Posted by Tom Grant on February 3, 2009
Saeed asks a great question: Why not have Development report to Product Management? PM is supposed to own the business argument for what Development is doing. Why not, therefore, make the development team formally accountable to the people who are supposed to give thumbs up or thumbs down to what the developers are doing?
Saeed is asking about the pros and cons of that arrangement, but undoubtedly he's urging us to question the unquestioned assumptions about technology companies. Provocative language like this has an old vintage. For example, a couple of centuries ago, the Enlightenment thinker Denis Diderot, trying to get shake up his readers' attitudes about the divine right of kings, said, "The arbitrary rule of a just and enlightened prince is always bad." Zing!
I urge you to read the comments section of Saeed's post. Here, I'll add a few thoughts of my own:
- Technical oversight is a major reason for the traditional arrangements. The people in charge of Development need to assess whether the team is doing a good job at technical tasks like builds, refactoring, and unit testing. The obvious conclusion: if you haven't been initiated into the Secret and Benevolent Order of Development, it's hard to supervise what the developers do.
- Some of the traditional arrangements are historical artifacts. New companies usually start with a couple of developers with a big idea. Many projects in big companies start the same way. Anyone coming later, like PMs, are going to report to the people already there, if they're going to join the product development effort.
- These arguments are not wholly convincing. Even if you're not a developer who has mutated into a product manager, it shouldn't be impossible to supervise teams with specialized knowledge you don't have. In the military, for example, you don't need to be an artillerist or a tanker to command them. Product management is its own dark art, so PMs have just as much a right to reject management by people who don't fully understand what they do.
The most important reason to take Saeed's question seriously is the high cost of making bad product decisions. Buyers were already becoming more discriminating; now, with the economic downturn, technology companies can afford even less to have Development working in its walled garden, only emerging occasionally to deliver some basket full of things that people on the outside may or may not want.
If shortening the cycle of market testing product testing will only happen with radical alterations to the structure and operations of technology companies, then maybe it's worth taking Saeed's question very seriously.
[Cross-posted at The Heretech.]