6 ms·
I tested it back when I ran a dev team. We were faced by constant pressure to develop customer stuff faster, and the team were doing regular 60+ hour weeks. I
by marcus_holmes 13y ago
I tested it back when I ran a dev team.
We were faced by constant pressure to develop customer stuff faster, and the team were doing regular 60+ hour weeks. I put in some new rules on hours, including:
- no dev worked more than 40 hours on scheduled projects. Overtime was for disaster recovery or slippage repair to meet a deadline
- if the salesperson agreed to a new deadline, then we went over their other tasks and worked out what needed to shift to make room. If they insisted on overtime, then we insisted on lieu time to pay it back, from their scheduled project time.
- if someone came in to fix a disaster, we gave them time of in lieu. If we really couldn't do the time off, we paid them overtime for the time.
It took a few months for this to settle in, and I got to test the productivity changes (by measuring project estimates vs actuals). Productivity improved as the hours worked decreased.
The productivity gains came mostly from more realistic project estimates, but a significant part from increased dev productivity... they just wrote better code that needed less correction, faster.
I would post the data, but it's commercially sensitive and about ten years old now.
- grandinj 13y agoI'd like to second this - I ran the same stats once, while contracting for a customer with excessive demands. I found that a week of working long hours was about my max, and after that I started making mistakes to the point that the extra hours actually became counter productive.