It's not a productivity problem. It's a connectivity problem.
Another week, another bank calling its people back to the office. NatWest has told around 13,500 "remote-first" colleagues, who currently come in twice a month, that from February they will be expected in at least twice a week. A further 17,000 hybrid workers have been asked to come in more often. Barclays delayed its own move to three days after staff pushed back, and TSB staff are preparing legal action over a three-day requirement.
I understand the temptation. If you lead an organisation, the office feels like the one lever you can pull. You can see who is in. You can feel the energy in the room, or its absence. And when results are under pressure, wanting your people where you can see them is a very human instinct.
But before more leaders follow suit, I want to offer some food for thought. Because I think many organisations are answering the wrong question.
The question leaders are asking
Most return-to-office decisions are framed around productivity. The worry is that people working from home are doing less, and that bringing them back will fix it.
The evidence does not support that, at least for hybrid working. The most rigorous study we have - a randomised trial of more than 1,600 staff at Trip.com, published in Nature in 2024 - found that people working from home two days a week were just as productive, and just as likely to be promoted, as colleagues in the office full time. Resignations in the hybrid group fell by a third.
There is research suggesting fully remote working can cost output, particularly for junior staff, so this is not an argument for never coming in. But here is the more uncomfortable truth: most organisations are not measuring productivity well enough to know whether they have a problem. Many decisions are being made on how things feel, not on what the data says. And for some leaders, if we are honest, it is harder to trust what you cannot see.
The question they should be asking
In my work with organisations across sectors, I am not seeing a productivity drop. I am seeing a connectivity drop.
By connectivity I mean two things:
How connected are people to their colleagues - do they know each other, trust each other, learn from each other?
And how connected are they to the organisation itself - its purpose, its values, the reason it exists?
Both have taken a hit since 2020, and that is a real problem. It shows up in weaker collaboration, slower onboarding, thinner loyalty, and junior staff who never get to learn by watching how experienced colleagues handle a difficult client or a tricky decision. That last point is the strongest argument the banks make for bringing people back, and it is a fair one. But it only works if the experienced people are in the room too, on the same days.
Which is where a lot of mandates fall down. I recently spoke to someone who travels from Manchester to her office in London because she has to be there two days a fortnight. When she arrives, hardly any of her colleagues are in. She could have been more productive at home, and she leaves feeling the whole thing is performative. That is not connection. It is attendance.
Why the current fixes are not moving the dial
I have seen organisations spend hundreds of thousands of pounds redesigning their offices to tempt people back: new furniture, breakout zones, hot desking. In most cases it has not moved the dial.
The reason is that they redesigned for individuals when the problem is teams. Hot desking can even make connection worse. If you cannot count on sitting near the people you work with, why make the journey?
Blanket mandates carry a further risk. Your people may simply say no thanks and go down the road to a competitor. The people most likely to leave are often those with the most options, and those for whom the commute costs the most - parents, carers, people who live a distance from the office.
The cost of losing them is well documented. Oxford Economics estimated that replacing an employee cost more than £30,000, most of it in lost output while the new person gets up to speed. That was in 2014, so the figure will only have risen. For public and third sector organisations, which already struggle to compete on pay, the risk is sharper still.
None of this means you should not ask people back. It means that before you do, you should have worked through who might leave, what it would cost, and whether the change will actually solve the problem you think you have.
What to do instead
If connectivity is the real issue, invest in connectivity. Some of the organisations I work with are already doing this well.
Make the days count - One client brings everyone on site for two days each quarter to plan. Teams work through what they are each doing, where it overlaps and what they need from each other.
Coordinate who is in - Another publishes a schedule of which teams are on site and when, so people come in to see the colleagues they actually work with. Those days are spent on internal problems, not client work that could be done anywhere.
Design space for teams, not desks - If you are redesigning, put the money into places where groups can work together, not more seats for people to sit alone.
The common thread is that the office becomes somewhere you go for a reason, not somewhere you go to be seen.
There is no settled normal yet
Six years on, I am not sure we have settled on a new normal for how we work, and I do not think we will for some time. Technology will keep changing what is possible, faster than most policies are written.
So rather than fixing a number of days and defending it, treat this as something you design, test and adjust. Before you decide how many days people should come in, decide what they will be doing together when they are there.
Sources
NatWest forces thousands of remote workers back to the office - NatWest numbers, Barclays delay, TSB legal action
Unite opposes NatWest return to office plans - Personnel Today
Hybrid work is a win-win-win for companies and workers - Stanford, on the Trip.com study in Nature
It costs over £30K to replace a staff member - Oxford Economics, 2014
Comments