My company started doing a serious WFH 'experiment' about 18 months before COVID hit. The team I oversee consists of a dozen or so teams of about 40-50 developers each, with a manager that reports directly to me. I don't manage the devs directly, but deal with dev productivity and budgets for (amongst other things) staffing. So our N is not small, but not huge either.
50% of our devs - randomly selected - were offered the chance to WFH, and about 95% of them wound up doing WFH at least 3 days a week, rising to 5 days a week after a few months. The other 50% stayed in the office. The teams that went WFH and the other 50% don't work on the same projects or in the same location.
After about 6 months, we started looking at productivity metrics using a couple Pinpoint-like tools that we built. Simple metrics like: backlog change time, on-time delivery, workload balance, LOC/Checkins ratio, etc. We also tracked things like non-code calendar hours, time logging delays, and other non-development related activity.
Neither group knew what we were tracking or how often. The results were pretty clear as soon as 6 - 9 months in: WFH numbers sucked pretty hard. Everything was down across the board, and it seemed like WFH devs were spending 15-30% more time in non-development tasks, their PR reject rate was as much as 40% higher than office-based devs.
Things were so bad that if COVID had not hit, we were going to abandon the idea of letting people WFH altogether outside of extreme circumstances.
We aren't a unique shop. We're not doing anything out of the ordinary - we're developing on a variety of web and RT device platforms, but doing the same stuff in similar ways as other companies. We have a pretty flat team structure, and teams have a lot of flexibility and decision-making power, which has always worked well for us.
However, it's clear to me and the other execs that WFH absolutely does not work for us.
50% of our devs - randomly selected - were offered the chance to WFH, and about 95% of them wound up doing WFH at least 3 days a week, rising to 5 days a week after a few months. The other 50% stayed in the office. The teams that went WFH and the other 50% don't work on the same projects or in the same location.
After about 6 months, we started looking at productivity metrics using a couple Pinpoint-like tools that we built. Simple metrics like: backlog change time, on-time delivery, workload balance, LOC/Checkins ratio, etc. We also tracked things like non-code calendar hours, time logging delays, and other non-development related activity.
Neither group knew what we were tracking or how often. The results were pretty clear as soon as 6 - 9 months in: WFH numbers sucked pretty hard. Everything was down across the board, and it seemed like WFH devs were spending 15-30% more time in non-development tasks, their PR reject rate was as much as 40% higher than office-based devs.
Things were so bad that if COVID had not hit, we were going to abandon the idea of letting people WFH altogether outside of extreme circumstances.
We aren't a unique shop. We're not doing anything out of the ordinary - we're developing on a variety of web and RT device platforms, but doing the same stuff in similar ways as other companies. We have a pretty flat team structure, and teams have a lot of flexibility and decision-making power, which has always worked well for us.
However, it's clear to me and the other execs that WFH absolutely does not work for us.