I think code is just one tiny piece of the puzzle, but here's an example: I'm one guy who programs in his kitchen. I would up at 6:12 AM last Saturday and, unable to get back to sleep, decided to write a new feature into my website. I found it easier than I expected and eventually ended up writing two new features. Then I set up A/B tests for both of them. I then got done at 10:37 AM and spent the rest of my day playing videogames.
I was able to "write" two features in 4.5 hours because it was a matter of snapping two bits of OSS code together with a bit of UI glue drizzled on top. (The A/B testing framework is also OSS, though I wrote it.) My customers don't care about that, though: they care that they couldn't do X or Y last Friday and now they can. (Well, if they're in the A/B test group.)
You know what isn't anywhere in the above description? A four hour planning meeting between six people discussing whether X or Y should go into the next version of the product. I've been at that meeting before. I think we all have. It is a tremendous waste of time when I can just ship the feature in an A/B test and tell you a week later how 5,000 users reacted.
(Answer: feature X increased task success. Feature Y decreased it. Neither effect was statistically significant. As I had been delaying making these primarily out of fear that featureitis would hurt task completion, in the face of repeated customer requests for both features, I'll be shipping them to everybody.)
And I'm one guy doing this from his kitchen table in his spare time in a not-too-rich niche. There is nothing about these techniques that make them less effective for bigger, better funded, more focused teams in richer niches.
I think code is just one tiny piece of the puzzle, but here's an example: I'm one guy who programs in his kitchen. I would up at 6:12 AM last Saturday and, unable to get back to sleep, decided to write a new feature into my website. I found it easier than I expected and eventually ended up writing two new features. Then I set up A/B tests for both of them. I then got done at 10:37 AM and spent the rest of my day playing videogames.
I was able to "write" two features in 4.5 hours because it was a matter of snapping two bits of OSS code together with a bit of UI glue drizzled on top. (The A/B testing framework is also OSS, though I wrote it.) My customers don't care about that, though: they care that they couldn't do X or Y last Friday and now they can. (Well, if they're in the A/B test group.)
You know what isn't anywhere in the above description? A four hour planning meeting between six people discussing whether X or Y should go into the next version of the product. I've been at that meeting before. I think we all have. It is a tremendous waste of time when I can just ship the feature in an A/B test and tell you a week later how 5,000 users reacted.
(Answer: feature X increased task success. Feature Y decreased it. Neither effect was statistically significant. As I had been delaying making these primarily out of fear that featureitis would hurt task completion, in the face of repeated customer requests for both features, I'll be shipping them to everybody.)
And I'm one guy doing this from his kitchen table in his spare time in a not-too-rich niche. There is nothing about these techniques that make them less effective for bigger, better funded, more focused teams in richer niches.