Ready, set develop: How to create a six hour startup

Ready, set develop: How to create a six hour startup

Posted by John Cook
October 18, 2007 4:37 p.m.

Earlier this month, Brian Dorsey set out on a mission: develop and launch an Internet service in less than six hours.

PictureThe 33-year-old Seattle software developer wasn’t looking to set a
Guinness World Record. He just wanted to explore whether something
worthwhile could be built on the Web in a short period of time.

Brian Dorsey

Dorsey — who enlisted the help of about a dozen acquaintances from
entrepreneurial groups in town — nearly made his target. About seven
hours after the group gathered in the living room of his Phinney Ridge
home on the morning of October 6, Tagmindr — an online bookmark reminder service — was born.

Call it a startup on speed.

Dorsey’s project, part of an effort by a group of entrepreneurs and developers known as Seattle Saturday House,
might be an extreme case. But it speaks to a larger trend. The cost and
time it takes to build new Internet applications is shrinking rapidly
— changing the dynamics of the technology industry which for years has
been plagued by slow development times and countless hours of product
testing.

These new projects (it is a little hard to call them full-fledged
companies) take an entirely different approach from the software and
Internet companies of the late 1990s. The new ethos among
entrepreneurs: throw something out there quickly and see if it sticks.

Getting an Internet service off the ground has never been so easy,
with open source software and low-cost Web site hosting available at
the click of a mouse. Advice, support and — in some cases development
talent — can be easily obtained through blogs or online groups. And
new online platforms, such as Facebook and MySpace, provide fertile ground where entrepreneurs can immediately test the value of their services.

Of course, turning these small Web applications into meaningful
businesses is an entirely different matter. And it is unclear whether
any of them will ever make money. But this new breed of entrepreneurs
— armed with little more than an idea and some coding skills — are
quickly trying to create new features that enhance online photo, music
or storage services.

It didn’t take much to get Tagmindr
up and running. Just an initial idea — which Dorsey came up with
earlier in the month and couldn’t get out of his head — a domain name,
and a few smart people. All told, about $30 was invested in the
project, including the snacks that Dorsey purchased to fuel developers
during the coding marathon.

In creating an online application in an afternoon, Dorsey said he
simply wanted to get more practice in product development and
programming.

"It was kind of the idea of fail early and fail often," said Dorsey,
who was astonished that 12 people devoted a Saturday to his project.
"Ironically, I think the Tagmindr one was amazingly successful."

And Tagmindr isn’t the only quick start startup to launch in recent weeks.

Mark Maunder recently created Feedjit.com
— an online application that allows bloggers to track the geographical
location of users — in less than half a day. Powered by one "finely
ground" cup of coffee, the Seattle developer worked from 4 p.m. to 2:30
a.m., noting in a blog post that it took him 10.5 hours from the "first time my hand touched the keyboard until I fixed the last bug."

After spending three months creating LineBuzz and eight months on GeoJoey, Maunder said he just wanted to have some fun with Feedjit. So he set a 12 hour time limit on the project.

"You can get obsessed with picking just the right colors or layouts.
I didn’t worry about that," said Maunder. "I said: ‘this is just going
to be functional’ and I got it out the door."

That’s also what has driven Ben Curtis, a freelance software
developer who spent a portion of his time earlier this year on an
online job applicant tracking service called Catch the Best.
The 31-year-old sunk about 125 hours and less than $5,000 into the
project, setting a launch deadline for the end of September.

"I am a developer so, by nature, I want to add the gold-plated
features and continue and continue and continue," he said. "(A
deadline) pushes you to get something done and get some feedback on it."

Both Curtis and Maunder worked alone, a strategy some believe is the best approach for these quick efforts.

"I think the bigger the team gets, the more challenging it is to
execute quickly," said Maunder. "I’ve been in so many companies where
things just seem to move so incredibly slowly. I generally move quicker
on my own or with really small teams."

Dorsey agreed that a larger team could derail a project, saying that
Tagmindr probably could have been created by one very focused person in
six hours. Multiple contributors also open up potential problems in
terms of ownership of the idea.

"Being a developer, that was the thing I had thought the least about
upfront. And, in that respect, I have already learned a lot from this
project," said Dorsey, adding that Tagmindr’s code was released with a
liberal open source license.

That wasn’t a problem for Maunder, who is now trying to keep up with demand.

More than 5,000 bloggers in over 100 countries have adopted the
Feedjit service, an adoption rate that surpasses that of Maunder’s
previous three startups which he spent months building. Feedjit is also
attracting more than 11,500 page views a day.

"It is weird thing because the faster I execute and the faster I get
it to market, the more successful it seems to be," he said.

Now, Maunder is tinkering with ideas of how to build a business
model around Feedjit. He wants to expand the concept to provide more
real time statistical information to bloggers.

"I didn’t set out with the idea that it is going to be a long-term
thing," he said. "It was just a one or two day experiment, but it
gained a heck of a lot of traction and it has gone viral."

In its second month, Maunder is already dealing with the "geometric"
growth of the service. He built the site using all open source
technologies — from the programming language to the Web server to the
database engine.

"You don’t need to raise $1 million so that you can buy Oracle or
buy a license for three Windows Web servers," said Maunder. "The only
thing you pay for now is for your hardware and your bandwidth." He
estimates that it took $690 to build Feedjit, including $10 for the
domain name.

And what was the big lesson that Maunder learned during his 10.5 hour development push?

"Get your product, your idea, your service to market as quickly as
possible and get feedback from customers," he said. "If you work in a
vacuum for six months or a year without getting that feedback, you tend
to … forget about the customer."

Curtis — who sold an online conference application last year for a
slight profit — added that he’s learned the importance of developing
products that matter to people.

"The big trick is not to build something," he says. "But to build something that people will actually pay you for."

>BackTrack

Leave a Reply

RSS Daily Search Trends