Every organiser starts with a spreadsheet, and rightly so. With twenty companies, someone who knows the sector builds a better schedule than any system: they know that this buyer is already a client of that supplier, and that those two have not spoken since 2019.

The problem is not the spreadsheet. It is the point at which it stops working, which usually arrives without warning.

Where the manual approach stops working

The number of combinations grows far too fast

With 20 companies there are 190 possible pairs. With 60, there are 1,770. With 100, almost 5,000. No human evaluates five thousand possibilities, so what happens in practice is something else: pair the obvious ones, and fill the rest alphabetically or by whatever slot is left.

A last-minute change becomes total rework

Three cancellations the day before do not affect three meetings: they affect the entire chain of everyone who would have talked to them. Redoing that by hand, with the event the next day, is the classic source of a sleepless night.

Nobody can prove the criteria

When an exhibitor complains about the schedule, the answer has to be better than "that is how it worked out". Explicit criteria are what turn a complaint into a conversation, and what allow adjustments for the following year.

What you gain by automating

Scale without losing the criteria. The rule applied to the tenth company is the same one applied to the hundredth, which rarely happens once fatigue enters the equation.

Redoing is cheap. A last-minute cancellation stops being a crisis and becomes a new run of the pairing.

Distribution. A system can balance everyone's schedule at the same time, which is precisely what a person cannot do in their head.

What you lose, and how to compensate

You lose contextual knowledge: the system does not know that two companies already know each other, or that one of them had a problem with the other. That is compensated with two simple things: letting a company mark who it does not want to meet, and keeping human curation as the last layer, adjusting a few pairs instead of building them all.

The mistake is treating the choice as either/or. The design that works is automated at the base and manual at the exception.

How to decide

Two questions settle it: how many companies take part, and how often the event repeats. Below thirty companies, at an annual event, the spreadsheet delivers. Above that, or at a recurring event, the cost of doing it manually stops being the time to build it and becomes the poor schedule that half the participants receive.

The page on B2B matchmaking explains the calculation behind the pairing, and the guide on how a business round works shows the format from the participant's side.