How We Decide an Ingredient Has Failed

A founder’s view on evidence, rejection, and the discipline of saying “not yet.”

There is a peculiar habit in the supplement industry: once an ingredient becomes popular, it is very difficult to make it disappear.

A molecule acquires a reputation. A few promising studies are published. Influencers repeat the same conclusions. Manufacturers build products around it. Consumers begin to recognise the name.

Eventually, the ingredient develops something resembling commercial momentum.

At that point, science often becomes secondary.

I have always been uncomfortable with that process.

When I look at an ingredient for MindHeaven, I do not begin by asking whether people know it, whether it is trending, or whether a competitor is already selling it successfully. I begin with a much less exciting question:

What would have to be true for this ingredient to deserve a place in a formula?

And just as importantly:

What would make me reject it?

That second question is essential.

If you only define the conditions under which you will accept an idea, but never define the conditions under which you will abandon it, you are not really conducting an investigation. You are looking for confirmation.

Failure does not mean “no effect”

An ingredient can fail for many reasons.

Sometimes the evidence simply does not show a meaningful effect.

Sometimes an interesting biological mechanism exists, but the mechanism never translates into a measurable human outcome.

Sometimes the effect appears only in a population very different from the people for whom a product is being designed.

Sometimes the studies are too small.

Sometimes the results are inconsistent.

Sometimes the dose required in research is impractical for the formulation.

Sometimes the ingredient works, but not well enough to justify the complexity, cost, safety considerations or interactions it introduces.

And sometimes the scientific idea remains interesting, but the evidence is simply not mature enough yet.

That last distinction matters.

I do not consider every rejected ingredient a bad ingredient.

Sometimes the correct decision is simply: not yet.

I am particularly cautious with mechanisms

Mechanistic explanations are seductive.

A compound crosses the blood-brain barrier. It influences a neurotransmitter. It interacts with a receptor.

It changes a biomarker associated with cognition.

All of this can be scientifically interesting.

But I have learned to treat the word “therefore” with suspicion.

A substance may affect a biochemical pathway and still fail to produce a meaningful improvement in memory, attention or performance.

The human body is not a linear machine.

It is a system of compensations, feedback mechanisms, redundancies and interactions that we frequently understand only partially.

So when I see a beautiful mechanistic explanation, my next question is not: “Does this make sense?”

It is:

Has the predicted human outcome actually been demonstrated?

If the answer is no, the mechanism remains a hypothesis. Not a benefit.

Small studies are not useless

But they are easy to overinterpret.

A trial with twenty participants can be valuable. It can reveal a signal worth investigating.

What it cannot do is carry the weight of certainty that marketing often places upon it.

If a small experiment produces an impressive result, I want to know whether someone else reproduced it.

I want to know whether the effect appears in a larger population.

I want to know whether the endpoint was defined before the experiment began.

I want to know how many outcomes were measured.

I want to know whether the reported effect was the primary objective or one favourable result among twenty analyses.

Most importantly, I want to know what happened when another group tried to do the same thing.

Replication is not glamorous.

But it is one of the things I trust most.

An ingredient can also fail because the comparator was wrong

This problem appears frequently in combination studies.

Imagine that a study compares Ingredient A + caffeine against placebo, and the combination performs better.

That result tells us something about the combination. It does not necessarily tell us what Ingredient A contributed.

Perhaps caffeine produced most of the effect. Perhaps the combination genuinely performed differently.

We simply do not know unless the study includes the appropriate comparison.

For a claim such as “Ingredient A improves the effects of caffeine,” I want to see Ingredient A + caffeine versus caffeine alone.

This may sound obvious. In supplement marketing, it often is not.

Safety can defeat efficacy

An ingredient may work and still fail.

If the effect requires doses that introduce disproportionate adverse effects, I am not interested in calling the experiment a success.

If long-term safety is poorly understood, that uncertainty belongs in the decision.

If an ingredient interacts with common medication, has a narrow practical use case, or creates unnecessary formulation complexity, we have to evaluate that honestly.

The purpose of formulation is not to fit as many scientifically interesting molecules as possible into a bottle.

It is to make responsible choices.

Sometimes we reject something we wanted to use

These are the most important decisions.

It is easy to reject an ingredient you never liked.

The real test comes when the ingredient fits the concept perfectly.

Perhaps it would make the formulation easier to explain. Perhaps customers already recognise it. Perhaps competitors have made it almost mandatory in that product category. Perhaps I personally find the mechanism fascinating.

And then the evidence does not survive scrutiny.

That is precisely when the standard matters.

Because if I change the evidentiary threshold depending on how much I want a conclusion to be true, there is no standard at all.

What does failure mean at MindHeaven?

For me, an ingredient has failed when its evidence is no longer strong enough to justify the claim, dose, form or role we intended for it.

That conclusion may be: Reject. It may be: Hold. It may be: Use, but make a much narrower claim. And sometimes: Remove it from the formula entirely.

I do not regard any of these outcomes as wasted research.

Rejecting a popular ingredient for a good reason may be more valuable than discovering another ingredient to sell.

Every rejection sharpens the formulation. Every unanswered question defines what we still need to learn.

And every failed hypothesis protects us from building a product around something we merely wanted to believe.

That is why I do not ask whether an ingredient is fashionable.

I ask a simpler question:

If I did not sell this ingredient, would I interpret the evidence in exactly the same way?

If the answer is no, I know I need to look again.