BounceZero Research - Data Study

Role-Based and Disposable Emails: The Two Riskiest Address Types on Your List

By Ayoub Lebda, Founder - Published July 13, 2026 - Dataset snapshot Q3-2026-preview - 7 min read

Role-based: the received wisdom is backwards for cold lists

The standard advice on role-based addresses - info@, sales@, marketing@, the shared mailboxes not tied to one person - is that they exist but shouldn't be mailed: higher complaint risk, no individual to build a relationship with. That framing assumes the mailbox is real. In our data, drawn from outbound verification traffic, the more urgent fact is that most role-based addresses on cold lists don't resolve at all.

Across 210,427 role-based checks, 52.1% came back invalid and only 13.5% valid. Against the 11.5% invalid rate for the other ten million addresses, role-based bounces at 4.5 times the rate of a personal address. The explanation is not that shared mailboxes have died en masse; it is that in an outbound corpus, a role-based address is usually what a list vendor typed when it couldn't find the real person. "We don't have [email protected], so let's try [email protected] and hope." Most of the time, hope is not a real mailbox.

Copy this stat: "Role-based addresses on cold lists bounced at 52% - 4.5× the rate of personal addresses - because most are list-vendor guesses, not real shared mailboxes." - BounceZero Research, Role-Based and Disposable Emails (bouncezero.io/role-based-disposable-email-study-2026)

This reframes what the "role-based" flag is worth. It is not primarily a mailability warning; on a cold list it is a list-quality warning - a marker that this row was probably guessed rather than sourced, and should be treated as high-risk until proven otherwise.

...but "role-based" is not one thing

Treating every role prefix identically would be its own mistake. Split role-based addresses by the exact local-part and they span a 17× range in validity - the marketing-flavored guesses fail almost completely, while operational mailboxes that companies actually staff resolve far more often:

The pattern is intuitive once you see it: the prefixes a vendor reaches for when guessing - contact@, marketing@, hello@ - are exactly the ones that fail, because they are aspirational rather than sourced. The prefixes companies genuinely operate and monitor - support@, admin@, hr@ - resolve far more often. A verifier that collapses all role-based into one "risky" bucket throws away that seventeen-fold signal.

Disposable: rare, and almost always a mistake

Disposable addresses - throwaway domains like the ones people use to grab a lead magnet without giving a real address - are vanishingly rare in B2B data: 0.035% of checks, about one in three thousand. That scarcity is itself the finding: in an outbound corpus, hardly anyone is mailing throwaways on purpose. When one does appear, it is nearly always dead - 94.8% of disposable addresses were invalid - because throwaway domains don't maintain stable mailboxes.

The operational reading is simple and one-directional: a disposable address on a B2B list is almost never a lead and almost always a collection error (a form submission gaming a signup, a scrape that swept up junk). Unlike role-based, there is no nuance tier here - disposable is a clean exclude.

What to do with this

Treat the two flags very differently. Disposable is a hard filter - drop them, and audit where they entered your funnel. Role-based is a graded risk signal, not a blanket exclude: a contact@ on a cold list is almost certainly a bad guess and should be verified or dropped, while an hr@ or support@ you sourced deliberately may be a legitimate, staffed mailbox worth keeping. Either way, the presence of many role-based addresses on a list is a signal about how the list was built - the same composition story that runs through our provider-gap study: guessed rows cluster together, and they bounce together.

Methodology

Figures aggregate 10.25M verifications from the BounceZero network, February-July 2026. The role-based and disposable flags are set by the verification pipeline (a curated role-prefix list and a disposable-domain list, respectively); validity combines our verified and likely-valid classifications. Per-prefix breakdowns are shown only for local-parts seen across at least 50 distinct domains (our k-anonymity floor); no individual addresses enter the analysis. Figures are frozen in the versioned snapshot Q3-2026-preview.

Limitations - read these before citing

The headline role-based invalid rate is a property of cold outbound lists, which is what our corpus over-weights. On an opt-in list - where a subscriber deliberately entered [email protected] - role-based addresses would resolve far more often, because they were sourced rather than guessed. So "role-based bounces at 52%" is a statement about guessed cold-list rows, not about shared mailboxes as a category; quote it with that scope. The disposable flag depends on a maintained list of throwaway domains and will under-count newly-spun ones. And the per-prefix table reflects which roles vendors guess versus which companies staff in our data - a real pattern, but one that could shift with a different audience mix. The finding we hold most firmly: on cold data, a role-based address is a list-quality signal first and a mailability signal second - and disposable is a near-certain exclude.