What was Semantria?
Semantria was the world's first massively scalable sentiment analysis API for processing social media content. All kinds of companies that were building apps on social data were using our API for sentiment, classification, named entity extraction, etc. We would add very valuable metadata to every tweet or Facebook post they'd have going through their pipes.
How did you get started?
I was 21 and high on life from being a hot-shot marketing and sales guy for a growing company. My financial situation was so-so, and I was getting bored with my repetitive job. I really wanted to start my own thing, and I kept on aimlessly brainstorming with people.
Initially I came up with a totally different idea — a Salesforce app that would process client and customer comments for sentiment and tell you which customers are unhappy. Completely accidentally I ran across an owner of a Ukrainian outsourcing company who believed in me and gave me a team to work with. We spent a year building the product that nobody would use — the Salesforce plugin. Afterwards, I realized that nobody wanted the Salesforce piece, but people were asking to connect to the API that powered it. Hence, I decided to throw out the plugin and focus on the API.
Also, right around the API pivot I got in a massive accident and had to execute the pivot from hospital bed for 3 months. I was mixing morphine and conference calls.
Where did you find the time and funding to work?
I was working nights and weekends. Funding came from all of the credit card debt I could get my hands on. I would also fix computers for people I'd find on Craigslist during nights and weekends.
This lasted for almost a year until we started getting our first revenue and I could pay myself a little bit after paying my people.
How did you grow? What did you do for marketing?
I was one of the first adopters of marketing automation in 2007, so I used Eloqua to fuel growth. We went to market with very sophisticated marketing automation routines that fueled growth for years to come. We tried pretty much every channel. We did a ton of PR for pay-per-click, and my company is still on the first SERP page on Google for "sentiment analysis." We didn't have any kind of official launch. Over time I started spending less time on PPC and focused more on SEO and personal networking. We ended up growing mostly organically.
How did you deal with incorporating and other legal issues?
I Googled it and incorporated myself. Don't do that. I would have paid 50% less in taxes in the end had I hired a proper lawyer.
How did you monetize? What was your revenue?
We made recurring revenue from API subscriptions. It took half a year to realize that selling one-time buckets of credits doesn't work for an API business.
Revenue growth was bumpy, because it depended on signing several huge customers. It peaked at ~$150k MRR when I sold the company.
I reinvested all of the money, except for paying myself $5K/mo and slowly paying off credit card debt.
What were your biggest challenges, fears, and mistakes?
The biggest challenge was simply not knowing what I was doing. Also, my health after the accident — physio was sucking up 5-10 hours a week and a lot of money.
My biggest fears were that I'd fail and have to declare bankruptcy. One bad move could kill the company and take me down.
My biggest mistakes were with hiring. First, I hired all my good friends. While this helped initially, it damaged some of the relationships. I also hired too many interns and trusted them with critical stuff. I couldn't afford experienced people until the very end.
What tools, apps, resources, skills, techniques, and habits were most useful to you?
I automated a lot of stuff via direct API integrations. I used Eloqua, Google Analytics, Google Apps, Ducksboard, Trello, Zapier, and Freshbooks.
I didn't have mentors, nor did I read too many books. There wasn't any time. I just kept on pushing through.
I paid contractors with handwritten checks, but having someone manage my schedule and financials was a huge help.
What's your advice for hackers setting out to be their own boss?
One: don't be afraid to lose it all.
Two: bootstrapping is not the fastest way to grow your business.
What was your tech stack?
We used Java and PHP.