Should I sign a three-year lease?

Cerno is a private AI boardroom that runs consequential decisions through a structured process and returns a brief like the one below. This example is illustrative; your own would reflect the actual terms, your margins and how long you've been trading.

How this brief was reached

The advisors formed independent positions, a Devil's Advocate challenged the leading one, and the vote landed 4–1 for the conditional lease — with the CFO's dissent kept on the record, because the CFO is right about the arithmetic.

What the majority weighed more heavily than the arithmetic was what kind of door this is. A three-year lease with no break is one of the few genuinely irreversible commitments a small business makes: the rent is owed whether trade holds up or not, and assigning or subletting a lease you no longer want is slow and rarely recovers the full cost. A break clause at eighteen months doesn't make the lease cheap to reverse; it makes it possible to reverse, once, at a known point. That single clause moves the decision from one category to the other, which is why the recommendation makes it a condition rather than a nice-to-have.

The CFO's dissent was not adopted, but it set the model in the actions. The lease is cheaper if the business stays steady. The question is what happens if it doesn't, and "model it against a quarter less turnover" is the honest version of that question.

Confidence is moderate. The board is fairly sure about the shape of the answer — reversibility is worth paying for here — and less sure about the assumption doing the most work: that a business two years old, which has only ever been in flexible space, knows what it will need in three.

What the board weighed

Two years is not a long record. The CFO's case rests on "steady for two years". That's true and it's not much. The board reached for the base rate: not a published statistic, but the local, countable one — businesses like this one that took fixed premises nearby in recent years. The number that are still in that space, doing what they set out to do, is the honest prior for this decision, and the action is simply to go and count.

The premium is a price for information. Twenty percent more, for a year, on the rolling space, buys a year of evidence about what the business actually needs — how much space, in what location, with what growth. Whether that's expensive depends entirely on how likely the evidence is to change the answer. For a business that has never held fixed premises, the board judged that likelihood high enough to be worth the money.

Location is a bet on where customers will be, not where they are. The lease assumes the location is right for three years. If the customer base is shifting — more remote, more of a different kind of client, a different part of town — the lease locks in yesterday's answer. This was the assumption the board was least able to test from the brief, and it's flagged as such.

"We'll sublet if we have to" is not a plan. It came up. The board's view: it's the reversal everybody assumes and almost nobody achieves at the price they hoped. If the break clause isn't available, the rolling space is the reversal, and it costs a known amount.

If your situation differs

If you've been trading five years or more in premises of this kind, the biggest assumption is answered by your own record and the CFO's dissent is probably the majority. Take the cheaper lease — still with the break if you can get it.

If the space is a specific asset — a workshop with equipment installed, a kitchen with extraction, a fit-out you'd be paying for — the lease and the fit-out are one decision, and the reversal cost is the fit-out, not the rent. Treat it as more one-way, not less.

If the landlord offers a break at twelve months instead of eighteen, take it. The number is less important than the existence of a point at which you can stop.

The frameworks behind it

This decision turns on reversibility — the break clause is the smaller door — and on reference-class forecasting: the countable base rate of nearby businesses that did the same thing is the check on a two-year record that feels like more than it is.