The Seven-Step Delivery Model

A useful mental model for organizing deliverability work: every email program cycles through seven steps, and each step feeds the next. Steps 1–2 and 6–7 are primarily the sender's (marketer's) responsibility; steps 3–5 lean on the sending platform. The whole loop rests on the golden rule: send interesting and valuable content to people who have agreed to receive it — the same wanted/expected principle in Foundations of Email Deliverability.

        ┌─────────────────────────────────────────────┐
        ▼                                             │
1 Assemble → 2 Create → 3 Transmit → 4 Recipient → 5 Track
  the list     content     cleanly      issues       results
        ▲                                  │            │
        └────────── 6 Manage hygiene ◄─────┘   7 Fine-tune content ──► (2)

Step	What it means	Covered in depth
1. Assemble the right list	An audience that wants, consented to, and expects the mail, with accurate current addresses; segmented and targeted.	Consent Methods, List Hygiene, M3AAWG Senders BCP
2. Create great content	Well-written, valuable, recognizably branded, legally compliant content with truthful subject lines.	Content & Design for Deliverability, compliance articles
3. Transmit the message cleanly	Authenticated, RFC-compliant sending at rates each receiver will accept, from infrastructure with earned reputation.	Sending Infrastructure Practices, SPF/DKIM/DMARC, IP Warm-Up
4. Deal with recipient issues	Classify and act on complaints, bounces, throttling, filtering, unsubscribes (taxonomy below).	Complaint Feedback Loops, Enhanced Status Codes, DSNs
5. Track the results	Measure delivery, opens, clicks, and business outcomes; compare campaigns over time.	Metrics & Benchmarks (including open-rate measurement caveats), Reputation Monitoring
6. Manage list hygiene	Feed step-4 signals back into the list: remove hard bounces and complainers immediately, retire repeat soft-bouncers, sunset the unengaged.	List Hygiene & Sunset Policies
7. Fine-tune content	Use step-5 data to optimize what gets sent next — subject lines, offers, frequency, segmentation.	Content & Design, Two Worlds (the "less is more" economics)

The model's value is the feedback loops: recipient issues (4) must update the list (6 → 1), and results (5) must update the content (7 → 2). A program that runs steps 1–3 but never closes the loops accumulates stale addresses and unengaging content — exactly the profile mailbox providers score down.

Division of responsibility

The white paper's split holds up as a way to structure sender↔platform conversations: marketer-controlled reputation inputs are complaints, hard bounces, spam-trap hits, and content; platform-controlled inputs are RFC/technical compliance, authentication mechanics, transmission-rate management, and bounce/FBL processing. Neither side can compensate for the other's failures — and a platform is not a list-cleaning service: it automates ongoing maintenance, but starting with a properly consented list is the sender's job.

Recipient-issues taxonomy (step 4)

Once a message is handed off, six kinds of things go wrong, in roughly descending order of reputation impact:

Issue	What happens	Reputation effect	Sender action
Spam/junk complaints	Recipient flags the message; the provider both junks future mail for that user and counts the complaint against the sender's IP/domain. Reaches the sender via feedback loops where offered (Gmail exposes only aggregate rates via Postmaster Tools, not per-message reports).	The single strongest negative signal.	Suppress the complainer immediately; fix the acquisition or content problem generating complaints. Make unsubscribing easier than complaining — a visible unsubscribe (top of long messages, or top and bottom) plus one-click List-Unsubscribe, now mandated by Gmail/Yahoo/Microsoft for bulk senders.
Hard bounces	Permanent failure (5xx): address or domain invalid or unreachable.	Providers monitor hard-bounce rates closely and may block a campaign mid-send past their threshold.	Remove after a single hard bounce; never reactivate (recycled-trap risk — see Spam Traps).
Soft bounces	Transient failure (4xx): full mailbox, greylisting, timeouts, congestion, per-domain/IP throttling.	Neutral if handled; repeated soft bounces to the same address signal staleness.	Retry per RFC 5321 semantics; suppress addresses that soft-bounce across multiple consecutive sends.
Transmission rate errors	Sending too much too fast — especially from IPs with little history — triggers timeouts, throttling, and blocks.	Damages IP reputation directly.	Respect per-provider rate expectations; warm up new IPs; let the platform adapt retry/throughput per receiver in real time.
Client-level spam filters	Beyond the provider's gateway filtering, per-user filters (blocked-sender lists, learned personal models) can junk mail that was accepted at SMTP time.	Invisible to the sender except as absent engagement.	Earn per-user trust: recognizable From identity, replies, address-book adds.
Deletes-without-reading	The message arrives but the recipient deletes it unopened.	Providers increasingly weight engagement, so chronic non-engagement erodes reputation even with zero complaints.	Step 7: better targeting, subject lines, frequency; sunset the chronically unengaged.
Unsubscribes	Recipient opts out through the provided mechanism.	Not harmful — an unsubscribe through your own link is the healthy exit path and infinitely preferable to a complaint.	Process instantly and without friction (no login, no retyping, no delay); continuing to mail after opt-out is illegal in most jurisdictions.

The diagnostic habit: when delivery degrades, first classify which of these is happening — a transmission issue, a list/content issue, or accumulated reputation damage — because remedies differ completely. See Reputation Monitoring for the monitoring stack and recovery workflow.

The noreply@ problem

The From address is part of the content (step 2) and the recipient-issue surface (step 4). Sending from noreply@ — an address that is unmonitored or doesn't exist — tells subscribers "please don't talk to us" and costs the sender concretely:

1. Mail from noreply@ is more likely to be ignored, deleted, or flagged as spam — modern clients display the address, not just the display name.
2. Replies are lost customer feedback, including reply-based unsubscribe attempts; a missed "remove me" reply often converts into a spam complaint.
3. Replies are a positive signal: a recipient replying often auto-adds the sender to their contacts/safe-senders, improving placement of all future mail for that user.
4. Observed effects: lower open rates, higher unsubscribe and complaint rates.

Two authorities back this. CAN-SPAM requires From/Reply-To and routing information to be accurate and identify the initiating business (see CAN-SPAM); the M3AAWG Senders BCP requires senders to be able to process unsubscribe requests sent to the From and Reply-To addresses. Use a monitored, real address (hello@, news@, info@), and keep functioning abuse@ and postmaster@ mailboxes on the sending domain.

Aged claims in the source, corrected

The 2018 white paper predates several shifts; where its advice has aged, the current reality is:

• Authentication: it listed "SPF/Sender ID, DomainKeys and DKIM" with none dominant. Sender ID and DomainKeys are long obsolete; the current baseline is SPF + DKIM with DMARC on top, and major providers require authentication for bulk senders — see Gmail, Yahoo, Microsoft.
• Third-party accreditation/whitelisting (Goodmail, Habeas, Return Path certification): these programs are defunct or marginal; providers now rely on their own reputation systems. Recipient-level trust (contacts/safe-sender lists) still matters; the industry term for sender-side lists is now "allowlist."
• Keyword-based spam scores ("avoid words like free, casino"): modern filters are holistic and reputation-driven; trigger-word lists are largely a myth — see Content & Design for what actually matters in content.
• Open tracking: since Apple Mail Privacy Protection (2021), opens are inflated and unreliable as an individual signal — see the measurement caveats in Metrics & Benchmarks.

Related

• Foundations of Email Deliverability — the wanted/expected principle underlying every step
• Two Worlds of Email Deliverability — the strategic frame for steps 5–7
• Consent Methods and the List-Quality Spectrum — step 1 in depth
• Metrics and Benchmarks — the numbers behind steps 4–6