growth: v2, and a world that suddenly went live

v2 launched in May 2020, just as the pandemic pushed events and creators online. Streamers who were suddenly getting more traffic wanted to reach audiences on every platform at once, and Splitstream was ready for them.

For the three months after launch, sign-ups doubled month over month, and sales grew 70% month over month. 2020 became our biggest year.

I ran marketing myself, alongside product and engineering: Google, Facebook and Instagram ads, plus regular organic posts. Splitstream also earned a place in the market on its own. Forums, guides and “Restream alternatives” roundups listed it alongside established competitors like Restream, Castr and Switchboard. A 2022 guide to running a digital congress recommended it as a low-cost way to stream to a conference platform and YouTube at once.

Who paid: 980 customers around the world, and no single customer the business depended on. Paying users included people at Everton FC, Kaspersky and the German Bundestag.

What went wrong

After two strong years, growth reversed. Four things drove it.

1. A tiny team can’t ship fast enough. Customers stayed for a few months on average. The main reason was the pace of new features: bootstrapped, and effectively the only full-time engineer, I couldn’t add what customers asked for quickly enough to keep them.

2. The pandemic tailwind faded. The surge that made 2020 our best year didn’t last.

3. A piracy incident took down our European server. In November 2022, a user relayed FIFA World Cup matches through Splitstream, even though our terms explicitly prohibited piracy. Our European server was flagged and taken down. Dealing with it and recovering cost us time, money and users at the worst possible moment. It taught me that for an infrastructure product, trust and safety can’t be an afterthought: you need abuse monitoring and automated detection of copyrighted content before you need them.

4. The co-founder partnership wasn’t working, and I waited too long to act. Splitstream needed a hard call that I delayed.

Winding down

By 2023, with growth reversed and my own time split with a full-time product role, Splitstream was wound down. It had 21,000 registered users in its database when it closed.

The lessons stayed with me. Almost everything I’ve done since, from writing specs that engineers can build to thinking about retention from day one, traces back to running Splitstream.

What I’d do differently

Make hard co-founder calls faster. The longer a misaligned partnership continues, the more it costs the company. I’d address it within weeks, not years.

Hire sales and marketing early. I built the product, ran the ads and handled sales. The product was the part I was best at, and the other two got whatever time was left. Professional sales and marketing would have done more for growth than any feature I shipped.

Build abuse prevention from day one. Monitoring for misuse and detecting copyrighted content should have been part of the platform, not something we learned we needed after a server went down.

Raise a seed round to support growth. Bootstrapping kept Splitstream lean and proved people would pay. But it also capped how fast a tiny team could build. After 2020’s momentum, with a proven paid product, that was the moment to raise, hire engineers and compete on features.