Paul Graham penned this great
article on how to get a startup going (courtesy of Ross Mayfield
of SocialText fame).

It definitely reminded me of what my “dream” startup would entail. I’ve had
the opportunity to work with three classic startups so far. One spectacularly
flamed out – I keep the option certificates they started granting us in lieu of
paychecks framed on my office wall as a reminder that it could always get worse.
One peaked, tanked (most of the company got laid off on the morning of
9/11…ouch), and is now seemingly coming back to life. Another stayed tiny and
exited on the right foot, and is how I ended up at my present company. All
taught me some lessons about what I’d like to work on in the future.

1. It needs to be “social” – an abused word these days, but still valid. I
want to work on something that helps people discover each other, interact, and
make their lives more interesting. “Interesting” should NOT equal more
stressful, which I fear many of our social technologies tend to do by layering
more data on people without simplifying anything.

2. It needs to be something I can explain to my Mom (even better, that she
would want to use) – I love what I do now, which is more on the enterprise side
of the shop, but if I were to do my own, I want to be able to demo it to people
only vaguely aware of computers, and have them still “get it” and love it.

3. It needs to actually simplify and enrich your life, not force you to learn
something entirely new or change your way of doing things (to point #1). Too
many new technologies force you to change to their model, versus working with
you. Newsgator is a fantastic example of
a good play here – it simply plugs right into my Outlook, which is where I spend
90% of my day anyway. Look at how people work and play today, and offer them
something that enriches THAT experience, rather than expecting them to adopt an
entirely new one.

4. It needs to be a fun environment – life is too short to spend your whole
time obsessed with metrics, process, and 80 hour work weeks. Work hard and be
disciplined, of course. But if it’s not fun, why spend time away from your
family because of it?

5. It should be relatively virtual – collaboration technologies are simply
too good right now to force yourself to be bound to a specific office or
geographic location (for software companies at least – this definitely isn’t
applicable to most industries). Telligent
and SocialText are two great
examples of employees collaborating from around the country/globe. Plus, having
everyone in the same place tends to result in endless, needless status,
planning, and review meetings. Because you can meet, you do, and efficiency and
productivity suffer.

6. The technology and business model should support and encourage others to
build off it and extend it. To use the Microsoft buzzword du jour, it should
support an “ecosystem”. It should be something that has an open API and flexible
licensing, plus be viral enough (wait…buzzword bingo!) – and cool enough
– that people want to play with it. Flickr
is a cool example. Firefox extensions is also a nice idea on this front. Get
other people contributing to the success of your product by making it their
own.

7. There should be a free component. Not a 30–day trial, but something
lightweight and free that is immediately useful to your customer. Give em a
little, help them get dependent on it by keeping it fun, cool, and useful, and
then offer them something even better for a fee. In today’s noisy online world,
you need something for people to talk about. If it’s only available for a fee,
it’s hard to get noticed (with many exceptions…still valid I feel).