What’s especially helpful to such nuanced and detailed debates is data. So my question is why isn’t data being brought out in support of the RTO argument?
Two pretty obvious potential reasons are that they don’t have data or the data doesn’t support the argument they’re making, which reflects rather poorly on management and their decision-making.
I’m personally skeptical of the explanation “they have the data and it supports their argument but they just don’t wanna show it to their employees as they beg them to come back to office.” Just makes a lot less sense than the other explanations IMO.
No one’s bringing out data to support WFH either, so I really don’t understand this demand. WFH at scale only has around 3 years of history where in office has decades and decades.
There are plenty of anecdotes by employees stating that they are definitely, for sure, 100% more productive at home. But I have seen very few by top level execs stating that this is true and that it also scales to teams, divisions, etc. Even the definition of productivity seems very, very loose in these discussions - is it that an employee feels more productive because they don’t have to commute? Because they write more code or better code? Because they can do the laundry at lunch rather than outside work hours?
On the other hand you have people saying “in office is more productive.”
Note how it only makes sense to ask one of those parties for data to substantiate their position.
There really isn’t a debate here that requires scientific studies backing it. It’s like asking whether Elon Musk had any studies showing X was a better name than Twitter. Companies can do what they want and employees can vote with their feet. If the job market weakens then it becomes a harder vote in the negative.
WFH is really no different a perk than free lunches or unlimited PTO. it will be offered when the labour market demands it, but I think we are heading into a weaker market and employees will be up shit’s creek.
That is irrelevant to the conversation going on here.
Imagine, for a moment, that I have collected metrics on the members of the teams that I manage that measure productivity against work from home, things like sprint burn down versus work from home, and things of that nature. Suppose that I have this information for 50 people. I can’t share it externally due to corporate policy, and I can’t extend its scope to other teams that don’t use Jira in the way we do.
I can use it to try to discuss WFH with my teams directly, but even starting this conversation is liable to upset WFH stalwarts, but if I do I cannot present this information in a way that is both deidentifiable and satisfyingly comprehensive. If I aggregate it I get something like “we’re 10% better at task completion while in the office”, I have people who claim to be an exception to the rule (and the loudest complainers are disproportionately actually the worst when it comes to WFH productivity), and if I present it unaggregated then I have people complaining about being called out publicly as under-performers (rightfully so) and people complaining since their productivity didn’t slip 10% they should be able to work from home.
I cannot manage 50 different WFH policies due to incoming complaints about fairness, mandating RTO upsets the team, allowing unrestricted WFH nets a productivity loss, and anyone who says “fire the people who are underperforming from home” knows nothing about management.
So my personal response here is to just allow the productivity loss: 10% productivity means less to me than the major impact on the happiness of the teams, but it must be recognised that this pick is the “best of bad options”, if you’re trying to maximise business output.
Clearly, at some level of management, the happiness is less important that the profit, but that’s beyond my pay grade.
It isn’t beyond your pay grade. If your team underperforms because of your decision, then you risk the entire team being let go in the next downturn.
I also have access to other more obvious task-tracking metrics like the number of pull requests, or Jira task completion, or stack deploys.
I also have team calendars where team members plug in whether they're working from home or from the office.
It's fairly trivial to punch a bunch of this data into BI tooling to get an approximation of how work from home makes a difference to how much actually gets done. Is it 100% accurate? Certainly not. Would I ever use it for an actual performance conversation? Nope. But it does help as a component of the gestalt that is my understanding of WFH performance.
The other major component of my understanding is more direct. I go out for lunch with my teams or for beers with my teams, and I listen to what they have to say. I know that a reasonable number of my team members head out at 3PM to pick their kids up from school every time they work from home, because they tell me that to my face. The guy with the new baby is not 100% focused on work when he's home, because he's got tiny human to deal with.
It's also very important to note that performance is squishy and hard to quantify at the best of times, before you even get into things like Goodhart's Law. For example. employee retention is important and as a rate of turnover fairly easy to measure, and is a significant factor for total velocity, but causes of turnover are almost always orthogonal to the root cause of job dissatisfaction... More importantly, though, do we decrease the negative impacts of turnover commensurate to the loss of productivity by allowing WFH? Who knows!
I have way way way more that I could say on the topic, but I might be being a bit hypocritical by continuing to talk about it while at work. ;]
I think in these discussions everyone oversimplifies to one single conclusion that should apply to all companies in all locales and all employees. There’s way more nuance to this.