ClayValet in hindsight
I just shut down my startup, 22 months after I started it in January ‘07. I want to tell you how I decided to close the company, what problems we encountered, and what I’ll change next time.
ClayValet was an attempt to build a shopping recommendations business on top of Amazon’s Mechanical Turk, funded with seed capital from one angel investor. We created a website for shoppers to ask for and browse recommendations. Unlike search engines, our service was powered by real people, so we could answer ambiguous and detailed questions (within a day), and we could filter out useless answers.
Our website did not get much traffic. After several revisions aimed at improving usability and adding content, the metrics improved, but growth was still too slow. We tried to convert our recommendation service to a white-label product and sell it to online retailers, as a way to engage their customers and convert more sales. I didn’t close any sales. I explored the possibility of raising another round of funding to hire a sales team and develop this product, but our lack of traction was a problem for investors.
With a few months left in our runway, I focused on finding a potential acquirer for ClayValet. I contacted a couple dozen companies, and ended up having substantial conversations with three of them. In the space of a week, two of them went from lukewarm to actively interested. At the end of that week, all three told us that they would be happy to hire us, but “can’t structure the deal as an acquisition at this time.”
We could have continued to find acquirers, sell our service, or fundraise. At this point I’d tried all of the above, and didn’t think we had a good chance of success. Rather than burning through the rest of our funding, I decided to close the company and return the remaining cash to our investor.
Overlapping learning curves
We were doing too many things for the first time. Before ClayValet, none of us had built a consumer website, promoted a consumer website, attempted enterprise sales, or run a company. It’s possible to learn everything, but learning is not instantaneous, and it detracts from focus and momentum. I found it hard to plan for a process I didn’t understand and to hire for positions I didn’t understand. One of the best bits of advice I’ve heard lately is “do mostly what you know, and change as few things as possible”.
The cold start problem
I drastically underestimated the difficulty of creating an audience for a new shopping site. Our recommendation service was new and often impressive, but it was in the end a feature and not a full shopping experience. Websites are easier to build today, but expectations are higher. Early user feedback slammed us for not having features that people had come to expect: conversations, profiles, a wider catalog, instant search. Shoppers wanted us to email them more. Shoppers wanted us to provide a more interactive service. Even worse, shoppers forgot about us because they didn’t often have the problem we offered to solve.
The process of enterprise sales
You want to have an “in” with your customers and a clear benefit, and even then the sales cycle takes months and a lot of engagement with the customer. We were pitching an interesting but unproven feature that didn’t obviously solve a problem they cared about. We figured that out and adjusted the pitch, but we still didn’t have a credible track record to back it up.
TO DO NEXT TIME
Running a new company is a lot of work for one person, and having a great cofounder can drastically reduce risk. I would look for someone intelligent, trustworthy, pragmatic, and responsible, with experience that complements mine. I believe that the choice of cofounder will be the most important decision in my next business, followed by the choice of market.
In doing ClayValet, I met many experienced businesspeople: entrepreneurs, investors, domain experts, and more. I sought out their help, and when I ran into difficulties I’d ask a couple of people for their advice. This was much better than no guidance, but it was too sporadic: the people I spoke with only knew about my immediate questions, and they couldn’t help with “big picture” decisions because they weren’t observing the company regularly. Next time I’ll form a board of advisors who review our progress every month.
After doing so much for the first time at ClayValet, I can plan with more confidence. For example, I’m still no expert on enterprise sales, but I know roughly how long it takes, how buying decisions are usually made, and the kind of people I would need to help me. As always, there’s a balance: excessive planning impairs flexibility, lack of planning magnifies risk. I’ll be closer to the right balance next time.
Specific revenue model
In my initial business plan, I will have a specific target customer for our first sale. Waving my hands about markets and demand won’t cut it.
In my initial business plan, I will have a specific distribution channel for reaching the next group of customers after the first sale. The distribution channel will not be “cold calling”, “word of mouth”, “organic traffic”, or TechCrunch.
I’m wrapping up some ClayValet business and deciding what to do next. I’m considering starting a new project, joining a local startup, or taking on some consulting projects.