Email infrastructure
Most outbound email runs on rented infrastructure. A seat on a shared platform, a pool of domains someone else controls, a reputation tied to senders you will never meet. When that reputation degrades or the platform suspends the account, the programme stops. We took a different path. We built and operate our own sending infrastructure so that a private equity operating team can run sustained outbound programmes across multiple portfolio companies without depending on a third party platform that treats your volume as a liability.
Owned infrastructure begins with hardware and network resources we provision directly, not a resold seat on a marketing automation platform. We run our own mail transfer agents, configure our own sending policies, and maintain our own relationships with mailbox providers. No intermediary can throttle volume, suspend the account, or change the terms because a different customer triggered a spam event. For an operating partner managing outbound across several portfolio companies, that separation matters. A deliverability problem inside one portfolio company should not silence the others, and a platform wide policy change should not force a rebuild of your entire outreach stack.
The alternative is the model most agencies use. They buy a subscription to a sending platform, warm a handful of domains inside that platform, and call it a managed service. The platform owns the IP reputation, the domain pool, and the relationship with the receiving mail servers. If the platform decides your industry is too risky, or if another user on the same IP space gets blocklisted, your programme suffers consequences you did not cause and cannot fix. We avoid that dependency entirely. The infrastructure is ours, the reputation is ours, and the decisions about volume, cadence, and domain rotation are made by people who answer to the operating partner, not to a platform vendor.
This approach costs more to build and takes longer to stand up. It requires technical staff who understand mail server administration, DNS configuration, and feedback loop processing. It is not the right answer for a firm that wants to send a few hundred messages a week and call it done. It is the right answer for a firm that intends to run sustained, high volume outbound as a permanent capability across multiple portfolio companies, and who cannot afford to have that capability turned off by a third party.
We never send cold outreach from a company's primary brand domain. The brand domain carries the company's email, its customer communications, its transactional messages, and its internal correspondence. Burning that domain's reputation with cold outreach that generates spam complaints would be an unforced error. Instead we operate one hundred and sixteen sending domains, each configured with its own authentication records, its own warming schedule, and its own reputation trajectory. A portfolio company might use several of these domains, each mapped to a specific campaign, geography, or buyer segment.
Domain separation serves two purposes. First, it contains risk. If one domain's reputation dips because a particular list or message underperforms, the other domains continue unaffected. Second, it allows us to match the sending domain to the context of the message. A domain that references industry specific terminology in its name can improve open rates for campaigns aimed at that industry. A domain that appears to originate from a particular region can improve response rates in that region. These are small effects individually, but across a large programme they compound.
Each sending domain has its own website. Not a thin parked page, but a real site with an indexed presence, a privacy policy, a contact mechanism, and content that explains what the domain is for. Mailbox providers crawl these sites. A domain that resolves to a genuine web presence with clear identity signals passes more reliably through spam filters than a domain that resolves to nothing. We have shipped two thousand six hundred and twenty five indexed pages across the properties we operate, and that body of content supports deliverability in ways that a naked domain never could.
A new mailbox cannot send high volume on its first day. Mailbox providers treat sudden volume from an unknown sender as a strong spam signal. Warming is the process of sending a gradually increasing number of messages over several weeks, establishing a positive reputation through engagement signals before the mailbox is asked to carry a full campaign load. We warm five hundred and sixty seven mailboxes daily, every day, regardless of whether a particular mailbox is currently assigned to an active campaign.
Warming is not a set and forget process. Each mailbox provider has its own rate limiting thresholds and its own engagement metrics. A warmup schedule that works for Google Workspace recipients may be too aggressive for Microsoft recipients, or too conservative for a list that is mostly consumer addresses. We adjust warming curves per mailbox based on the destination mix, the bounce rate, the spam complaint rate, and the reply rate. When a mailbox shows signs of stress we flatten the curve and let the reputation recover before increasing volume again.
A mailbox that sits idle loses its reputation. Mailbox providers interpret silence as abandonment, and a cold restart is often harder than the initial warmup. We keep every mailbox in our pool active, sending a low baseline volume even when no campaign is assigned. That way a mailbox is always ready to scale up when a new programme launches. For an operating partner, this means the infrastructure is warm before the deal closes. We can begin outreach within days of a new portfolio company coming on board, not weeks.
Every sending domain we operate has SPF, DKIM, and DMARC records configured and monitored. SPF authorises our mail servers to send on behalf of the domain. DKIM signs each message with a cryptographic signature that receiving servers can verify. DMARC tells receiving servers what to do when authentication fails and provides a reporting channel so we can see who is sending mail that claims to be from our domains. We review DMARC aggregate reports weekly and investigate any unauthorised sending attempts.
Authentication alone does not guarantee inbox placement. We monitor placement across the major mailbox providers using seed lists that report back whether a message landed in the inbox, the spam folder, or was blocked entirely. When placement degrades for a particular provider, we diagnose the cause before the problem spreads. Common causes include a sudden spike in bounce rate from a stale list, a subject line that triggers lexical spam filters, or a domain reputation dip from low engagement. Each cause has a different fix, and applying the wrong fix makes the problem worse.
We also process feedback loops. When a recipient marks a message as spam, the mailbox provider sends a feedback report to the address we have registered for that purpose. We ingest those reports, suppress the complaining addresses immediately, and analyse patterns across complaints to identify messages or lists that need adjustment. A complaint rate above roughly one tenth of one percent is a leading indicator of deliverability trouble. We aim to keep complaint rates well below that threshold, and we adjust campaigns proactively when the rate begins to climb.
Every message we send includes a physical postal address in the footer. This is a legal requirement in several jurisdictions, but we treat it as more than a compliance checkbox. A real street address that matches the domain's identity signals to mailbox providers and to recipients that the sender is a legitimate organisation with a physical presence. We use a verifiable address, not a virtual office or a registered agent that appears on a hundred other domains.
The opt-out mechanism is a single click link that resolves to a suppression page we host. No recipient is asked to log in, confirm an email address, or complete a survey to unsubscribe. The suppression takes effect immediately and is written to a global suppression list that applies across all campaigns and all domains we operate for that client. A recipient who opts out of one campaign will not receive another message from any domain in our pool, regardless of which portfolio company the campaign serves. This is a harder standard than the legal minimum, and we apply it because a recipient who asked to stop and then receives another message is a justified spam complaint that damages deliverability for everyone.
We also maintain a manual suppression process for recipients who reply asking to be removed rather than using the link. Those requests are processed within one business day. The reply itself is routed to a human who reads it, and if the reply contains a removal request alongside other content, the removal is actioned and the other content is passed to the appropriate person. No automated system decides what a human reply means.
Platform suspensions happen. A sending platform decides that a particular industry vertical is too risky, or that a particular sending pattern looks like spam to its automated review systems, or that a complaint rate crossed an internal threshold that was never disclosed to the customer. When that happens, the account is frozen. Outreach stops. The domains the customer was using, often registered through the platform itself, may be locked. The customer has no access to the mail server logs, no ability to diagnose what went wrong, and no leverage to get the account reinstated quickly.
When you own the infrastructure, there is no platform to suspend you. The mail servers are ours. The domains are registered in our name with a registrar we control. The authentication records point to our infrastructure. If a mailbox provider blocks our traffic, we have the server logs, we can see the SMTP conversation, and we can diagnose and fix the problem directly. We do not file a support ticket and wait for a platform's trust and safety team to decide whether we are worth their time.
This matters especially in private equity, where outbound programmes often target senior executives at large organisations. Those recipients are more likely to mark a message as spam than a consumer recipient, and a platform that optimises for consumer marketing may view a complaint rate that is normal for executive outreach as a policy violation. We understand the difference between a campaign that is genuinely abusive and a campaign that is reaching a difficult audience with a relevant message. Our infrastructure is configured to handle the latter without triggering automated suspensions that a platform would impose.
The tradeoff is that we carry the operational burden ourselves. We maintain the mail servers, we patch the software, we monitor the blocklists, we handle the abuse reports. That burden is real, and it requires staffing and process that a platform offloads. For an operating partner who wants to set up outbound once and forget about it, a platform may be the better choice. For an operating partner who needs outbound to work reliably across multiple portfolio companies over multiple years, owning the infrastructure is the only way to guarantee continuity.
Most sending platforms pool their customers onto shared IP addresses. The platform's deliverability team manages the pool's reputation, and individual customers benefit from the aggregate sending behaviour of everyone else on that IP. When the pool is healthy, this works. When one customer sends a poorly targeted list that generates high complaints, the pool's reputation degrades and every customer on that IP suffers. The platform will eventually remove the offending customer, but the reputation damage can take weeks to repair, and during that time your campaigns are landing in spam.
We do not share sending infrastructure with other organisations. The IP addresses we use send only our clients' traffic. The domains we warm send only our clients' campaigns. The reputation of our sending pool is determined solely by the quality of the lists we mail and the relevance of the messages we send. If a reputation problem occurs, we caused it and we can fix it. There is no anonymous third party whose behaviour we have to investigate and no platform support team we have to persuade that the problem is not ours.
Isolation also means we can configure our sending infrastructure for the specific needs of private equity outbound. We can set custom rate limits per domain that match the tolerance of executive inboxes at large enterprises. We can implement sending patterns that mimic human sending behaviour rather than the bursty patterns that marketing platforms use. We can adjust retry schedules, connection timeouts, and TLS policies without asking a platform administrator for permission. These are small technical advantages individually, but they add up to a deliverability profile that a shared platform cannot match.
A cold email that generates a reply is a success, and that reply deserves a human response. We route every reply to a person who reads it and decides what to do. If the reply is a question about the product or service, it goes to the appropriate sales or support person. If the reply is an objection, it goes to someone who can address it. If the reply is a request to be removed, it is actioned as described above. No reply is discarded by an automated filter, and no reply receives an auto response that begins with thank you for your interest.
This is labour intensive. It requires staff who can read a reply, categorise it correctly, and route it to the right person within a few hours. For a high volume programme, that can mean processing several hundred replies a day. We staff for that volume. The alternative is the noreply address that most marketing platforms default to, which tells the recipient that the sender does not want a conversation. For outbound aimed at senior executives, where the goal is to start a conversation, a noreply address is self defeating.
Reply handling also feeds back into deliverability. Mailbox providers track whether recipients reply to messages, and a positive reply rate is a strong engagement signal that improves inbox placement. When a recipient replies and receives a thoughtful human response, the engagement signal is reinforced. When a recipient replies and receives silence or a bounce, the engagement signal is weakened. Routing replies to humans is not just good practice. It is a deliverability tactic.
Owned infrastructure is not the right choice for every situation. It is expensive to build and maintain. It requires technical staff who understand mail server operations, DNS management, and deliverability diagnostics. It takes time to warm a new domain or a new mailbox, and that time cannot be compressed without risking reputation damage. For a firm that needs to send a few hundred messages a month, or that needs to launch a campaign next week with no prior preparation, a shared platform or a managed service on someone else's infrastructure will be faster and cheaper.
It is also the wrong answer for a firm that does not have the operational discipline to maintain list hygiene, honour suppression requests, and monitor complaint rates. Owned infrastructure gives you more control, but control without discipline produces worse outcomes than a platform that enforces its own rules. If your team cannot commit to cleaning lists before every send, removing hard bounces within hours, and reviewing complaint patterns weekly, you are better off on a platform that will suspend you when you fail to do those things.
We tell prospective clients this directly, and some of them decide that owned infrastructure is not what they need. That is a good outcome. The worst outcome would be signing a client who is not ready for the operational commitment, watching their reputation degrade, and then having to explain why the infrastructure they were promised did not work. We would rather lose a deal than deliver a programme that fails because the client was not a fit.
Questions
A new domain typically requires four to six weeks of gradual volume increase before it can sustain full campaign loads. The exact timeline depends on the destination mix, the quality of the initial seed list, and the engagement signals the early sends generate. We can accelerate the timeline somewhat by using aged domains that have an existing reputation history, but we never skip warming entirely. A domain that sends high volume on day one will be blocked by most mailbox providers within hours. For an operating partner planning a post close outbound programme, we recommend beginning the warming process during the diligence period so the infrastructure is ready when the deal closes.
Domain level blocklists are a routine operational event, not a crisis. When a domain appears on a blocklist, we identify the list, determine the listing criteria, and take the corrective action required. Often this means submitting a delisting request after addressing the underlying issue, such as a stale list that generated high bounces. While that domain is delisted, we shift volume to other domains in the pool that are not affected. Because we operate one hundred and sixteen sending domains with independent reputations, a single domain's blocklisting does not stop the overall programme. The affected domain typically returns to normal operation within a few days.
A sending platform seat is cheaper because you are renting a share of shared infrastructure. That works until it does not. When the platform suspends your account, or when another customer on your shared IP damages the pool's reputation, or when the platform changes its acceptable use policy to exclude your industry, the cost of the disruption dwarfs the savings. Owned infrastructure costs more upfront but eliminates those single points of failure. It is the difference between renting a seat on a bus that might not go where you need and owning the bus. If outbound is a peripheral activity for your portfolio companies, a platform may be sufficient. If outbound is a core revenue function, the economics of ownership make sense over any reasonable holding period.
We configure suppression, data handling, and consent mechanisms to meet the requirements of the jurisdictions where recipients are located. This includes GDPR for European recipients, CAN SPAM for US recipients, and other regional frameworks as needed. We maintain a global suppression list, process data subject access requests, and document the lawful basis for processing for each campaign. We are not a law firm and we do not provide legal advice. We recommend that each client have their own counsel review the programme's compliance posture. What we provide is the technical implementation: the suppression mechanisms, the data handling procedures, and the audit trails that demonstrate compliance in operation.
We strongly advise against it and will not do it as a standard practice. Using a company's primary domain for cold outreach risks the deliverability of the company's regular business email. If the cold programme generates spam complaints, the domain reputation suffers and internal email, customer communications, and transactional messages may begin landing in spam folders. We use separate sending domains specifically for outreach. If a client insists on using their primary domain, we will explain the risks in writing and require explicit acknowledgement of those risks before proceeding. In nearly every case, the client concludes that the risk is not worth taking.
Related
Next step
Cold email infrastructure is not a commodity. The difference between a rented seat on a shared platform and infrastructure you own is the difference between a programme that stops when someone else decides and a programme that stops when you decide. We built our own because we could not find a platform that gave private equity operating teams the control, the continuity, and the isolation they need to run sustained outbound across multiple portfolio companies. The cost is higher, the operational burden is real, and the setup takes longer. For the operating partner who needs outbound to work reliably over a multi year hold period, those are investments worth making. For the operating partner who needs a quick campaign on a small list, we are probably not the right answer, and we will say so.
Start a conversation