How long should pre-close diligence take?

The honest answer to the question in the title is that nobody credibly knows, and you should be suspicious of anyone who tells you otherwise.

That is worth saying plainly because the alternative is common. Diligence software is sold with confident cycle-time claims, and the sample behind them is usually a handful of customers who were unusually disorganised beforehand. A figure gathered that way tells you about the sample, not about you.

Why no benchmark exists

Three reasons, and they are structural rather than a gap someone will fill next year.

The deals are not comparable. A £5m refinance with a borrower you have lent to twice, a first-time unitranche into a group with three foreign subsidiaries, and a grant application from a charity with two part-time staff are all "diligence". Averaging them produces a number that describes none of them.

Nobody instruments the internal half. Even firms that track how long a counterparty takes to respond usually cannot say how long a document sat with Legal after it arrived, because nothing in the process records it. A benchmark built on the visible half is measuring the wrong thing by roughly half.

The start line is ambiguous. Does the clock start at the first conversation, at the mandate letter, at the day someone opened a checklist? Firms answer this differently, which means even two honest numbers are not comparable.

Measure your own instead

The measure worth holding is days from opening a deal to IC-ready, and its value comes almost entirely from being measured the same way twice.

Do it once before you change anything. That baseline is the only thing your later number can be compared against, and the temptation to skip it is exactly why so many process improvements are unprovable a year later.

To reconstruct a baseline from deals you have already closed, take one deal and its email thread and write down, for each requirement, two dates: when it arrived and when someone accepted it. It is tedious for one deal and impossible for twenty, which is the real reason nobody does it — not indifference, but that the data exists only as a reconstruction.

What firms tend to find

Three patterns come up often enough to be worth expecting, though they are observations rather than research and should be held loosely.

It is a scatter, not a stall. The delay is rarely one dramatic block. It is twenty small gaps of three or four days each, none of which anyone would have called a problem individually, plus one genuine week-long wait sitting behind a dependency nobody had named out loud.

Roughly half of it is internal. Documents that arrived and waited. This is the half nobody has numbers for, and it is also the half you control.

The bottleneck is not where the team guessed. Legal takes the blame for a lot of delay that turns out to be a document going back and forth because the first version was unsigned, or an item that was never actually requested because it lived on the checklist and not in any email.

The one number to avoid

Resist "percentage of items complete". It rises steadily, it feels like progress, and it tells you almost nothing: the last three items on a deal are routinely worth more elapsed time than the first thirty, because they are the ones that were hard.

A completion percentage that reads 90% for eleven days is a number that has stopped measuring anything. Days open is uncomfortable in exactly the way a useful metric should be.

What a target looks like

Set one only after two honest measurements, and set it as a distribution rather than a mean — "no deal above X weeks" is a more useful commitment than an average, because the deals that hurt are the tail.

And expect the number to be argued about at renewal. That is a feature: a metric nobody argues about is one nobody is using.