A recent trend is the advent of quite a few "lab" companies. Milk is the most relevant, Color is another. I think almost all of these are doomed to fail because of flaws in the core of the concept that prevent really big, stable companies from being born.
Switching too quickly
If a product dies within a labs company, the company doesn't die. I think this is actually a bad thing. It will be harder to get yourself and your employees pumped up about a product that may be scrapped in two months. You'll work fewer late nights.
It already is so hard to convince users to use your startup's product. Imagine if your users knew that this was just an experiment! For example, I refuse to use Oink because I know it may not exist in the near future. This could lead to a self-fulfilling prophecy. This is implicit with any startup, but becomes much more real when the founder himself discloses the fact that even he is not totally all-in on this idea.
Lack of execution
If you know another idea is just around the corner, less time is spent on the details. I won't name names, but I've been disappointed more than once by the lack of polish of products produced by labs companies. People run the fastest they've ever run when death is right behind them.
Hiring
Bring on awesome people is probably the biggest challenge of any early-stage startup. Raising money has become easier. It is a lot easier to hire someone that believes in your vision, and it is hard to have a vision when your goal as a company is to produce many things. PayPal had a clear vision, and for engineers that wanted to work on a revolutionary payments company, PayPal was number one. When you don't have a specific mission, your startup becomes one of thousands.
Equity imbalances
There is less incentive to join a lab company early. This makes hiring your initial employees more difficult. But it also leads to a skew in equity that may lead to significant drama in the company's future. Again, I won't name names, but email me if you want some concrete examples.
Mis-matched passions.
If someone is really passionate about mobile, and then their company decides that it is too hard to make money on mobile, what does that person do? Stick around, working on a product he doesn't like — or quit. I've seen many people quit labs companies for this reason.
Every startup is already a "lab"
A startup starts with one (big) idea. The idea doesn't pan out after a couple of years, and the startup fizzles out or works on something new. This is the path of a startup. I don't think you need to make this explicit.
It is already really, really, hard to build a big company. I think the founders, the employees, and the investors need to rally around just one powerful idea at the very beginning. You need to go all-in.
Sahil Lavingia (@slavingia) is co-founder of Crate and Rmmbr on the web, and Dayta and Color Stream on the iPhone.
