Email Deliverability for Agent Outreach: Setting Up SPF, DKIM, DMARC and Domain Warmup Correctly
Email deliverability in cold outreach refers to the probability that agent-generated outreach emails actually land in the inbox rather than in spam. It depends on authentication (SPF, DKIM, DMARC), separate sending infrastructure, gradual domain warmup and consistent monitoring of bounce and complaint rates. Without this setup, any volume outreach burns sender reputation.
Key Takeaways
- ✓Volume outreach without SPF, DKIM and DMARC reliably lands in spam at Gmail and Microsoft 365 or gets blocked entirely - authentication is mandatory, not a nice-to-have.
- ✓Cold outreach never runs through the main domain, but through separate sending domains or subdomains with their own mailboxes, so that the brand domain's reputation stays protected.
- ✓Domain and mailbox warmup realistically takes 4 to 8 weeks; after warmup, roughly 30 to 50 cold emails per mailbox per day is a conservative upper limit.
- ✓A bounce rate below 2-3 percent and a spam complaint rate within the Google Postmaster target of below 0.1 percent (hard limit 0.3 percent, as of 2026) are the reputation thresholds an agent stack must monitor continuously.
- ✓In the DACH region, technical deliverability is only half the battle: UWG §7 (DE), TKG (AT) and revDSG (CH) restrict B2B cold outreach legally more strongly than US playbooks suggest.
Email deliverability in cold outreach refers to the probability that agent-generated outreach emails actually land in the inbox rather than in the spam folder. It depends on four levers: correct authentication (SPF, DKIM, DMARC), separate sending infrastructure, gradual domain and mailbox warmup, and ongoing monitoring of bounce and complaint rates. Without this foundation, any volume outreach burns sender reputation - and the best AI-generated text is worthless if no one sees it.
This is exactly where agent-driven outreach fails. AI agents push the marginal cost per email towards zero and tempt teams to crank up volume. Yet providers do not assess the content first, but the sender. A stack that ignores this logic produces the failure pattern most frequently observed in practice in DACH B2B sales: deliverability collapse from AI-generated outbound at scale.
The three quick answers up front:
- Authentication is mandatory. Without SPF, DKIM and DMARC, Gmail and Microsoft 365 (as of 2026) effectively block volume outreach or sort it into spam.
- Never via the main domain. Cold outreach runs via separate sending domains or subdomains with their own mailboxes, so that the brand domain's reputation remains untouched.
- Warmup requires patience. Four to eight weeks of gradual increase; after that, a conservative upper limit of around 30 to 50 cold emails per mailbox per day applies.
Why volume outreach lands in spam without a proper setup
Receiving servers decide within milliseconds whether an email goes to the inbox, to the spam folder or straight to rejection. This decision relies on signals that almost all have nothing to do with the email text:
- Authentication status - do SPF, DKIM and DMARC pass?
- Domain and IP reputation - is there a clean sending history, or does the domain appear for the first time with high volume?
- Engagement signals - are emails opened, replied to, or marked as spam and deleted?
- Volume patterns - does the sending volume rise organically or in sudden jumps?
A freshly registered domain that sends hundreds of nearly identical messages without any lead-up meets every classic spam pattern simultaneously. The fact that an AI agent has individualised the text changes little - the rejecting decision is made at the reputation level, before the content is even weighted. To make matters worse, DACH B2B inboxes quickly recognise templated AI text as such and manually mark it as spam, which drives up the complaint rate and damages reputation further.
SPF, DKIM and DMARC: the authentication foundation
These three DNS-based standards are the entry ticket to the inbox. They only work together.
SPF (Sender Policy Framework) is a DNS TXT record that specifies which servers are allowed to send on behalf of a domain. The recipient checks the sending IP against this list.
DKIM (DomainKeys Identified Mail) cryptographically signs every outgoing email. The public key resides in the DNS; with it, the recipient verifies that the email was not altered in transit and actually originates from the authorised domain.
DMARC (Domain-based Message Authentication, Reporting and Conformance) connects SPF and DKIM, defines via policy how unauthenticated emails are handled, and provides aggregate reports. The policy has three levels: none (observe only), quarantine (to spam) and reject (reject). For your own sending domains, an enforced policy is the goal.
Record | Purpose | Example value (pseudocode) |
|---|---|---|
SPF | Define permitted sending servers | |
DKIM | Cryptographically sign emails | |
DMARC | Link policy + reporting | |
MX | Incoming mail / bounce handling | |
Custom tracking domain | Link tracking without reputation leak | |
A common mistake: running tracking via the generic domain of the sending tool. These domains are often already on blocklists due to other users - a dedicated tracking subdomain avoids this reputation leak. As of 2026, the major mailbox providers explicitly require SPF, DKIM and DMARC as well as a low complaint rate from bulk senders; the setup is therefore a minimum standard, not an extra.
Sending infrastructure: domains, mailboxes and IP strategy
The second pillar is the separation of infrastructure. Three principles:
Separate sending domains instead of the main domain. Cold outreach never belongs on the brand domain. If the outbound damages reputation, neither transactional nor internal emails may be affected by it. Common choices are specially registered domains that resemble the brand (such as company-team.de, try-company.de) or a dedicated subdomain. Each sending domain carries its own reputation.
Multiple mailboxes per domain, inbox rotation. Typically two to three mailboxes are set up per domain. The agent stack distributes the sending volume across all available mailboxes (inbox rotation) rather than overloading a single one. Scaling is achieved through breadth - more domains and mailboxes - not through the depth of a single mailbox.
Dedicated vs. shared IPs. With the major mailbox platforms via standard inboxes (Google Workspace, Microsoft 365), you send via their shared, high-reputation infrastructure anyway. A dedicated IP only pays off with very high, constant volume over your own sending infrastructure, because a dedicated IP needs its own warmup and, at low volume, performs worse than an established shared IP. For most DACH mid-market outreach setups, the shared approach via standard mailboxes is the more pragmatic path.
The warmup process with a concrete timeline
Warmup means building a credible sending and interaction history for a new domain and its mailboxes before genuine cold volume flows. During the warmup phase, the mailboxes - often automated via warmup pools - exchange emails with other inboxes that are opened, replied to and rescued from spam. These positive signals build reputation.
A robust, conservative timeline:
Phase | Period | Volume per mailbox/day | Activity |
|---|---|---|---|
Domain maturation | Day 0-14 | 0 | DNS setup (SPF/DKIM/DMARC) completed, domain ages, no cold sending yet |
Warmup start | Week 1-2 | 2-5 | Automated warmup traffic only, no real prospects |
Increase | Week 3-4 | 5-20 | Warmup continues, first small genuine campaign tranches |
Ramp-up | Week 5-6 | 20-40 | Genuine cold volume dominates, warmup in the background |
Operation | from week 7 | 30-50 | Stable plateau, warmup share remains for reputation maintenance |
Important guardrails: the domain itself should exist for at least two to three weeks before the first cold sending. The volume is never increased in sudden jumps, but in small steps. And warmup does not end after ramp-up - a certain share of warmup traffic is maintained permanently to support reputation.
Volume limits, spam triggers and monitoring
After warmup, a conservative benchmark is an upper limit of around 30 to 50 genuine cold emails per mailbox per day. Anyone wanting to send more adds mailboxes and domains.
Typical spam triggers an agent stack must avoid:
- volume per mailbox that is too high, or sudden jumps in increase
- nearly identical texts across many recipients (templated AI output)
- many links, images, tracking pixels or attachments in cold emails
- poor list quality with many invalid addresses (hard bounces)
- aggressive spam trigger words and purely sales-driven language
Metrics that are monitored continuously:
- Bounce rate below 2-3 percent. Higher values indicate poor list quality - verify addresses before sending.
- Spam complaint rate ideally below 0.1 percent. This is the target value recommended by Google for bulk senders (as of 2026); the hard limit documented in Google Postmaster Tools, beyond which deliverability tips over, is 0.3 percent.
- Reply and positive reply rate as an engagement and reputation signal.
- Domain and IP reputation in Google Postmaster Tools and in Microsoft SNDS.
If the complaint or bounce rate rises, volume is throttled immediately and the cause is investigated, rather than simply continuing to send.
Example: scaling from 0 to 500 emails per day
An agency wants to deliver 500 personalised cold emails per working day to verified DACH B2B contacts for a mid-market client.
- Infrastructure: At 40 emails per mailbox per day, roughly 13 mailboxes are required. Split across about 5 sending domains with 2-3 mailboxes each, this provides enough redundancy for inbox rotation.
- Setup: SPF, DKIM, DMARC, a dedicated tracking subdomain and clean bounce handling are configured on all 5 domains. Domains mature for 2 weeks.
- Warmup: 6 weeks of gradual increase per mailbox according to the timeline above. Only from week 7 does the full campaign run.
- Monitoring: Daily checks of bounce rate (target below 2-3 percent) and complaint rate (target below 0.1 percent, hard limit 0.3 percent) across all domains.
- Realistic time horizon until full operation: around 8 weeks - 2 weeks of maturation plus 6 weeks of warmup. Anyone skipping this lead-up risks burned domains and a fresh start.
On the tooling side, this stack typically includes specialised outbound platforms such as Smartlead, Instantly or Lemlist for sending, rotation and warmup, complemented by enrichment tools such as Clay or Apollo. For the DACH market, Dealfront (Karlsruhe) is the most firmly established, GDPR-native source for sales intelligence and contact data. These tool mentions are as of 2026.
The DACH reality check: deliverability is not compliance
A good deliverability setup only determines whether an email arrives - not whether it is legally permitted. In the DACH region, this is the decisive point. UWG §7 (Germany) as well as the parallel provisions under the TKG (Austria) and the revised DSG (revDSG, Switzerland) restrict B2B cold outreach considerably more strongly than US outreach playbooks suggest. Presumed consent is narrow and legally contested. Technical deliverability and legal admissibility must therefore always be planned together; the compliance mechanics in detail are covered by the dedicated cluster article on cold outreach compliance in DACH.
On top of that: LinkedIn is enforcing its terms of service against automation tools increasingly hard, which is why a pure volume approach is risky there too. The robust DACH path is rep-in-the-loop augmentation with clean infrastructure, genuine personalisation and conservative volume - not maximum throughput at any cost.
For agencies and B2B teams
For marketing agencies, robust sending infrastructure is the prerequisite for any scalable outbound service: those who set up sending domains, warmup timelines and monitoring cleanly protect the reputation of every client account and reproducibly deliver measurable delivery rates. For B2B decision-makers in the DACH region, it means planning two weeks of maturation plus around six weeks of warmup and a parallel compliance review (UWG/TKG/revDSG) before any agent outreach initiative - only then does the agent stack pay off as a pipeline driver rather than burning domain reputation. Anyone wanting to build cold outreach infrastructure and DACH-compliant agent pipelines should treat setup, warmup and legal review as a single joint project from the very start.
FAQ
Why does agent-generated cold outreach land in spam without a proper setup?
What is the difference between SPF, DKIM and DMARC?
How long does a domain warmup for cold outreach take?
How many emails may I send per mailbox per day?
Which metrics must I monitor in outreach monitoring?
Is good deliverability sufficient for cold outreach in the DACH region?
Want to go deeper?
Get new analyses straight to your inbox – or see how we put this knowledge to work for companies.