A verified contact without a date is an opinion

Two rows in the same column both say deliverable. One was probed this morning; the other was copied from a list bought in 2023. To every query, every report and every dashboard they are identical. They are not the same fact, and the difference between them is a single field that most lists do not store.

What a verification result actually claims

An SMTP probe is a conversation with a mail server at a moment in time. The server answers about the mailbox as it exists right then: it accepts mail for this address, refuses it, or declines to give a verdict at all. Nothing in that exchange is permanent. The moment the answer is recorded, it starts becoming history - which is why the useful unit is not "verified" but "verified at".

A static resold database can hand you a row and a purchase date. It cannot hand you a probe date, because the probe is not what produced the row - the resale is.

Why dates decide what you can do with the list

• Sending order. Working the freshest contacts first is a decision you can only make if freshness is stored. Without the field, the only honest option is to re-probe everything, which is the same work you thought you had already done.
• Re-probe scheduling. A verdict is worth what it costs to refresh it. With a date on the row, a pipeline can re-check what is old and leave alone what is current instead of redoing the whole file on a fixed calendar.
• An honest accuracy number. "99% accurate" means nothing without saying accurate when, over what sample, and how the answer was obtained. A timestamp is what turns a claim into a measurement.
• Attribution when things go wrong. If a bounce arrives, a row with a probe date tells you whether the verdict was stale or whether the address died inside a week. Rows without dates make every post-mortem a guess.

What decays under the dates

Contact data is not a slower-moving asset than it looks. Three published figures set the clock:

• 23% of contacts change jobs annually (ZoomInfo, 2025). Every job change can retire the address you hold.
• 40% of email addresses are dead within two years (NeverBounce). Time alone is enough to kill almost half a file.
• 42% of B2B data goes stale within a year (Harvard Business Review, 2024). The people are still there; the records about them are not.

A list sold as a snapshot carries the age of the snapshot on every row with no way to separate the fresh part from the tired part. Checked contacts carry a published 2-5% bounce rate against 10-35% for static, resold databases - and the difference is mostly explained by when the answer was taken, not by who took it.

How to store a verdict

1. Keep the state, the time and the source on the same row. A state without provenance cannot be audited and cannot be re-ranked.
2. Never rewrite a verdict in place without changing its date. Overwriting makes the row look current while the underlying answer stays old.
3. Treat unknown as a first-class state. Catch-all domains and unanswered probes are not failures; they are rows that need a decision at send time.
4. Probe at the point of send. The only verdict that cannot be stale is the one taken while the message is going out.

What TAPAC returns

TAPAC probes the mailbox inside the search that found the contact and returns the state, the time of the probe and the source page or profile the address came from - so every row in the result is a dated statement rather than a badge. Pay-per-use at $0.10-0.50 per contact, from 100 free searches; no stored snapshot to inherit.

npx -y @tapacapi/mcp        # tools: tapac_find_contacts, tapac_status
# or hosted: https://tapacapi.com/mcp
# or REST:   POST https://tapacapi.com/v1/contacts/search

Related: what an accuracy claim is measured against, how often a list needs rechecking, and why delivered is not delivery.