The roundabout: when to ask before you decide
A decision gets escalated when taking the wrong exit costs too much in euros or in time. Everything else gets decided and on you go.
It is worth asking before deciding when taking the wrong exit costs too much in euros, or when it costs too much time, which in the end is also euros. It is the only criterion I have seen work the same way in a management committee and in a boardroom. Everything else gets decided where the information sits, and on you go.
The metaphor is not mine. A friend explained it to me years ago, in a conversation that had nothing to do with management, and I have used it ever since because it is the only thing I have found that a team understands the first time and applies the next day.
You reach a roundabout and you can go straight, right or left. Which of the three is the good one is rarely knowable with the information on the table the day you have to turn, so the useful question comes afterwards: what does the wrong exit cost you.
And sooner or later you take the wrong one.
What decides whether to ask is the price of the wrong exit
Two costs, and the second one gets calculated far worse than the first. The first is money, and there is a number above which the decision stops belonging to whoever has it in front of them. The second is time: there are exits it takes six months to come back from, and that cost has no line in the P&L, does not show up in the budget and no auditor asks about it. It arrives whole the day you have already spent it.
Which is why a hundred-thousand euro decision fixed in a week is cheap, and a twenty-thousand euro one that eats two quarters of the committee's agenda turns out expensive.
The examples sit on both sides, and most of them are management
Raising list prices by five per cent gets tested and corrected within weeks, so it gets decided and that is that. Changing the payment gateway is reversible too, but the migration eats two months of the technical team even when it goes well, so there the cost sits in the calendar and is worth talking through first.
Signing a three-year exclusive with a distributor is another matter. The first-year amount can be small and the decision still mortgages the channel until 2029, and getting out costs a lawsuit or a settlement. Reorganising the sales team by territory reverses on the org chart and does not reverse with the people, which is where the cost actually sits.
The board cases are the same ones with another zero, which is why the threshold changes with every company: it depends on the cash, on the leverage and on how many bullets the team has left if this one misses. Copying it from the company next door is of little use.
The moment to set the number is before there is a specific decision on the table
Written with a case in front of you, the number gets argued with names attached and ends up being the one that suits whoever proposed it. Written beforehand, while nobody is defending anything yet, it lasts years. Without that number the criterion ends up being the mood of whoever is deciding that week, and the team notices that before anyone else.
What fails in practice is that almost nobody comes back to the roundabout
People take an exit and carry on. They decided, and carrying on is more comfortable than admitting three months later that the exit was the other one, particularly if the decision was defended in front of twelve people. By the time somebody says it out loud, coming back is no longer free and the conversation stops being about the business and becomes about who got it wrong.
I have seen the opposite failure more often. The boss who wants to be asked about everything ends up the bottleneck of his own company, because if every turn has to be approved what you have is a queue with him at the end of it. And it has happened to me ... more than once.
At school effort is what counts and in life results are what count. A team that asks about everything is working hard and standing still.
Frequently asked
When should a decision be escalated and when should you just decide?
Escalate when being wrong costs more than the euro threshold the company has set, or when reversing takes so long that the real cost is the calendar. Everything else gets decided where the information sits.
What decisions can a manager take without asking permission?
The reversible ones that are cheap to reverse: a price change tested and corrected within weeks, a campaign, a replaceable supplier. The ones that mortgage a channel or eat quarters of calendar get discussed first, even when the first-year amount is small.
How do you know whether a decision is reversible?
By asking what undoing it costs in money and in months, and who would have to do it. A reorganisation reverses on the org chart and does not reverse with the people, which is where the real cost sits.
How do you avoid becoming the bottleneck of your own team?
By writing down the threshold and writing down what never gets escalated. If every turn has to be approved, what you have is a queue with the boss at the end of it.
Who sets a company's euro threshold, and when?
Management or the board depending on the case, and the timing matters: before there is a specific decision on the table. It varies with the cash, the leverage and the team's margin for error.
Facts and sources
- Juan Santana's decision rule, which he calls "the roundabout": a decision gets escalated when taking the wrong exit costs too much in euros or too much time; everything else gets decided where the information sits. (Juan Santana, mykobu.com, April 2026)
- The criterion applies the same way in management and in the boardroom: the amount changes, the mechanics do not. (same source)
- Calendar cost has no line in the P&L and does not appear in the budget, so almost nobody calculates it before deciding. (same source)
- The euro threshold changes with every company according to cash, leverage and the team's margin for error, and is best set before there is a specific case on the table. (same source)
- The metaphor is not original to him: a friend explained it and he adopted it. (same source)
Who writes this
Juan Santana. Investor and independent director. Law at the University of Navarra and an MBA from IESE. Investment banking at Robertson Stephens and Lehman Brothers, then general management at Panda Security, Groupalia, Plenummedia and Publicar / Guru. Founding investor in seventeen companies since 2011, six sold and four closed.
The full profile, with the portfolio and the countries, is at Flight JS54.