Innovative Commercial Strategies Innovative Commercial Strategies CRE Advisory + AI Systems

November 19, 2026 · Daniel Lawton

What high confidence means in a research brief, and why a good one tells you what it does not know

Two briefs can report the same sentence: “The seller was a defendant in a 2019 fraud suit.” In one brief that sentence is a verified fact you can take to a closing table. In the other it is a name match on a common surname that nobody checked. On the page they look identical.

The difference is the confidence level, and a brief that does not state one is asking you to guess. Worse, a brief that sounds equally sure about everything is telling you its author did not separate what the record shows from what the author assumed.

This is what the three levels mean in an ICS brief, how a finding earns one, and why the section listing what we do not know is often the one that saves the client money.

Three words with specific meanings

Every finding in a brief carries one of three confidence levels. They are not moods.

LevelWhat it rests onWhat you can do with it
HighPrimary documents, identity confirmed by more than a name, no credible contrary evidenceAct on it. Put it in front of your attorney.
ModerateGood sources with a gap: identity probable but not proven, or one solid source uncorroboratedAsk a direct question. Cover it in the contract.
LowA single weak source, a bare name match, or an inference from patternTreat it as a lead. Do not repeat it.

Two things the scale is not. It is not a measure of how serious the finding is: a low confidence finding can be about something grave, and a high confidence finding can be trivial. And high does not mean certain. Records contain errors, clerks mis-index, and people share names. High means the evidence is strong enough to act on and specific enough to check.

The source grade describes the input, not the conclusion

Before a finding gets a confidence level, each piece of information under it gets a grade. ICS uses the standard two-part scale: a letter from A to F for how reliable the source has proven, and a number from 1 to 6 for how credible this particular item is. The glossary entry on source grading has the full table.

A certified court docket is about as good as a source gets. An anonymous online review is near the other end. A local news story sits in between, and its grade depends on whether the reporter cites documents.

The trap is assuming a strong source means a strong conclusion. It does not. A real docket for a real lawsuit against a “Michael Sullivan” is an excellent source for the claim that someone named Michael Sullivan was sued. It is no source at all for the claim that your Michael Sullivan was sued, until something ties the two together. Grades go on the inputs. Confidence goes on the conclusion. A brief that blurs them will overstate.

A worked example: one lawsuit, three confidence levels

A composite. A client is buying a building in Bridgewater from an individual seller with a common name. A Superior Court search returns a 2019 civil suit alleging fraud against a defendant with the same first and last name. Asking price: $1.2 million.

As first found: low confidence. Same name, same county. Plymouth County has dozens of adults with that name. At this stage the correct sentence in the brief is “a person with the seller’s name was sued in 2019; identity not established.”

After pulling the complaint: moderate. The complaint names the defendant as manager of an LLC. The Secretary of the Commonwealth’s record for that LLC lists a manager with the seller’s name and a business address in Bridgewater. Two independent records now point the same way. Probable, not proven: the middle initial differs between the two.

After pulling the deed: high. The LLC named in the suit is the grantor on a 2016 deed, signed by its manager, conveying a different parcel to the seller individually at an address that matches the seller’s current one. Three primary documents, linked by entity, address, and signature. The middle initial discrepancy is noted and explained as a clerical variation that appears in one document only.

Then the part that matters to the client: the docket shows the case was dismissed with prejudice by stipulation in 2021. The finding in the brief reads: “High confidence the seller was the defendant in a 2019 fraud suit. High confidence the case ended in a stipulated dismissal. Terms of any settlement: unknown, not public.”

That took four documents and under two hours. The low confidence version, repeated at a negotiating table, could have cost the client the deal or invited a defamation claim. This is what entity resolution is for.

What moves a finding up, and what moves it down

Confidence goes up with:

  • Primary documents over accounts of them. The docket, not the article about the docket.
  • Independent corroboration. Two sources that do not rely on each other. Five websites repeating one press release are one source. See corroboration.
  • Identity beyond the name. An address, an entity, a signature, a middle name, a date that lines up.
  • Recency. A status checked this week.

Confidence goes down with:

  • A single source, however good.
  • Evidence that cuts the other way and has not been explained.
  • Inference from pattern. “Entities like this usually” is a reason to look, not a finding.
  • Age. A lien found in a 2022 search may have been released.

A good researcher also goes looking for the evidence that would prove the finding wrong, and reports whether any turned up. If nobody tried to break the conclusion, the confidence level is a guess.

The section that says what we do not know

Every ICS brief has a gaps section. A sample from a composite file on a contractor:

  1. State court search covered Plymouth, Norfolk, and Bristol counties, 2005 to present. Earlier cases and other counties were not searched.
  2. The subject operated in New Hampshire from 2012 to 2016. No New Hampshire records were searched under this scope.
  3. One federal case is sealed. Its existence is known, its subject matter is not.
  4. Ownership of the LLC is not in the public record. The manager is known, the members are not.
  5. Two former addresses could not be confirmed.

Read that list as a menu. Items 1 and 2 can be closed with a few more hours of searching, roughly $400 to $900 of additional work in a typical scope. Item 3 cannot be closed by anyone lawfully. Item 4 can be closed with a question: ask the counterparty to represent in writing who the members are.

Without the gaps section, “no litigation found” reads as “no litigation.” With it, the client knows exactly which risk is still open and can decide whether it is worth four hours to close. That is the decision the brief exists to support.

How to tell a good brief from a confident one

Three checks you can run on any research product, ours included.

  • Does anything rate below high? If every finding is certain, nobody was grading.
  • Does everything rate low? A brief hedged on every line protects the author and tells you nothing.
  • Can you check it? Each finding should cite a document you could pull yourself: court, docket number, book and page, filing date.

A fourth check is about the gaps themselves. Some gaps stay gaps on purpose. ICS will not raise a confidence level by pretexting, meaning calling a bank, an employer, or a relative under a false identity. We do not surveil private individuals. We do not obtain medical, credit, or private financial data. And no ICS brief is produced for, or may be used for, employment, credit, or residential tenant decisions. When the lawful public record runs out, the brief says “unknown” and stops. What to do about an unknown that matters is a conversation to have with your attorney.

What to do this week

  1. Take the last report anyone gave you about a counterparty and mark each claim high, moderate, or low yourself.
  2. For each claim, write down the document it rests on. If there is none, it is low.
  3. For every claim about a person, ask what ties it to that person besides the name.
  4. Write the list of what was not searched: which counties, which years, which states.
  5. Decide which open gap is worth paying to close before the deal moves.
  6. If the deal is a building, put the confirmed findings into a free Buyer Brief and leave the low confidence ones out.

Common questions

What does high confidence mean in a due diligence brief?
It means the finding rests on primary documents, such as a court docket or a recorded deed, that the identity match is established by more than a name, and that no credible evidence points the other way. It does not mean certainty. It means you can act on the finding and your attorney can check it.
Is a source grade the same thing as a confidence level?
No. A source grade rates an input: how reliable the source has been and how credible this particular item is. A confidence level rates a conclusion drawn from one or more inputs. A reliable document can still support only a low confidence conclusion if it is unclear that it concerns the right person.
Why would a research brief list what it did not find?
Because an absence only means something if you know where the researcher looked. No lawsuits found in three Massachusetts counties is a different statement from no lawsuits anywhere. A gaps section tells you which risks were tested and which remain open, so you can decide whether to pay to close them.
Should I act on a low confidence finding?
Treat it as a lead, not a fact. A low confidence finding is worth a direct question to the other side or a targeted follow-up search. It should not be repeated to third parties, used as leverage, or published. If it matters to the decision, the right move is to pay to raise or kill it.

Keep reading

Want this applied to your business?

Book a free strategy call — we'll turn the ideas in this article into a concrete plan with numbers attached.