How not to fail a plant trial

Here is the uncomfortable truth about first trials: most that fail do not fail on the chemistry. They fail on everything around the chemistry. The material was late. The instructions assumed a PhD. The success metric drifted. The founder changed the formula halfway through to chase a better number.

A plant trial is an operations test as much as a science test, and the buyer is watching both. This is the field checklist we walk every founder through before an introduction.

Before the trial

Ship what you said, when you said it. A plant scheduled line time for you. Being a week late does not just delay the trial, it tells the buyer what working with you will be like at scale. On-time and as-specified is the first thing they are grading.

Write instructions a shift supervisor can follow. Not a chemist, a shift supervisor. Mixing ratios, cure times, temperatures, and handling in plain language, on one page. If your protocol needs you on the phone to execute, it is not ready.

Bring the safety data. A safety data sheet and clear handling guidance are table stakes. A plant cannot and will not put an unknown material on their line without it. Missing SDS is the fastest way to turn a yes into a "let us revisit next quarter."

Agree the success metric in writing. Before anything is poured or sprayed, both sides sign off on the number that defines success and how it will be measured. This is the single most protective thing you can do. It prevents the goalposts from moving after the result comes in.

During the trial

Do not change the formula mid-trial. It is tempting. The first panel looks off, and you want to tweak. Do not. A trial on a moving formula produces a result you cannot defend and the buyer cannot trust. Lock it, run it, learn.

Leave a control. Treated versus untreated is the entire point. A result with no control is an anecdote. Insist on the control even if the buyer is relaxed about it, because you are the one who will need to defend the number later.

Be present without hovering. Answer questions fast. Do not stand over the operator's shoulder implying they will get it wrong. The plant's team running your material successfully is itself part of the proof.

The goal is not a perfect result. It is a trustworthy one. A trustworthy no beats a suspicious yes every time.

After the trial

Read the result honestly. If it worked, confirm the pre-agreed follow-on and move fast. If it did not, resist the urge to relitigate. Ask for the specifics, thank the team, and treat the no-go as data.

Write it up. A one-page trial summary, control versus treated, method, and metric, is an asset you will use in every future conversation, with buyers and investors alike.

Protect the relationship. The plant engineer who ran your trial is now the most valuable contact you have in that company. Even a no-go done well leaves a door open for the next product.

The founder failure modes, ranked

In our experience the trials that go wrong cluster around a short list, and none of them are about the molecule:

1. 1Late or off-spec delivery.
2. 2Instructions that assume the founder is in the room.
3. 3No agreed success metric, so the result is disputed.
4. 4Formula changes mid-trial.
5. 5No control.
6. 6Chasing a better number instead of accepting a clean no.

Every one of these is preventable. That is the good news. The science is the hard part, and you have already done it. Do not lose the trial on the easy part.

This is exactly the work First Yes does before an introduction: pressure-testing your readiness, structuring the protocol, and making sure the first trial with a major buyer is one you can win, or lose cleanly.