# Evidence for a contact address

An address on a prospect record carries its origin the way every other claim does — where it was published, and when it was read. That provenance is what separates a usable address from a plausible string, and it is the thing to check before approving a first message, because an address belonging to a third party is a real address that is still the wrong one.

- **Status:** Available
- **Audience:** both
- **In the app:** #/prospects
- **Last verified:** 2026-09-10
- **Canonical:** https://connectbyjbrh.com/docs/prospects/contact-evidence/

## Three tests an address has to pass

1. **Found.** It appeared in public material, rather than being derived from a name or a domain.
2. **Attributed.** The place it appeared is stored with it, so a reader can see the origin without re-running research.
3. **Belongs to this organisation.** Not to a directory operator, a web agency, a registrar, or a marketplace that lists the business.

The third test is the one that quietly removes the most addresses. A published contact on a listing page is often the listing company's, and the business you researched never sees mail sent there. It is a genuinely sourced address, and it is not a way to reach the prospect — which is why provenance is stored rather than just the string.

## What the record holds

| Kept | Used for |
|---|---|
| The address itself | Sending, and matching against existing identities |
| Where it was published | Judging whether it belongs to the business |
| When it was read | Freshness — an address read a year ago is a weaker claim |
| What kind of address it is | A named person's, a role address, or a general enquiry route |
| Whether a person supplied it | Distinguishing your own knowledge from research |

An address you supply yourself is stored the same way, with you as its origin. That is deliberate symmetry: a person's knowledge is a stronger source than a public page, and the record should say which one it had.

## Matching against people you already know

Before an address is treated as a cold contact it is checked against the identities already in the workspace. An **Identity** is one address on one channel that resolves to a **Person**, and a person may hold many. If the address resolves, this is not a prospect — it is somebody you already know, and the cold path ends there.

That check is cheap and the failure it prevents is expensive. Writing a cold introduction to an existing customer reads as evidence that nobody in your business is paying attention, and no amount of good research afterwards repairs it. [Resolving a prospect to an existing relationship](/docs/prospects/prospect-identity/) covers what happens on both sides of the match.

> **Careful** A match on the organisation without a match on the address is still a match worth reading. A new person at a company you already serve is a relationship task with a warm route, not a cold prospect who happens to share a domain.

## When provenance goes bad

**The source page is gone** — The address stays with its history, marked as no longer supported. It is not silently deleted, because a message may already have been sent to it.
**Two sources give different addresses** — Both are kept. Choosing one silently would hide the fact that the business presents itself two ways, which is exactly the sort of thing a person should see before writing.
**The address bounces** — That is recorded against the address rather than against the business. The prospect survives; the route does not, and it becomes an uncontactable prospect with a next action.
**The address turns out to be a third party's** — Correcting it removes the route and records why. This is the failure the third test exists to catch, and catching it late is still catching it.

None of these produce a replacement address. A bounced or wrong contact does not trigger a search for a likely alternative, because there is no mechanism to produce one. The prospect moves to a different channel or waits for re-research.

## Questions

### Is a role address worse than a named person's?

Different rather than worse. A published enquiry address is a legitimate route and often the one a business actually reads. A named person's address is better for a specific ask and worse if that person has left — which is why the record stores which kind it holds.

### Why does the record show when an address was read?

Because addresses decay. A contact read this month is a stronger claim than the same string read a year ago, and freshness uses that age to decide what is worth re-reading before anything is sent.

### Can I see the source without leaving the prospect?

Yes — origin sits with the address on the record, the same way every other claim carries its backing. Checking it is a glance rather than an investigation, which is the only reason anybody does it before approving a message.

## Related

- [Connect does not guess email addresses](https://connectbyjbrh.com/docs/prospects/no-guessed-email/)
- [Contactability](https://connectbyjbrh.com/docs/prospects/contactability/)
- [Resolving a prospect to an existing relationship](https://connectbyjbrh.com/docs/prospects/prospect-identity/)
- [Evidence on a prospect](https://connectbyjbrh.com/docs/prospects/evidence/)
- [Finding a contact address, end to end](https://connectbyjbrh.com/docs/workflows/contact-discovery/)
- [Relationships in Connect](https://connectbyjbrh.com/docs/relationships/)

## What this page is based on

- docs-source/sources/CHANNELS.md §4 — contact discovery and evidence capture
- docs-source/sources/CHANNELS.md §5 — Person, Identity and value_key
- `docs-source/facts.py` — CANONICAL_TERMS Identity; CAPABILITY_STATUS cross_channel_identity