Hacker News Reader

Comment by donw | Hacker News Reader

I'd argue that the prototype is worthless, unless it's validated by an active user set and/or profitability, and has been through at least one iteration where features were added or changed.

I've seen a good number of technical prototypes built by non-technical people, nominally via outsourcing, and in every case the only path forward was a full rewrite. Crazy interdependency and tight coupling make it easy to build a 'demo' version of a codebase, but they also make it very expensive to bolt on new features.

Similarly, I think the percentages are high, and don't account for any business skills that the technical founder brings to the table -- what about a tech guy who's gone through the VC dance, and built successful products in the past?

View on HN

I see what you're saying about the value of the prototype. How would you account for whatever risk the non-technical cofounder took on to build it? Risk has to coordinate with value added though, so if a prototype didn't really generate any new value, it's worthless. Perhaps there is some value in "lessons learned" from messing up the initial prototype, but maybe future investors/partners shouldn't be penalized for them.

You're right about not accounting for business skills the technical cofounder brings to the table. I need to state that somewhere. Great point.

Wow, you're on the ball. :)

We have a very convenient way of expressing value here. If the prototype has proven solid enough to support an active and growing user population, then it has value, because it's cleared the first hurdle of software design -- making something people want and can use.

A few iterations on said prototype before bringing in a technical cofounder is also a big plus-point. Those iterations are where you find out whether or not your prototype will be flexible enough to turn into a real product.

If it's a good prototype that proves the concept and acts as a target, then it mightn't matter if there's any reusable code there.

I know a few people who have a startup that's based on the biggest mess in the world, that took off like a rocket when it was launched, and pulled in a big pile of funding.

Problem is, now they need to change direction slightly, and they can't, because their codebase and underlaying data model were built without anybody asking the question, 'So, can we fix this later?'

I'm all for throwing a rough prototype together that's buggy as all hell and lacks features, but don't build bits of it that you can't easily change later. This isn't really as hard as it sounds, but it takes foresight from somebody who's built a lot of crap over their career and seen how their bad decisions made things fall apart. :)

Reply on news.ycombinator.com