Everyone has business ideas. Some people quickly forget them, others remember and wonder what could have been, and then there are the crazies who spend 23 hours of the day, working before and after their 9-to-5 jobs (unless they are lucky enough not to have a 9-to-5), breaking their necks implementing their ideas, failing, succeeding, learning, bleeding out of their ears from stress and worry, and ultimately loving it. In this article I want to share with you my experiences of 10 Web companies I have started in the past and what happened to them.
Attempt No. 1: Virtual Socrates
My first attempt at entrepreneurship was a long and winding failure that started during college when I was studying computer science and realized I was really interested in philosophy as well.
Problem I tried to solve: I wanted to make it easy, fun and exciting for others to learn philosophy.
Attempted Solution: Automate Socratic dialogs so software would act as Socrates and the user would try to discuss the topics of the original dialogs. This interactive and simpler way of learning the core concepts in philosophy was going to make people more interested in philosophy.
Failure Points: The usability of this project turned out to be the toughest to implement. I partnered with a very senior, but kind of flaky engineer, and we spent about a year implementing the solution. The complexity of the project grew as we were developing it, business direction was lacking, and the project was never released to the public and collapsed after my partner left the project.
Now I can laugh about it because my current approach to starting projects is to have one person be able to complete the Proof of Concept in less than three months. If it takes longer, a red flag goes off in my head.
Attempt No. 2: Semantic Shopping and Fashion Search
One thing that came out of the philosophy project was the monetization question that would not really be addressed by doing just philosophy. Towards the end of that project I had realigned it to focus on ecommerce, now dealing less with Socratic questions of "What is virtue?" to something more like "What shoes are cool?"
Problem I was trying to solve: Rigid and "unnatural" current online shopping experience where people can't find items according to their subjective taste.
Attempted Solution: Semantic Web implementation of a search that "understands" subjective concepts like cool, sexy, and a number of styles in order to help people shop in a way that is closer to how they would naturally think of the items they want.
Failure Points: By many counts this project was a success. I put together a great team of real experts, we created the semantic fashion search, and the quality of the search was slowly improving. The project was even making money helping people shop! The problem we ran into was acquiring new shoppers in the hyper competitive apparel market. Plus having a product that was 20% better than something like Amazon search that currently exists was not enough to get people to switch over how they shop.
The technology was also not simple and we spent too many of our resources maintaining and refining the very complex semantic search consisting of NLP and ontology technologies. Plus sales margins were razor thin and there did not seem to be a light at the end of the tunnel in terms of generating enough profit to maintain a 3-5 person team. Ultimately, the project did not survive.
Attempt No. 3: Selling the Semantic Fashion Search to Established E-Commerce Sites (B2B)
We tried to salvage the situation by packaging the search to help bigger companies add an interesting wrinkle to their existing sites. This was quite an uphill battle. Meetings were difficult to get, sales cycles were painfully long and suffocating even in the best cases. After a number of partnerships (not clients) we realized we did not have the resources to see these partnerships through.
What made the sales cycle so difficult in this case was that we did not have a product that was mature enough to really excite the customers. And a big enterprise sale, everything has to go right: scalability, much improved customer experience, etc.
Lesson learned: Understand available resources and make sure the ambitiousness of the project matches the available resources (this was pretty much echoed in every project).
Attempt No. 4: Blogging as a Business
Blogging is a very interesting business. The best bloggers are great at writing great content, SEO, picking the right niche, and being patient, as there is quite a bit of work that goes into creating a well-read blog before it takes off.
My blog was about the Semantic Web community in NYC. It did reasonably well, and an interesting nuance of the blog is that an editor from DevX noticed the blog and invited me to write for DevX. Due to copyright issues I decided to discontinue the blog and focus on my DevX writing.
At DevX I get to have a job where I do what I love: write about startups and technology, so I consider the blog have been a success as it has gotten me here.
Had I continued with my blog, I would have likely found that my topic had been too niche.
Attempt No. 5: Startup Consultancy
Something that occurred naturally during all my startup work was that I was gaining all sorts of very rounded skills and was learning management, developing a business savvy that was complimenting my existing programming skills very nicely. This caused a number of people to come to me for help. As a rule I am always happy to help, but after a while I just got overwhelmed and had to start a small startup consultancy which made a little money, let me do what I loved to do: work with startups and help them grow.
A major problem was that I was helping these entrepreneurs do everything free and fast, and since I preached keeping risk low and spending little, this made my pay pretty low. Plus, I never felt comfortable taking money from people who didn't have much to begin with. So I did not attempt to grow my practice, but still always offer to help as much I reasonably can.