How GrowthPal builds
a sending list.
Who fits, who is reachable, and who is deliberately left out. The verification step, the bounce threshold it exists to protect, and the market size below which none of it works.
Updated
A list is built in two passes. The first decides who fits, from the client's closed-won data rather than their stated ideal customer profile. The second decides who is reachable, by verifying every address with Reoon and BounceBan before a single message is sent. Average bounce rate across campaigns stays under 2%. Below roughly 30,000 reachable contacts, the approach does not apply.
Pass one: who fits
The starting point is the client's closed-won data, not the ideal customer profile they describe in the kickoff call. Those two disagree more often than not. A company that believes it has one buyer type usually has several, and only some of them are worth chasing. The 412-campaign study found that targeting the profile the closed-won data showed, rather than the stated one, moved reply rate by a further 1.8 times on top of offer specificity.
Pass one therefore produces a definition, not a list: the firmographic band, the buying role, and the trigger that makes now the right week to write. The industry pages each carry their own version of that definition, because the buyer and the trigger differ by market. Those are worked examples of this step rather than restatements of it.
Pass two: who is reachable
A definition becomes a list only once every address on it has been checked. Two tools do different jobs here, and the order matters.
| Tool | What it does | What it claims |
|---|---|---|
| Reoon | Bulk verification. Its documentation says it asks "the recipient mail server whether the exact mailbox exists, without sending an email", and matches each address against disposable, role-account, free-provider and known spam-trap databases. | More than 99% accuracy. This is the vendor's own figure and the page cites no independent test, so it is quoted here as a claim rather than a measurement. |
| BounceBan | Catch-all resolution. Its site states that it does not "simply mark all catch-alls as risky like others do", and that it identifies which catch-all addresses are deliverable without sending to them. | Reliable verification for 85 to 95% of catch-all, greylisted and gateway-protected addresses. Again a vendor claim with no published methodology. |
The catch-all problem is the reason the second tool exists. A catch-all domain accepts mail for every address at it, real or not, so an ordinary verifier cannot tell a working mailbox from a typo. On most B2B lists catch-all domains are a substantial share of the better-known companies. Discarding them removes exactly the accounts a client most wants to reach. Sending to them blind is how a campaign arrives at a double-digit bounce rate in week one.
Both figures above are vendor marketing. The number that is actually measured is GrowthPal's own: bounce rate under 2% on average across campaigns, counted against addresses attempted.
Why the bounce threshold matters more than it looks
A bounce rate is not only a wasted send. It is the earliest signal a mailbox provider uses to decide what a sender is. Google's sender guidance tells bulk senders that "if messages start bouncing or start being deferred, reduce the sending volume until the SMTP error rate decreases", which is to say that the provider has already started throttling by the time a human notices. Yahoo asks senders to track hard bounces, soft bounces and inactive recipients separately, and to remove invalid recipients promptly.
The connected constraint is the spam rate. Google requires bulk senders to keep "spam rates reported in Postmaster Tools below 0.30%" and recommends staying "below 0.10%". Yahoo sets the same 0.3% requirement and notes it is calculated against mail delivered to the inbox rather than against everything sent, which makes the real budget tighter than it first appears. A list that has not been verified burns that budget on people who were never going to read the message.
Market size: why 30,000
Testing 10 to 15 message angles requires enough contacts that each angle reaches a sample big enough to separate a real difference from noise. Below roughly 30,000 reachable contacts there are not enough people to go round, and the exercise becomes guesswork wearing the clothes of a test. At that size precision account-based work beats volume, and GrowthPal says so rather than taking the engagement.
Reachable is doing work in that sentence. It means contacts who pass pass two, not companies in a market-sizing deck. A market of 30,000 companies is not a market of 30,000 reachable contacts.
Who is deliberately left out
- Role accounts. info@, sales@, support@ and the rest. They reach a shared mailbox rather than a person, and they correlate with complaint rates.
- Catch-all addresses that BounceBan cannot resolve. Unresolved means unknown, and unknown goes in the discard pile rather than the send.
- Known spam traps and disposable domains, which Reoon screens for.
- Anyone who has unsubscribed, across every client campaign, permanently.
- Markets below the size threshold above, which is a reason to decline the work rather than a filter on the list.
The legal position
Cold B2B email is lawful in the United States and the United Kingdom under different rules, and the difference shapes the list rather than the copy. The FTC's CAN-SPAM guidance sets requirements that apply to commercial messages regardless of consent, including an accurate From line and a working opt-out. The ICO's position is that business-to-business marketing to corporate subscribers is treated differently from marketing to individuals, and that sole traders and some partnerships are treated as individuals rather than as businesses. That last point is a list-building constraint, not a compliance footnote: it decides which company types can be on the list at all.
What is not documented
| Gap | Status |
|---|---|
| The share of a typical list that is catch-all | Not measured and published, so no figure is given. It varies by market. |
| Bounce rate by market or by list source | Only the overall average, under 2%, is published. |
| How many addresses a typical verification pass discards | Not published. |
Frequently asked questions about list building
Does GrowthPal buy lists?
No. A list is built to a definition derived from the client's closed-won data, then verified address by address. A bought list fails on both counts: it was not built to the client's buyer definition, and its addresses have not been checked against the mail servers that will receive them.
Why use two verification tools instead of one?
They solve different problems. A bulk verifier can tell a working mailbox from a dead one on an ordinary domain, but it cannot tell either from a catch-all domain, which accepts everything. BounceBan exists to resolve that second case, which is where many of the best-known companies on a B2B list sit.
What bounce rate should a cold email campaign expect?
GrowthPal's own campaigns average under 2% against addresses attempted. Neither Google nor Yahoo publishes a numeric bounce threshold; Google's instruction is to reduce sending volume when the SMTP error rate rises, which means the provider reacts before any dashboard shows a problem.
Is a smaller market really a reason to be turned down?
Yes. Below roughly 30,000 reachable contacts there are not enough people to test 10 to 15 angles against, so the central method does not work. Taking the engagement anyway would mean charging for a test that cannot reach a conclusion.
Sources for this page
- Reoon Email Verifier (vendor documentation). Accessed 2026-10-24.
- BounceBan (vendor documentation). Accessed 2026-10-24.
- Google, Email sender guidelines (primary source). Accessed 2026-10-24.
- Yahoo, Sender Best Practices (primary source). Accessed 2026-10-24.
- FTC, CAN-SPAM Act: A Compliance Guide for Business (primary source). Accessed 2026-10-24.
- ICO, Business-to-business marketing (primary source). Accessed 2026-10-24.
Last reviewed 2026-10-24. A plain-text version of this page is at /research/methodology/how-lists-are-built.md, and the site index for language models is at /llms.txt.