Evidence on a prospect
Every claim on a prospect record arrives attached to where it came from. A statement with nothing behind it does not reach the record at all — it is not softened, flagged or marked uncertain, it is simply not written. That single rule is what makes a prospect list still worth reading months later, and it is the reason the list is shorter than one built by a scraper.
The rule, stated once#
A claim and its backing travel together. When research reads that a business opened a second site, the record holds the sentence and the place it was read, not the sentence alone. When nothing supports a sentence, the sentence does not exist.
The alternative — writing the claim and marking it unverified — sounds kinder and is worse. Unverified claims are read as facts by whoever opens the record next, because a busy person reads the sentence and not the badge beside it. Refusing to write it is the only version of the rule that survives contact with a hurried Tuesday.
What counts as backing#
| Kind | Example of a claim it supports | Strength |
|---|---|---|
| The organisation's own published material | What it does; where it trades; what it announced | Attributed — it is what the business says about itself |
| A public listing or registration | Trading name; registered place; trade category | Strong for identity, weak for current activity |
| A public announcement carrying a date | A recent change worth mentioning in an opening | Strong, and it ages — see freshness |
| A page naming a person and a role | Who appears to hold a role | An inference, recorded as one, never asserted |
| Nothing | Anything | The claim is not written |
Notice that the strongest kinds are still attributed rather than absolute. A business describing itself as the leading supplier in its region is recorded as a thing the business says, with the page attached. Reading that line as a fact about the region is the reader's error, and the record is built so the reader can avoid it in one glance.
Where it shows up, and why you should look#
The backing sits with the claim on the prospect record, so nothing has to be hunted for on another screen. Three moments make it worth opening:
- Before approving a first message. The opening is grounded in one or two of these claims. If the claim is stale or attributed rather than factual, you want to know that before it is quoted back to a stranger.
- When a claim surprises you. A surprising sentence is either the most useful thing on the record or a misread source, and the source settles it in seconds.
- When somebody disagrees. *Where did that come from* is answerable, months later, without anybody reconstructing a run.
The same trail is what makes a fit judgement arguable rather than oracular. A score you cannot interrogate is a number somebody eventually stops believing; a score whose inputs are each traceable can be disagreed with specifically, which is how it gets better.
Ageing, contradiction and removal#
- A claim gets old
- Freshness governs re-research: a record is re-read when its age makes a claim doubtful rather than on a fixed calendar. See Keeping prospect research fresh.
- Two sources disagree
- Both are kept with what each says. Silently picking a winner would destroy the only information that tells a reader to be careful here.
- The source disappears
- The claim is marked as no longer supported. A vanished page does not retroactively make yesterday's research dishonest, but it does stop the claim being used to ground a new message.
- A person says a claim is wrong
- The correction is recorded against the record and outranks the research. A person's direct knowledge of a business is better than a public page, and Connect is not built to argue with it.
Questions#
Can I add my own knowledge to a prospect?
Yes, and it is treated as what it is: a person's statement, recorded with you as its origin. That is stronger than a public page for anything you actually know, and it is kept distinct from research so a later reader can tell the two apart.
Why does one prospect have five claims and another twenty?
Because businesses publish different amounts, and because research spend is routed to where it can change a decision. Twenty claims on a candidate you were never going to write to is money spent on certainty you did not need.
Does an unsupported claim get stored somewhere out of sight?
No. It is not written to the record at all — there is no hidden tier of lower-confidence statements waiting to be promoted. What you can see is what exists.