What is a seed network in email warm-up?
Every warm-up tool has one. Very few explain it. The seed network is the pool of mailboxes your warm-up emails are sent to — and the quality of that pool determines whether your warm-up score means anything at all.
Key takeaways
- The seed network is the pool of mailboxes your warm-up emails land in; their engagement generates the signals your domain needs.
- ISPs identify synthetic seed networks quickly — typically within weeks — and discount or exclude the engagement they generate.
- Network size is not network quality: a network of real business inboxes is harder to distinguish from genuine traffic than a large synthetic one.
How warm-up works — and why the network is the whole thing
Email warm-up works by creating a pattern of positive engagement signals from your sending domain: emails sent, emails received, emails opened, emails moved to inbox when they land in spam. These signals are picked up by ISP machine learning models (Gmail, Outlook, Yahoo) and used to assign your domain a reputation score that affects where future emails land.
The mailboxes on the receiving side — the ones your warm-up emails land in — are the seed network. Their engagement behaviour generates the positive signals your domain needs. Whether those signals are meaningful to an ISP depends entirely on what the seed mailboxes look like.
Real business inboxes vs synthetic addresses
Most warm-up tools use one of two approaches. The difference is not a minor technical detail — it determines whether the positive engagement your warm-up generates is recognised as genuine by ISP models.
Real business inbox
Synthetic seed address
Why ISPs identify synthetic networks
Gmail and Outlook train machine learning models on billions of real inbox interactions. A synthetic mailbox that receives 300 warm-up emails per day, opens every one, never marks anything as spam, and has no other email activity produces an engagement profile that is statistically impossible from a genuine business inbox.
ISPs identify this pattern quickly — typically within weeks of a large synthetic network launching. Once a seed pool is fingerprinted, the positive engagement signals it generates are discounted or excluded from the reputation calculation entirely. The warm-up appears to be working. The score climbs. The actual reputation is not changing.
“A mailbox that receives 500 warm-up emails a day, opens every one, never reports spam, and has no other email activity has an anomalous engagement profile. It looks like what it is: a warm-up farm.”
Network size is not network quality
Warm-up tools frequently advertise network size: 500,000 inboxes, 1 million, 4 million. Larger is promoted as better. The opposite is often true for synthetic networks.
A synthetic network of 500,000 addresses shared by 100,000 senders means each seed address receives warm-up traffic from hundreds of different senders every day. The engagement profile becomes more obviously artificial as the network scales, not less. ISP models converge on it faster.
A real inbox network of 5,000 inboxes, each receiving warm-up traffic from only a small number of senders at modest daily volumes, is harder to distinguish from genuine business email traffic. Quality — specifically, real engagement from indistinguishable mailboxes — is what drives results.
Seed networks and placement tests: the same problem
Placement tests use seed networks too — they send a test email to seed mailboxes and report where it landed. If the seed panel for the placement test is 35 inboxes, the result is based on 35 data points. If those 35 inboxes are clustered in one provider, the test says nothing about the others.
The same question — are these real inboxes, and are there enough of them across the right providers — applies to placement testing as much as to warm-up. A placement tool reporting 100% inbox placement based on 5 seed inboxes is not giving you a reliable signal.
Five questions to ask any warm-up tool about its network
A warm-up tool that cannot answer “who is in my network and how do I know?” is not giving you a warm-up score. It is giving you a score from a test it designed itself, against mailboxes it controls.
See what network InboxPlease uses — for free
Run a free placement test. No account required. Your results include the sample size and provider breakdown behind every score.
Run free testFounder & CEO of InboxPlease. Over a decade building email infrastructure for SaaS businesses; writes about deliverability, inbox placement and sender reputation.