The answer everyone gives, and no one has measured
The collaboration figures are green. The decision is still open. Only the first is measured, so a decision that will not land looks like a coordination problem — and the remedy is another forum, billed to the people already in everything.
Somewhere in your quarterly operating review is a slide about how well the organisation works together. It is probably green.
Cross-functional forums are up. Response times are down. Participation in the planning cycle is in the nineties, and the engagement survey’s collaboration questions moved four points.
Two slides earlier, the same review carried the pricing question for the enterprise segment. It has been carrying it since February. Nine teams now hold a piece of that question, each progressing the part it can safely progress, which is why the pre-read runs to sixty pages and nobody has read all of it. The question is where it was in February.
Both slides are accurate. The first is not evidence about the second.
Organisations routinely treat collaboration activity as evidence of collective thinking. It is not. A system can improve coordination, participation and information flow without changing what a group understands, decides or produces together. Two different results. One is measured to two decimal places. The other is not measured at all — and when it fails, the reflex is to buy more of the first.
What coordination actually solves
Start with what the software is for, because it does that job well.
Coordination is the problem of work people do apart: who holds what, who needs it, what is blocking whom, what happens next. Getting it wrong costs duplicated effort, late handoffs and decisions taken without the number that would have changed them. Thirty years of software have been pointed at that, and the difference is not marginal: information reaches whoever needs it without anyone having to ask.
The measurements arrived with the software and measure the same thing: volume, participation, latency, reach. Honest numbers about coordination. Traffic — how much of it, how fast, how evenly spread.
The trouble starts one step later. Coordination governs what a group does when it is apart. There is a second result, and it is only produced while the group is together: an understanding of the trade-off all of them will still be working from on Wednesday. Every function can hold the same pricing analysis and still leave with a different view of which risk the company is accepting. That understanding cannot be circulated. It is made out of the differences between them, and either made while they are in the conversation or not at all.
Nine teams and a sixty-page pre-read is one shape of that failure. Three hours of complete engagement that ends with everyone holding what they walked in with is another. The collaboration figures are healthy in both.
What has been measured, and on whom
The load has been measured, and carefully. Rob Cross, Reb Rebele and Adam Grant, drawing on two decades of data and network research across more than 300 organisations, found that time spent by managers and employees on collaborative activities had grown by half or more, and that 20 to 35 per cent of value-added collaboration comes from 3 to 5 per cent of employees. Notice what that measures: quantity, and its distribution — that collaboration is expensive and unevenly borne. Nothing about what any group produced.
Now ask what your organisation uses to judge whether a leadership conversation worked. Almost always, the opinion of the people in it — usually the person who called it. Steven Rogelberg, across multiple studies of employees in a range of industries, found that whoever leads a meeting rates it more favourably than those attending, and that the more someone talks the better they judge it. Take that verdict as your reading on whether the conversation worked, as most organisations do, and your reading is a self-report from the most invested and most talkative person in the room.
Group performance itself has been measured, in laboratories, on problems with one right answer. Ethan Bernstein, Jesse Shore and David Lazer gave three-person groups a routing puzzle and varied a single thing: how often each member could see what the others had come up with. Continuously, at intervals, or never.
Continuous visibility raised the group’s typical answer and lowered its best one. People copied one another onto something decent and stopped searching alone, so the optimal solution turned up less often. Groups that saw each other only at intervals had it both ways: the higher typical answer, and the optimum found as often as by people working alone.
Three people on a routing puzzle is not a leadership team choosing between product lines, and carrying it into your room is our inference, not their finding. What survives the distance is the shape of it: contact is not a dial you turn up. More of it bought a better average and cost the group the best thing it had.
Which is where this article’s title overstates it. All of that is real measurement. None of it touches the question your operating review is actually deciding: whether another forum around the pricing question improves what the group makes of it. Every proposal to add one assumes an answer nobody has.
Why the answer is always more
Follow the pricing question into its third quarter. Nobody concludes that the group did not think well together, because that is not one of the available explanations. What is available is a coordination fault, and those have obvious causes: the wrong people in the room, the analysis arriving late, functions not aligned beforehand. Each has a fix, and every fix is more collaboration. Add two attendees. Add a pre-alignment session. Stand up a working group.
That is not stupidity. It is the only remedy available for a problem the organisation has no measure of, and it moves the numbers it does have. The bill goes to the few people in everything.
Here is the part worth arguing with. A capital request of that size would not survive this review without a named result and someone accountable for it. A request for more collaboration is approved in the room where it is made.
The reflex runs hardest in the rhythms you can add to — the weekly exec sync, the project review — and costs most in the rooms you cannot: board preparation, the session after something has gone wrong. Akheon works on that second result, reading what a group is producing together beneath the conversation while it is still live. Whether that can be done well is a different argument.
Refuse to approve the reflex unpriced. Two questions, neither of them about the meeting:
- What does this group have to produce that none of them can produce apart? Not the minuted decision. The understanding of the trade-off each function will still be working from next month.
- Who will be able to tell you whether that happened, and when? If the only answer is asking the people who were in the room how it went, that is Rogelberg’s finding, not an observation.
If the second has no answer, you are buying traffic and calling it thinking. Traffic was already working.
Keep funding it on those terms and the question you have carried longest will still be open next February. The slide will be greener.
Sources
Rob Cross, Reb Rebele and Adam Grant, “Collaborative Overload,” Harvard Business Review 94(1), January–February 2016: 74–79. Two decades of data; network research across more than 300 organisations.
Steven G. Rogelberg, The Surprising Science of Meetings (Oxford University Press, 2019), and “Why Your Meetings Stink — and What to Do About It,” Harvard Business Review, January–February 2019. Survey and interview work with over 5,000 employees across a range of industries.
Ethan Bernstein, Jesse Shore and David Lazer, “How intermittent breaks in interaction improve collective intelligence,” PNAS 115(35), 2018: 8734–8739. Three-person groups, travelling-salesperson task, laboratory setting.