Cold email bounce rate by recipient security gateway: Mimecast, Sophos, Proofpoint, Microsoft 365, Google
We resolved the MX records of every bounced and accepted domain in two campaigns. Domains behind Mimecast bounced 90 to 1; Sophos 8 to 0; Proofpoint 0 bounces on 21 domains. Half of all our bounces across 5,650 sends were Mimecast.
| Gateway in MX | Campaign A (marketing leaders, hiring signal) | Campaign B (DTC brands) | Other campaigns |
|---|---|---|---|
| Mimecast | 90 domains bounced / 1 accepted | 30 bounced / 1 accepted | about 99% bounce |
| Sophos | 8 bounced / 0 accepted | same pattern, small volume | 100% in the sample |
| Proofpoint | 0 bounced / 21 accepted | 0 bounced | about 1% |
| Microsoft 365, Google, Barracuda, Cisco IronPort, Trend Micro, Fortinet, Zoho, GoDaddy | about 1% | about 1% | about 1% |
| Campaign | Share of pool removed | Bounce per lead before | Bounce per lead after |
|---|---|---|---|
| A (hiring signal, marketing leaders) | 16.7% | 16.95% | 0.84% (116 of the 121 bounces were Mimecast) |
| B (DTC brands) | 1.3% | 1.76% | 0.72% predicted, 0.30% measured in production on the next sends |
Why no verifier sees it
Mimecast rejects at the SMTP handshake based on the reputation of the sending domain, not on whether the mailbox exists. A fresh sending domain (lookalike, recently warmed) is refused whatever the address. Verifiers test whether the recipient domain accepts a recipient at the SMTP stage, which Mimecast gladly does before rejecting the actual delivery. The result is a bounce on addresses scored 99/100 and not catch-all. In the platform logs these show up with a generic provider code rather than a Google or Microsoft code.
On a 5,650-send baseline, 51% of all bounces came from Mimecast-gated domains, which bounced at 98%. Outside Mimecast, addresses rated 'Deliverable' bounced at 4.0% and catch-all addresses scored 90 to 99 bounced at 2.6%.
The strict-verifier test that still bounced at 16.9%
Before we had the MX filter, 714 leads survived the strictest verifier gate we used (score 80 or more and not catch-all; that gate had already discarded 37% of the pool). Sent, they produced 121 bounces per lead, 16.9%. The bounced domains were large organizations (transport federation, health insurer, real estate, SaaS) behind Mimecast or Proofpoint-type gateways. Verification cannot fix a pool whose domains are the problem.
- A pool that skews toward large enterprises skews toward gateways and will bounce structurally on young sending domains, whatever the verification score.
- A hiring signal (open roles) pulls toward large organizations because they always have open roles; it selected exactly these domains.
Proofpoint is innocent
Earlier post-mortems blamed 'Mimecast/Proofpoint'. The MX breakdown shows Proofpoint-gated domains at 0 bounces on 21 accepted domains in campaign A. Blocking Proofpoint would discard clean volume. Only Mimecast and Sophos are filtered.
The filter
One DNS MX lookup per recipient domain, cached, about 25 in parallel, zero cost. A domain is removed when any MX host contains 'mimecast' or 'sophos'. Cost of the filter: 1 to 17% of a pool depending on the population (B2B enterprise lists lose far more than DTC brand lists). This step now runs before every send, before the verifier.
What this measurement does not say
- Measured on young sending domains (lookalikes under a few months old). An established corporate domain with years of reputation may pass Mimecast; we did not measure that.
- Per-gateway percentages are domain counts, not send counts; small gateways (Sophos: 8 domains) carry wide error bars.
- The 0.30% production figure for campaign B is from one week of sends after the filter was applied.
Questions
Should I exclude Mimecast domains from cold email lists?
If your sending domains are young, yes: in our data they bounced at about 99%. Exclude them with a DNS MX lookup before verification; it is free and removes about half of the bounces.
Does Proofpoint block cold email?
Not in our measurement: 0 bounces on 21 Proofpoint-gated domains. Do not filter Proofpoint.
Related measurements
- How much of a 'verified' B2B contact list is actually catch-all or dead: GetLeads, Apollo, LeadMagic measured
- AgentsVerify vs No2Bounce vs BillionVerify on the same 455 real bounce outcomes
- The 'companies that are hiring' signal selected 283 software vendors and 36 real clinics: measured on a healthcare outbound list
Want this run on your pipeline?
We operate multichannel outbound (cold email, LinkedIn, paid retargeting, newsletters) for B2B companies, and we measure it the way these pages do.
Book a call