Can I mix single-match and multi-match records in the same send?

No — and the reason is the multi-match set is a superset that includes historical addresses. Mixing them means the send volume includes records that are by construction less likely to be deliverable. Run them as separate flows with separate sender domains if you run them at all: single-match on the warmed tenant, multi-match through a lower-volume waterfall-append pattern.

What bounce rate should trigger a pause on a campaign?

Over 4% bounce rate on a single send is a red flag — it suggests the underlying list is stale or contaminated. Over 5% should pause the send program while the tenant cools. The durable target on a warmed tenant with a current list is under 2%; anything consistently over 3% indicates a procurement-quality problem upstream.

How does this interact with B2B email procurement?

B2B email has the same underlying mechanics but different baseline numbers — B2B bounce rates run higher because employees change roles and domains. Our companion piece on B2B prospecting in 2026 covers the B2B-specific pattern; much of the advice there is that bulk cold email at scale is no longer the viable motion.

Is buying email data still legal and in-policy for the major ESPs?

Buying email data remains legal under CAN-SPAM provided the buyer manages consent, opt-out, and identification correctly; the constraint is increasingly the ESP-level policies. Most major ESPs prohibit sending to purchased lists on shared IPs and restrict or prohibit it on dedicated IPs too. Buyers who run a purchased file typically run it through their own SMTP infrastructure on a warmed dedicated domain, not through a mainstream ESP.

Email Data: Multi-Match vs Single-Match Buying

Buying consumer email data in 2026 looks nothing like buying it in 2018. Sender-authentication tightening at Google and Yahoo, tenant-level reputation scoring at Microsoft, and receiving-side behavioral filters have raised the cost of a bad send materially. The data itself has split into two procurement shapes — single-match and multi-match — and the buyer who does not understand the difference between them buys the wrong file and watches their sender reputation evaporate. GSDSI's core email file publishes both shapes at scale (314M single-match, 585M multi-match records); this piece explains when each belongs in the mix.

Key Takeaways

Single-match email data is one verified email address per record, tied to a single confirmed identity — the shape buyers run for primary outreach.
Multi-match email data includes multiple historical email addresses per identity, used for waterfall append and reactivation, not direct bulk send.
The 2024 Google/Yahoo sender rules and Microsoft's tenant-scoring made sender reputation the binding constraint on email marketing economics.
The durable posture is a warmed sender, a verified list at or under 2% bounce, a narrow per-send volume, and consent documentation that survives compliance review.

What Single-Match Actually Means

A single-match email record pairs one verified email address with one confirmed consumer identity — usually HEM-plus-name-plus-postal, with a recency timestamp on both the email verification and the identity anchor. The email has been SMTP-validated (the address accepts mail at the domain level), the identity behind it has been resolved against a primary identity source, and the record carries a consent pedigree the buyer can document. Single-match is the file buyers run for primary outreach: welcome flows, transactional overlays, and the paid bulk sends that a warmed domain can actually deliver.

The operational anchor is verification recency. A single-match record verified 18 months ago and not re-checked is, in 2026, functionally a guess. The best-in-category files re-verify on a rolling 60–120 day cadence; anything beyond 180 days without re-verification belongs in a different bucket. The FTC's CAN-SPAM compliance guide frames the regulatory floor; Google's bulk-sender guidelines and Microsoft's sender-reputation best practices frame the operational one.

What Multi-Match Is Actually For

A multi-match email record pairs a single consumer identity with the list of email addresses that identity has been associated with over time — current, prior work, prior personal, lapsed and reactivated. The file is larger per identity (typical 1.8–2.2x the single-match count), but the use pattern is not bulk send. Multi-match is a waterfall-append tool: for reactivation campaigns, for when a primary address bounces or marks spam, for bridging a known identity to the current deliverable address without restarting acquisition. Using multi-match as a bulk-send source is the specific error that breaks sender reputation — addresses in the multi-match set are, by construction, historical and therefore overrepresented in the 'no longer monitored' cohort.

Why Deliverability Is the Binding Constraint

Email procurement arguments in 2024–2026 have converged on one point: the quality of the data matters because the cost of sending to bad data is no longer linear. A 5% bounce rate on a warmed tenant is not 5% wasted send; it is a reputation event that depresses inbox placement across every campaign for the following weeks. Sender-reputation scoring at Microsoft 365, Gmail, and the enterprise inbound filters now penalizes the whole tenant for the worst-performing domain in that tenant. The operational rules that follow:

Keep bounce rate under 2% on the warmed sender, which in practice means buying single-match with recent verification and running list hygiene before every send.
Cap spam-complaint rate at 0.1% — a threshold enforced by Gmail's bulk-sender rules and watched closely by Microsoft.
Warm new sending domains over 8–12 weeks; do not drop a $2M email acquisition into a cold tenant.
Segment by engagement — the durable revenue cohort is the recently-opened segment; suppress the 180-day-unopened set from high-volume flows.
Use multi-match only for waterfall append when the primary address fails, never as a bulk primary source.

Advertisers' general counsel offices have tightened the consent bar on email acquisition. The buyer asking 'where did this list come from' in 2026 is asking about documented opt-in provenance, not just opt-out compliance. The FTC's privacy enforcement record and state-level privacy laws (California, Colorado, Virginia, Connecticut, Texas) have made consent documentation a compliance artifact the buyer needs to retain, not just trust. Data providers who cannot produce the consent chain from end consumer to end buyer are increasingly the ones getting dropped in annual vendor review. A companion piece on what privacy-safe actually means when buying location data applies the same framework to the location-data side.

How to Decide Which Shape You Need

The decision rule: use single-match for any flow that will actually send to the address. Use multi-match to append into an existing first-party file, to rescue a bouncing record, or to reactivate a dormant account with a known identity. Never buy multi-match intending to run it as a bulk-send list — that is the single most reliable way to burn a warmed tenant in 2026. The B2C data solutions on GSDSI's catalog frame where single-match and multi-match fit into the larger audience-targeting and acquisition stack.

Frequently Asked Questions

Can I mix single-match and multi-match records in the same send?: No — and the reason is the multi-match set is a superset that includes historical addresses. Mixing them means the send volume includes records that are by construction less likely to be deliverable. Run them as separate flows with separate sender domains if you run them at all: single-match on the warmed tenant, multi-match through a lower-volume waterfall-append pattern.
What bounce rate should trigger a pause on a campaign?: Over 4% bounce rate on a single send is a red flag — it suggests the underlying list is stale or contaminated. Over 5% should pause the send program while the tenant cools. The durable target on a warmed tenant with a current list is under 2%; anything consistently over 3% indicates a procurement-quality problem upstream.
How does this interact with B2B email procurement?: B2B email has the same underlying mechanics but different baseline numbers — B2B bounce rates run higher because employees change roles and domains. Our companion piece on B2B prospecting in 2026 covers the B2B-specific pattern; much of the advice there is that bulk cold email at scale is no longer the viable motion.
Is buying email data still legal and in-policy for the major ESPs?: Buying email data remains legal under CAN-SPAM provided the buyer manages consent, opt-out, and identification correctly; the constraint is increasingly the ESP-level policies. Most major ESPs prohibit sending to purchased lists on shared IPs and restrict or prohibit it on dedicated IPs too. Buyers who run a purchased file typically run it through their own SMTP infrastructure on a warmed dedicated domain, not through a mainstream ESP.