Saturday, May 18, 2013

EBD: the right way to sell Agile

For quite some time, I've been summarizing agile development as "early and frequent delivery of value." But it has struck me that skeptics have a good reason to be unswayed by such a claim. Of course "early and frequent delivery of value" would be nice. Ditto "easy," "seamless," "fun to use," and all those other claims every product and methodology has promised since the beginning of marketing. It's quite right to treat such promises as so much hot air. What matters is what you deliver.

So instead, let's ask a few easy questions.
What's better for making decision, more evidence or less? 
Can I assume we all agree here? Next question.
What kinds of evidence are better:
  • user surveys or user observations
  • historical data or last week's results
  • expert opinion or actual measurements
Still with me?
Which kinds of evidence can you get from:
  • requirements
  • high-level design
  • implementation
  • testing
  • deployment
Final question.
How does the above flow of evidence compare to the evidence gained from weekly user tests on working code?
And this is how you explain agile to skeptics: It's EBD -- Evidence-Based Development.