Four kinds of knowledge. You keep all of them.
Most market work ends as a slide deck, and everything learned in making it is lost — so the next person starts from nothing. An engagement here ends as a structured record your company owns, in your own words, with every entry carrying where it came from and when it was read.
Learned once, and inherited by everything after it.
GeneralTrue of the chemistry wherever it sells
Who supplies it and at what scale, how it stands with regulators on each side of the Atlantic, what safety credentials it already carries, which forms the market has settled on.
Every application you look at afterwards inherits this, so it is paid for once and used for years.
ProductWhat you make, in your words
What your company actually sells today, described the way your business describes it — your segment names, your brands and divisions, your own vocabulary.
Without it, the markets you already serve come back to you as discoveries. With it, they are subtracted before anything is ranked.
MarketEach candidate, and the facts behind it
Every market considered, with the facts that support or sink it — each one carrying its basis, its unit and its source, so a number can be traced rather than believed.
Including the markets that failed, and why. A list nothing falls off is a list nobody checked.
TechnicalWhat the substance actually does
Properties as measured and published, never as claimed. Interfacial behaviour, stability, compatibility — and how it behaves biologically, because that is often what decides where it can sell at all.
Three things go down with every finding, or it does not go down at all.
Where it came from
Without exception. If we cannot point at the document a claim came from, we do not write the claim down — not as a weaker fact, not at all.
What it does not prove
Written next to the finding, not in a footnote at the back. What the evidence will not support is part of the finding itself, not a disclaimer bolted on afterwards.
Where it stands
Still current, replaced by something later, or waiting on an answer only you can give. When something is replaced, the old version is kept and linked to the new one, so you can see what changed and when.
The knowledge never feeds back into the searching.
This is the rule that keeps the record worth having. A conclusion a person has signed off is never used as evidence for the next search or the next score. If it were, the tool would start corroborating itself: its own earlier judgement would come back looking like independent support, and the second study would agree with the first for no reason other than that the first happened.
This is wired in, not just intended. An automatic check refuses to let the work go ahead if the searching side can reach the conclusions at all. Nobody can cross that line by accident, or by not knowing the rule was there.
Where chemicals could sell, and why nobody finds out.
Nothing about your work would appear here unless you put it in writing. Every study on this site is published with the client's written permission, and no study is published without it. That is why there are two and not twenty.