Against the demo

The best company we ever backed showed us a spreadsheet and a burn scar. The spreadsheet was forty-one rows of downtime, by machine, by shift, by cause. The scar was on the founder’s left forearm, from a furnace door, and he mentioned it only because we asked.

There was no demo. There was no product to demo yet. We wrote the check in eight days.

Why the projector lies

A demo is a controlled environment, and a plant is not one. The demo runs on a laptop, on good wifi, with clean data and a friendly audience. The customer has a forty-year-old line, a dead zone behind the press, and a night shift that will find every edge case by Thursday.

Software that looks good on a projector has usually been designed for the projector. Software that works on a floor has usually been designed around a complaint.

The infrastructure that matters most tends to look worst on a screen.

What we look at instead

We read the spreadsheet. We ask who made it, and whether it is still open on someone’s second monitor. We ask for a customer who will tell us what is wrong with the product, and we call that customer before we call any other reference.

We also visit. Both partners, usually, because we would like to see the machine. Hana once spent a morning watching a technician work around a feature the founder had been proud of. The founder rebuilt it in three weeks. That company is Agarwork, and the colony counter on its bench has no demo mode at all.

The question we ask instead

We ask what happened last Thursday. Not last quarter, not on average. A single day, with a time stamp. The founder who answers with a story about a specific machine, a specific technician and a specific delay has told us more than any slide could.

Then we ask what they would do if the customer called tomorrow and said it was not working. The best answers are boring. They have a name, a phone number and a drive time. Those are the companies that last.

For founders

Do not hide the rough parts. Show us the log from last Thursday, the error nobody has fixed, and the workaround the customer invented. We trust a team that can describe its worst day more than one that has only a best day. A founder who says everything works has not yet been to the plant.

The plant will teach you more in an hour than a pitch meeting will in a quarter, and we would rather hear what it taught you. If you must make a deck, make it short and make the first slide a number. If you can bring the burn scar, even better. We will not ask where you got it.