Hacker News Reader

Comment by hrabago | Hacker News Reader

This sounds like unit testing. It sounds good when you hear about it, and you think you probably should be doing it, but you figure you're okay even without it.

And then you actually try it and then BAM! you realize its value and the difference it makes.

View on HN

The first time I tried CD on a new product idea, I quickly realized that my idea was wrong. I can't thank Steve enough for saving me the weeks/months of development I would normally have put into a project only for it to be mediocre performer in the market. I honestly wrote no software at all for the first time, called/met with a ton of people, and learned a lot in the process about how wrong ideas can be :)

I feel like this mentality changes you in ways you can't imagine until you do it. Recently when approached by a local businessman about a really lucrative project instead of doing the coding internally, I simply outsourced all the customer development through elance. So far results are everything we thought they would be and more. And I've done almost no work.

This is lightyears ahead of where I was years ago - I remember blowing hundreds of thousands of dollars on projects that never even got to market! Now I'm spending ~500$ to outsource customer development (in specific cases - depends on business model of course) and getting better results.

I would really be interested in hearing something specific about one of these projects, as a kind of contrast. When you say you outsource the customer development onto Elance, what does that mean? Isn't the customer development the part you have to do?

For example, come up with a quantitative market hypothesis (e.g. X% of Y businesses contacted by method Z in will be interested in buying the "software". Also problem statements etc etc etc - anything quantitative. We setup a simple data collection form (a-la kissmetrics) for the minions at elance and set them at finding out the answer. They generate target lists (web scraping etc), contact them, collate the results and the rest is pretty simple. I've done it in stages so it's not all the same people doing the work.

Now we are starting to outsource the development of a repeatable sales process, which will be an interesting thing. It'll be the first project I've done almost no work on, but could be the most lucrative!

So we are still doing some of the CD, but the hard part (getting out there and finding early adopters, interviewing them, collating feedback etc) is automated.

Could you maybe write a blog post with a concrete example of a project you researched and rejected like this? I'd love to read that....

Yeah, I'd like to read that, too.

I'm not the blogging/twitter type. Never done it actually, and don't plan to. I'd rather be doing things than writing about doing them :)

Thats confusing to me also, did you mean the product development? I could see outsourcing the coding would be easier then having people run around and interview people for you.

Exactly like unit testing - Eric Ries likes to talk about customer validation as an additional step added to TDD or testing in general - for example a new feature: to constitute progress, first it has to work, then it has to matter.

Reply on news.ycombinator.com