What a Cold Email Agency Actually Does

Who's answering this, and the conflict I should flag

I'm Hlib Storchak. I build outbound systems for B2B founders and sales teams, and most of what follows comes from running this for clients, with 2000+ meetings booked for B2B clients along the way. Worth saying plainly up front: I compete with the agencies referenced in this article. Some of my clients came to me after leaving an agency relationship, a few of them specifically because of the issues this piece walks through. I'm not a neutral party, so treat this as an honest operator's read rather than a scored review of anyone by name.

What I can offer instead is a straight answer to a question most agency sales pages dodge: what actually happens, phase by phase, once you sign, what it should cost, and where the real risk sits.

Phase 1: ICP and list building

Every legitimate engagement starts here, and it's the phase that determines whether everything after it works. A real agency defines your ideal customer profile with you (industry, company size, tech stack, buying triggers), then builds a tiered account list against it, enriches contacts, and verifies email addresses before a single message goes out. This is also where the first red flag can show up: an agency that skips straight to "send volume" without a documented ICP is optimizing for activity, not replies.

A working list from this phase should come with a documented tiering logic (why account A is tier 1 and account B is tier 3) and a verification pass, not just a raw export from a data provider. If an agency can't explain how a given account landed on your list, that's worth asking about before the list goes into a sequence.

Phase 2: copywriting and sequence design

This is the phase clients notice most and the one that's easiest to judge on taste alone, which is a mistake. A sequence is usually four to six touches across two to three weeks, mixing a cold open, a value-add follow-up, a social proof or case-study touch, and a breakup email. What matters more than any single line is whether the copy is written against the ICP tiering from phase 1, a tier-1 enterprise account and a tier-3 SMB account shouldn't get the same email with a different logo swapped in.

Phase 3: sending infrastructure and domain setup

This is the phase with the most operational detail and the most room for an agency to cut corners you'll never see. It covers buying and configuring sending domains separate from your main company domain, setting up mailboxes, running a warmup period before real volume, and configuring SPF, DKIM, and DMARC. For sending and sequencing, this is the stack I run for clients: Salesforge for sending, Mailforge and Infraforge for domains and mailboxes, Warmforge for warmup. That's a personal preference built on what I've run for clients, not a verdict, Instantly and Smartlead both do this job well too and are worth evaluating on their own merits for your specific setup.

The part that matters more than which platform an agency uses is a question almost nobody asks at the sales stage: whose name is actually on the domain registration and the mailbox accounts. That question sets up the next section.

Phase 4: deliverability management

Deliverability stopped being a nice-to-have in 2026. Per EasyDMARC's own tracking of the rollout, Google and Yahoo began enforcing DMARC for bulk senders in February 2024, and Microsoft set its own enforcement deadline for May 5, 2025, with non-compliant mail now rejected outright rather than just filtered to spam. An agency that isn't actively managing authentication and sender reputation against those standards is running your outbound on borrowed time.

Leadium publishes its own internal deliverability standard, what it calls its underwrite, on its site: SPF, DKIM, and DMARC set to enforcement (not the weaker "p=none" monitoring-only mode), separate authenticated sending domains, a two to three week warmup period before volume, spam complaint rate kept under 0.10% as tracked in Google Postmaster Tools, one-click unsubscribe on every send, and a kill-switch that pauses sending automatically when reputation starts to slip. Whether or not you work with Leadium specifically, that's a genuinely useful checklist to hold any agency's pitch against, since it's more concrete than most vendors publish about their own process.

Phase 5: reporting and optimization

The last recurring phase is the one that tells you whether phases 1 through 4 are actually working: weekly or biweekly reporting on sends, opens (directionally, given how unreliable open tracking has become), replies, positive replies specifically, and meetings booked, followed by A/B testing on subject lines, opening lines, and CTAs. A report that only shows "replies" without separating positive replies from unsubscribes and out-of-office bounces is a metric built to look better than the underlying result.

What it actually costs

Per BuzzLead's 2026 pricing breakdown, cold email agency retainers cluster into four rough bands, plus a separate performance-based model:

The same BuzzLead breakdown lists pay-per-meeting pricing separately, roughly $200 to $600 per qualified meeting booked, or a hybrid of a smaller base ($500 to $1,500/month) plus $150 to $400 per meeting. Infrastructure is sometimes billed separately: domains at roughly $30 to $60 a year each, mailbox hosting around $6 to $12 per inbox per month, and warmup tooling around $30 to $100 a month. For a named agency's exact current rate, always check current pricing directly, published ranges shift and vary by contract length and volume.

Here's a cost model you can run with your own numbers, every input labeled as an assumption: 3-month cost = (monthly retainer × 3) + setup fee. Assume a $3,500/month mid-market retainer, a $1,500 one-time setup fee, and a realistic 4 to 6 week ramp before volume is meaningful. That's (3,500 × 3) + 1,500 = $12,000 for the first quarter. If that quarter realistically produces, say, 8 to 15 qualified meetings once ramp is accounted for (your own number will depend on ICP and list quality), that's roughly $800 to $1,500 per meeting in quarter one, higher than the $200 to $600 pay-per-meeting benchmark above, because the retainer model bills you through the ramp period and the performance model doesn't. That gap narrows fast in quarter two once the list and sequences are proven, which is exactly why comparing a single month's cost-per-meeting across pricing models without adjusting for ramp time is misleading in either direction.

The ownership question

Of everything in this article, this is the one worth reading twice. Per a 2026 outbound agency contract guide from Danish Lead Co, "contracts that retain agency ownership of critical assets like email domains, contact lists, and sending infrastructure are highly problematic." Some agencies set up domains, mailboxes, and CRM records under their own accounts by default, which is operationally convenient for them and genuinely risky for you: leave the engagement and you can lose the warmed inboxes, the verified list, and the sending reputation you paid to build, all at once.

The same guide's recommended language is worth copying into your own contract review: the client owns all domains, email accounts, and contact data generated during the engagement, transferable on termination, in writing, not implied. This is the exact clause I check first when I take over an account from a previous agency. The mistake I see most often isn't a bad sequence or a slow ramp, it's a client who never confirmed asset ownership and found out the real answer only when they tried to leave.

Other red flags worth checking

• Hidden costs. Undisclosed setup fees, data acquisition charges, or scope-creep clauses that let the retainer grow without a renegotiation.
• No performance language at all. A soft guarantee is normal. Zero accountability language anywhere in the contract is not.
• Reply rate without a positive-reply split. A 15% reply rate where most replies are "unsubscribe" or an out-of-office is a worse result than a 3% reply rate made of real conversations.
• Long lock-ins with no exit. Per the same Danish Lead Co guide, a 6 to 12 month minimum term with no performance-based exit is treated as a major red flag; a 3-month initial term moving to month-to-month with 30 days' notice is the more buyer-favorable structure.
• Bought or scraped lists. Violates most inbox providers' bulk-sender guidelines and is the fastest way to burn a domain's reputation before the engagement even ramps.
• Week-one result promises. Real infrastructure needs two to three weeks of warmup alone. Anyone promising volume and results in week one is skipping a step you'll pay for later.

Questions to ask before you sign

1. Who owns the domains, mailboxes, and contact list once the engagement ends, and is that in the contract, not just the sales deck?
2. What's the actual notice period and minimum term, in days or months, not "flexible" as a description?
3. What DMARC policy do you run on sending domains: monitoring-only ("p=none") or enforcement ("p=quarantine" or "p=reject")?
4. Will reporting separate positive replies from total replies, and can I see a sample report before signing?
5. Is there a kill-switch or equivalent that pauses sending automatically if sender reputation drops?
6. What happens to setup fees if I leave before a certain date?

Agency vs fractional GTM vs in-house

An agency buys you speed and a built process on day one, at the cost of less control and the ownership questions above. Building in-house buys you full control and a system you keep permanently, at the cost of a slower ramp and the real hiring risk of a first SDR or GTM hire. A fractional Head of GTM sits between the two: you get a senior operator setting up and running the same process an agency would, but reporting to you directly and building toward a system your own team can eventually run, rather than one the agency keeps.

The deciding factor is usually less about budget and more about time horizon. If you need pipeline in the next 60 days and don't yet have anyone who can own outbound internally, an agency is the fastest path. If you're building a durable GTM motion you'll run for years, in-house or fractional-to-in-house tends to pay off the ownership and cost tradeoffs an agency doesn't solve.

Which model I actually reach for, and when

For a first attempt at outbound with no existing process, I usually recommend done-for-you: get a working engine live fast, with clean ownership terms, before investing in a hire. Once the ICP, offer, and sequences are proven and repeatable, that's when I'd shift a client toward either an in-house hire or a fractional GTM setup, so the system you're now paying to run also becomes something you own outright. The wrong move I see most often is skipping straight to a full-time SDR hire before anything has been validated, which turns a hiring mistake into a much more expensive one than an agency retainer would have been.

FAQ