Ask for the conversion rate of a lead pipeline and you get one number. Ask where the contacts go and most teams point at verification, because that is the only stage with a familiar benchmark. It is usually not the biggest leak - it is just the only one with a number attached.
Seven stages, and only three of them have published rates
| Stage | What is lost here | Published rate? |
|---|---|---|
| 1. Discovery | Coverage - people the source cannot see at all | None. Depends on the source |
| 2. Extraction | Records with no usable address, or a role mailbox instead of a person | None. Count the empty fields in your own output |
| 3. Verification | Undeliverable addresses, and catch-alls that must be unknown
|
Yes: 2-5% bounce on checked contacts vs 10-35% on static resold databases |
| 4. Storage | Decay while the list sits - 23% change jobs a year (ZoomInfo, 2025), 40% of mailboxes dead within two years (NeverBounce) | Yes, both; together roughly 4% of a stored list a month |
| 5. Send | Bounces among addresses that were called deliverable | Not published. It is your own delivery report |
| 6. Inbox | Spam placement, blocks, throttling past the gateway | None universal |
| 7. Read and reply | Attention | Partial: 3.2x open rate with personalisation (HubSpot, 2025); community-sourced leads convert 30% better (McKinsey, 2025) |
The blank cells are blank on purpose. Those rates depend on your list, your domain and your offer, so a vendor number for them would be a number about somebody else's pipeline. That is also why stage 5 is the only honest test of stage 3: bounce rate on messages actually sent to addresses a provider called deliverable.
The stages multiply
End-to-end yield is the product of the stages, not the sum. Four stages passing 90% each leave about two thirds of the input. Improving the worst stage pays more than nudging every good one, which is an argument for finding the worst stage before optimising anything.
Where the money actually leaks
Sends are billed at the same rate whether they land or bounce, and a bounce also spends the sending reputation that stage 6 depends on. So the check belongs at stage 3, not at stage 5:
- verification costs $0.10-0.50 per contact with TAPAC, pay-per-use, 100 free searches
- the published gap is 2-5% bounce on checked contacts against 10-35% for static resold databases
- freshness compounds on top of that: fresh data outperforms stale databases by 42% (Harvard Business Review, 2024)
Which is stage 4 leaking into stages 5, 6 and 7 - the same contacts removed from a list look identical to a bad subject line until you measure them separately.
Instrument the stages with no benchmark
Concrete tests, in order of cost:
- Extraction: count records with an empty address field. That number needs no external benchmark.
- Send vs handoff: put a seed mailbox you control inside the send. If the seed never arrives, the problem happened before the recipient's server was involved.
- Bounce rate: read it from your own delivery reports among addresses reported deliverable. That is your real false-positive rate.
- Inbox vs content: split one send into a variant with the same list and a different subject. If both flatline, it is placement; if one moves, it was content.
A list that verifies well and still gets no replies is not a list problem, and re-checking it again will not show you that.
What it looks like implemented
TAPAC runs the mailbox check inside the search that found the contact, so stages 1 to 3 arrive in one call - deliverable, undeliverable, or unknown for catch-all domains - with the time of the call and the source page or profile attached, so stages 4 and 5 can be measured against your own cadence instead of assumed.
npx -y @tapacapi/mcp # tools: tapac_find_contacts, tapac_status
# hosted: https://tapacapi.com/mcp
# REST: POST https://tapacapi.com/v1/contacts/search
Top comments (0)