Mailgun vs a Dedicated Cold Email Platform: Which One Fits B2B Prospecting
Every quarter some team discovers that Mailgun sends email for a fraction of what outreach tools charge and decides to run cold campaigns through it. Three weeks later they are debugging spam folder placement and a suspended account. The comparison is worth making properly: an SMTP relay and a cold email platform are built around opposite assumptions about who you are emailing and why.
- Mailgun is transactional infrastructure: it assumes recipients expect your mail. Cold outreach violates that assumption by definition, and Mailgun's acceptable use policy restricts unsolicited sending.
- Raw send capacity is the cheapest part of outreach; the expensive parts are warmup, inbox rotation, sequencing, reply detection and list hygiene — none of which a relay provides.
- Cold B2B volume done right is small: hundreds of personalized emails to named decision-makers, not tens of thousands of blasts. That workload rewards a platform, not an API.
- A realistic stack comparison is Mailgun plus five tools plus engineering time versus one outreach platform — the API is rarely cheaper once you price the whole pipeline.
- Use both where each belongs: relay for receipts, alerts and product email; a dedicated cold email platform for prospecting.
What Mailgun is actually built for
Mailgun is an email API: you POST a message, it handles SMTP delivery, bounces, webhooks and logs. It is excellent at what it was designed for — transactional and application email. Password resets, invoices, shipping notifications, product digests to your own signed-up users. In that world the recipient expects the message, engagement is high, complaint rates are near zero, and the main engineering problems are throughput, latency and template management.
That design shows up in the product's assumptions. Sending reputation is tied to your account and domain and is expected to be uniformly good. Lists are expected to be opt-in. There is no concept of a multi-step sequence that stops when a human replies, because transactional email does not have conversations. There is no warmup scheduling, because a transactional stream ramps naturally with your user base. And the terms of service reflect it: unsolicited bulk email is against Mailgun's acceptable use policy, and accounts that generate complaint spikes get throttled or suspended — correctly, from Mailgun's point of view, because one bad tenant can poison shared IP ranges for everyone.
None of this makes Mailgun a bad product. It makes it the wrong layer of the stack for prospecting. Judging Mailgun for cold email is like judging a freight rail line for door-to-door courier work: impressive tonnage, wrong problem.
What cold B2B outreach actually requires
Address-based B2B outreach is a different discipline. You are writing to specific decision-makers at specific legal entities who have never heard of you. Volumes are deliberately small — a serious campaign is often 300–2,000 contacts, not 50,000 — and every message is personalized enough to survive scrutiny by a busy executive. The goal is a reply and a meeting, not a click.
That workflow has a hard requirements list. Dedicated sending domains, separate from your corporate domain, so prospecting risk never touches invoices and support mail. Gradual warmup of every new domain and mailbox over weeks before real volume. Sending limits per mailbox per day — healthy cold sending is measured in dozens per inbox, not thousands. Multi-step sequences with delays, where a reply, a bounce or an unsubscribe immediately halts follow-ups. Reply detection and classification, because the entire economic output of the campaign arrives as inbound email. Suppression lists, bounce-driven list hygiene, per-segment analytics down to reply rate and meetings booked.
Notice how little of that list is about sending. Delivery is maybe a fifth of the problem. The rest is orchestration and conversation handling — exactly the layer an SMTP relay deliberately does not have an opinion about.
The deliverability model: shared expectations vs earned reputation
Deliverability is where the two approaches diverge hardest. On a relay, your mail inherits the expectations of transactional traffic: mailbox providers see a stream that normally gets opened, clicked and never flagged. Inject cold email into that stream and the metrics move the wrong way fast — lower engagement, more complaints, unfamiliar recipient domains. The relay's automated protection reacts, and either your throughput collapses or your account does. Even before suspension, your placement erodes silently: cold mail from transactional infrastructure carries technical fingerprints (headers, return paths, sending patterns) that filters have long since learned.
A cold email platform builds reputation the opposite way: slowly, per domain, per mailbox. New identities send small friendly volumes first, ramp over weeks, and stay under conservative daily caps permanently. Volume is spread across several mailboxes and domains so no single identity ever looks like a blast source, and so one damaged domain does not take the whole program down. Sending patterns imitate a human: staggered timing, business hours in the recipient's timezone, throttled bursts.
There is also a compliance angle worth naming. The point of this machinery is not to trick spam filters into accepting junk — it is to let a legitimate, personalized business letter arrive looking like what it is. If your emails are irrelevant blasts, no infrastructure will save the campaign, and under regimes like GDPR or CAN-SPAM you carry legal duties (honest headers, identification, working opt-out, a defensible basis for processing contact data) regardless of which tool sends the message.
Sequencing and reply handling: where campaigns are actually won
In cold B2B, roughly half of all replies typically arrive on follow-ups rather than the first email, which makes sequencing logic core infrastructure rather than a nice-to-have. A platform runs the state machine for you: wait three business days, send follow-up two unless the prospect replied, bounced or opted out; stop the whole thread the moment a human answers. Build that on a raw API and you are writing and maintaining a scheduler, a reply detector, thread-safe state per contact, and edge-case handling for out-of-office autoresponders — a real engineering project that has nothing to do with your product.
Reply handling is the deeper gap. A relay can tell you a message was delivered; it has no idea the prospect answered, because replies land in a mailbox the relay does not read. An outreach platform ingests the inbox, matches replies to campaigns and contacts, separates interested answers from unsubscribes and autoresponders, and pushes the conversation into a CRM where a salesperson picks it up with full context. For an address-based B2B campaign this is the entire point: the deliverable is a conversation with a decision-maker, and the platform's job is to make sure no conversation gets dropped.
A useful benchmark for what all this machinery buys: well-targeted cold B2B campaigns with proper sequencing tend to land in the 3–8% reply range, while one-shot blasts through generic infrastructure commonly sit below 1%. The delta is not copywriting magic — it is follow-ups arriving reliably in the primary inbox and replies being caught and worked.
The real cost comparison: API plus everything else
The pricing argument for Mailgun evaporates once you price the whole pipeline. The relay fee covers delivery only. To run actual cold campaigns on top of it you still need: sequencing logic (built or bought), warmup tooling for every new mailbox, a reply-detection and CRM-sync layer, list verification, suppression management, per-domain deliverability monitoring, and an engineer who owns all of it. Each piece exists as a separate SaaS, or as internal code someone must maintain at 2 a.m. when follow-ups start double-sending.
A dedicated platform bundles that stack into one subscription and one accountable vendor. For a typical B2B team sending a few thousand personalized emails a month, the bundled option is usually cheaper in total cost and dramatically cheaper in attention — and attention is the scarce resource in an SDR team. The DIY route on a raw API only starts making financial sense at engineering-heavy scale, and even then most teams route around relay restrictions by self-hosting SMTP, which is a bigger project still.
There is also the risk line item people forget: an account suspension mid-campaign is not an inconvenience, it is the campaign. Sequences freeze mid-thread, follow-ups never go out, and the prospects who were about to reply go cold. Cheap infrastructure with a nonzero probability of stopping your revenue motion for a week is not cheap.
When Mailgun is the right answer — and how the stacks coexist
Keep the relay for what it is superb at. Transactional email: signups, resets, receipts, alerts. Product notifications and digests to your own user base. High-volume programmatic sending where recipients opted in. If your outreach motion is actually marketing to an existing opt-in list, an ESP or a relay-based setup is legitimately the right tool — that is a different discipline from cold prospecting and follows different rules.
The mature setup most B2B companies converge on is a split stack: corporate domain and relay for operational mail, a cold email platform with its own dedicated domains for prospecting, and a CRM as the meeting point where replies become deals. The two send streams never share a domain or an IP identity, so a rough week in prospecting can never bounce your invoices.
Decision rule in one paragraph: if the recipient asked for the email, use the relay. If the recipient has never heard of you and you want them to take a meeting, use a platform built for cold outreach — with warmup, rotation, sequencing and reply handling as first-class features, and small personalized volumes as the operating philosophy.
Quick sanity check: a 1,000-contact campaign at 30 emails per mailbox per day needs about 5 warmed mailboxes across 2–3 dedicated domains, a 3-step sequence over two weeks, and reply routing into your CRM. Count how many of those five components your SMTP relay provides: zero — it provides the SMTP.
FAQ
Can I use Mailgun for cold email at all?
Technically you can push any message through the API, but unsolicited outreach conflicts with Mailgun's acceptable use policy, and complaint-driven automation will throttle or suspend accounts that behave like cold senders. Even before enforcement, transactional infrastructure is a poor deliverability fit for cold traffic. Use it for the email your users expect, not for prospecting.
Is a cold email platform just Mailgun with extra features?
No — the sending philosophy is inverted. A relay optimizes for high-volume delivery of expected mail from one trusted identity; an outreach platform optimizes for small, personalized volumes from many warmed identities, with sequencing and reply capture around them. The features are not add-ons, they are the product.
What volumes are normal for cold B2B outreach?
Far lower than most people assume: commonly 20–50 emails per mailbox per day, scaled horizontally across several mailboxes and dedicated domains. A serious address-based campaign targets hundreds to a few thousand named decision-makers, personalized per segment. If your plan requires tens of thousands of identical sends, the plan is the problem, not the tooling.
Do I still need my own deliverability work if I use a platform?
Yes, but less of it. You still own list quality, relevance and legal compliance — honest identification, a working opt-out, a defensible basis for contacting people under regimes like GDPR or CAN-SPAM. The platform handles the mechanical layer: warmup, rotation, throttling, bounce processing and reputation monitoring.
When does building outreach tooling on a raw SMTP API make sense?
Almost only when outreach infrastructure is itself your product, or you have dedicated engineers and unusual requirements a platform cannot cover. For a normal B2B team, the build option means recreating warmup, sequencing, reply detection and monitoring — months of work to reach feature parity with off-the-shelf tools, plus permanent maintenance.
Can my transactional and cold email share a domain if the tools are separate?
Do not do this. Reputation is tracked per domain, so a prospecting misstep would drag down password resets and invoices. Run cold outreach from separate, dedicated domains (often variations of your brand name), keep the corporate domain for operational and one-to-one mail, and let the domains earn reputations independently.
Want to apply this to your outreach?
We will map it to your segment and product — before any work starts.
Talk to us