In the heady days of the first dotcom bubble, I left my safe, well-paid job with a
management consulting firm to join a start-up as the technical lead. The founders had
just secured funding from an incubator, had a great product idea, and needed someone to
help them build the web site. It was great! We worked crazy hours, cut deals with
suppliers, found developers and designers and a hosting provider, and generally did
what start-ups do. We went from starting the project to go-live in 10 crazy, crazy weeks,
and I’m reliably informed that some of the code I wrote back then is still in use today,
billing customers for their order.

I can’t emphasize enough how crazy our plan was. We designed and built a “mass customization”
web site, taking real money from real customers, in 10 weeks. We loaded tens of thousands
of products into our ecommerce catalogue, hooked the site up to a payment provider, and
deployed the whole thing to a brand new hosting facility. We tested the order process
to make sure the products the customer ordered came out to the fulfillment team in the right
way. We ran security, performance and scalability tests.

In 10 weeks.

That sounds crazy, right – but we did it. And (mostly) loved every moment of it. However,
one thing I didn’t enjoy much. Even as we were building the site – the financial engine
behind our company – we found ourselves in endless wrangles about what features we should
build. Quite often, there was a strong feeling we should include those features in the
first release. And that’s exactly what it was – a “feeling”. In the end, we included
a lot of cool, interesting, eye-catching content and functionality. We spent a lot of
time and energy on building it, mostly based on “but maybe, if people find this content
and enjoy it, they’ll buy something from us!” statements.

Guess what – they didn’t.

Guess what else? We didn’t put much energy or thought into measuring our users’
interaction with the site, so it took us months to find out that all this cool stuff
wasn’t attracting any visitors, and that the visitors it did attract nearly always left
without buying anything.

So, once we had learnt this lesson, the conclusion was inevitable. “People aren’t interacting
with this content because they can’t find it! Make it more prominent on the home page!”.
Our site was built in a hurry, by a team of people who didn’t really know each other,
so making this kind of change was expensive, and fairly painful. But we did it – again,
many late nights, many pizza dinners, and heroic amounts of caffeine.

We put the new homepage live, with a prominent link to our cool content. It worked – we
got more traffic to our content section. However, our sales dropped quite quickly – it
turned out that the idea that people who browsed our content would end up buying something
from us was not strictly true, and certainly not over the period we measured.

So, we put the old homepage back, and sales recovered.

Not long after our launch, the dotcom bubble crashed, and our prospects of getting a
second round of funding looked rather bleak. We did secure more money, but nowhere near
what we had anticipated; after a few months, I left because it was clear the company’s
short term finances weren’t going to support developing cool new technology, and I felt
it would be unfair to take a salary in those circumstances. They’re still around, though,
and – arguably – the market leader in their segment!

When I read The Lean Startup, this story came
back to me. I think we all struggled with allocating our precious resources in those days;
a lot of our decisions were based on plausible stories, rather than hard analysis. The Lean
Startup ideas help you avoid that trap – I’m sure our “Minimum Viable Product” would not
have included many of the features we launched with. It’s not a surprise they’re no
longer present on the current site…

Another book that helped me think this through is
Thinking, Fast and Slow, which explores how the human mind makes decisions. It was
obvious that we were substituting the question “is it a good idea to invest time and money
into feature x” – a hard question – with the easier question “can we make up a plausible
story that shows investing in feature x is a good idea?”

Updated: