3 ms·
Ah, I suspect you're right. How is this solvable? Management engineering??
by nfoz 13y ago
Ah, I suspect you're right. How is this solvable? Management engineering??
- zaidf 13y agoI think ultimately it requires having a strong champion empowered to make big calls. In product development, what I have seen happen is that everyone has strong views about what must be included for a product to succeed. With no one with either authority or guts to take a stance for and against features, a bunch of average shit gets shipped in hopes of somewhat satisfying everyone instead of really satisfying a core group of customers.
- sirkneeland 13y agoit is solvable in organizations with people who have an enormous amount of credibility within the company. The most commmon examples are people who either founded the company or can credibly claim to have been their transformative savior (or in the case of Steve Jobs, both). Of course, founders have a way of either dying (Bill Hewlett & Dave Packard, and Steve Jobs again) or "losing their touch" (Mike Lazaridis of RIM). I would be curious of any examples in which there was such an undisputed leader at a company who was neither its founder nor its savior.
- nooneelse 13y agoIn light of the "management engineering" slant, the way people organizations seem to get made reminds me of the old meme that buildings in the far past were 'designed' only to the extent that they didn't follow the exact design of some other building which had fallen down. And in that vein, the answers by zaidf and sirkneeland, are like: buildings above some size keep falling down, so add a big central column in the middle to support the weight. Make all of the building depend on that central column. But buildings evolved to have arches and open domes, no central column needed. Why can't the same be done with people-based control/oversight/management systems? Building a beautiful structure dependent on all/many the members around, not just a central support/dictator.
- zaidf 13y agoJust because you have one person who makes the final call does not mean he isn't reliant on other members of his team. The way Gates describes his process resonates with me: In terms of deciding what programs are going to do, a fairly large group makes suggestions. Then there’s a filtering process. Eventually I’ll decide which of the ideas makes sense, and I’ll make sure we have champions who are personally involved in making that product succeed. http://programmersatwork.wordpress.com/bill-gates-1986/ http://programmersatwork.wordpress.com/bill-gates-1986/