Cold email deliverability agency: 9 things that separate real infrastructure from rented reputation
A cold email deliverability agency lives or dies by bounce rate and reply rate, not open rates. Here's what to audit before you sign anything.
Cold email deliverability agency: 9 things that separate real infrastructure from rented reputation
Pick the wrong cold email deliverability agency and you will spend 60 to 90 days burning through sender reputation before a single qualified reply lands. The metric that tells you whether an agency actually knows what it's doing is not open rate, it's positive reply rate, and any agency that can't quote you theirs across clients should be disqualified immediately.
Why deliverability is the job, not a feature
Most cold email agencies treat deliverability as a setup checkbox: buy domains, warm them up, connect to the sending tool, done. It isn't. Deliverability is an ongoing operating task. Sender reputation decays under volume. Google and Microsoft update filtering heuristics constantly. A domain that was landing in the inbox in March can be sitting in spam by May if nobody is watching bounce rate and spam placement week over week.
We run inbox placement tests on every active sending domain across our clients, typically weekly during ramp and bi-weekly at steady state. If placement drops below 85% on Gmail or 80% on Outlook, we investigate before we send another batch. That discipline is the difference between a program that compounds and one that craters.
The 9 things worth auditing before you hire anyone
1. Do they track bounce rate per domain, not per campaign?
A hard bounce rate above 2% on a single sending domain is a deliverability red flag. The mistake most agencies make is averaging bounce rate across all their sending infrastructure, which masks a single rogue domain dragging everything down. Ask specifically: what is the per-domain bounce threshold, and what happens when a domain crosses it? If the answer is vague, that's your answer.
2. Do they use open rate as a health signal?
If yes, walk away. Apple's Mail Privacy Protection, launched in 2021, prefetches tracking pixels on iOS and macOS devices regardless of whether the email was ever opened. Open-rate data is systematically inflated and tells you nothing reliable about whether your email reached a real inbox. Any cold email deliverability agency still citing open rates as a performance metric either hasn't updated their thinking since 2020 or is using vanity numbers to obscure poor reply performance. We don't track open rates for this reason. The real signals are positive reply rate and bounce rate.
3. What's their positive reply rate benchmark across programs?
Across the programs we run, a healthy positive reply rate sits between 1% and 4% of contacted accounts, depending on audience specificity, offer clarity, and sequence length. A 3-step sequence targeting a tightly defined ICP with a sharp offer can hit 3 to 4%. A broader spray-and-pray list rarely cracks 0.5%. Ask any agency you're evaluating for this number. Not open rate, not click rate. Positive reply rate. If they don't track it or can't tell you, they are not optimizing the thing that pays you.
4. How many sending domains do they provision per client?
The standard operating model for a properly run cold email program is one root domain per brand, then multiple subdomain or secondary domain variants used for actual sending. We typically provision 3 to 5 sending domains per client depending on monthly send volume, so no single domain carries more than 40 to 50 emails per day. Volume beyond that on a single domain accelerates reputation decay. Any agency sending 500 emails per day off one domain is cutting a corner that will eventually cost you.
5. What's their warm-up protocol and how long does it run?
New domains need 4 to 6 weeks of warm-up before they can carry real outbound volume. During that window, automated warm-up tools send low volumes of emails between seeded inboxes to build sending history. The tradeoff: warm-up tools like Mailreach or Instantly's warm-up network help, but they are not a substitute for real engagement signals. A domain with 500 warm-up interactions and zero real replies looks different to Gmail's filters than a domain with 50 warm-up interactions and 20 real replies. Good agencies layer both.
6. Do they strip tracking links from outbound emails?
Tracking links, the kind that route clicks through a redirect domain to measure CTR, are a deliverability liability. Spam filters flag redirect domains. We strip click-tracking from the majority of client emails, which means we don't report CTR. This is the right call. A click on a tracked link tells you almost nothing useful compared to a prospect replying to book a call, and it adds measurable spam placement risk. If an agency insists on tracking links in outbound, ask them to show you inbox placement scores before and after enabling them.
7. Is there a dedicated team, not just a login?
Several agencies in this space hand you a Smartlead or Instantly login and call it service. You get access to a dashboard, maybe a template library, and a 30-minute onboarding call. What you don't get is someone watching your sending domains every Tuesday morning. Deliverability management requires a human being checking placement tests, reviewing bounce patterns, and making judgment calls about when to rotate infrastructure. A login is a tool. A team that operates the tool is the service. Ask exactly who will be monitoring your sending health and at what cadence.
For a European print-on-demand marketplace, we run a US-targeted outbound program managing 4 sending domains with weekly placement tests, which has kept bounce rate below 1.2% across 14 consecutive months. That consistency doesn't come from a tool. It comes from someone checking it.
8. Every part of the email is a test, not a guess
Subject lines, opening sentences, CTAs, sequence length, send days, send times: every variable is a hypothesis. The agencies that compound results over time run structured A/B tests on one variable at a time rather than rewriting the entire sequence when replies drop. The practical version looks like this: hold sequence structure constant, test two subject line variants over 500 sends each, measure positive reply rate, keep the winner, move to the next variable. That's it. No magic. On an NYC growth-equity firm's outbound program, two rounds of subject-line iteration lifted positive reply rate from 0.9% to 2.4% over six weeks, with no changes to the list or the offer.
9. Can they show you spam placement data, not just "we use warmup tools"?
Inbox placement testing means running test emails through tools like GlockApps or Maildoso to see what percentage land in inbox vs. spam vs. promotions across major providers. It's the closest thing to ground truth on deliverability health. Ask any agency you're evaluating to show you a sample placement report from an active client program. If they can't produce one, they're not running them. If placement tests aren't part of the workflow, they are flying blind.
What cold email deliverability agency pricing actually looks like
Most agencies in this space charge between $3,000 and $8,000 per month on a retainer model. That range covers everything from low-touch list-and-send operations to full-stack programs with dedicated copywriters, domain monitoring, and weekly optimization cycles. Some agencies add a setup fee of $1,500 to $3,000 on top for the first 6 weeks of infrastructure buildout and warm-up. Performance-based models exist but are rare, usually structured as a base retainer plus a per-meeting fee, typically $150 to $400 per booked call depending on the ICP.
The cost you don't see on any agency's pricing page is what a deliverability failure mid-campaign actually costs: burned domains, a month of stalled pipeline, and the 6 to 8 weeks it takes to rebuild sender reputation from scratch. That's the real reason to pressure-test the infrastructure questions above before signing.
If you're evaluating the full landscape of B2B outbound options, the B2B cold email agency pillar covers how to structure the brief and what to expect across different program types.
When deliverability alone isn't the problem
Here's where most deliverability conversations go wrong: the agency fixes the infrastructure, bounce rate sits at 0.8%, placement tests come back green, and replies still don't come. That means deliverability was never the constraint. The problem is offer, targeting, or copy.
Deliverability gets your email into the inbox. Copy and offer get a reply. Both have to work. An agency that only sells deliverability and hands you a template to fill in yourself has solved half the problem and charged you for the whole thing.
The programs that actually generate pipeline combine clean infrastructure with sharp ICP definition and sequences rewritten based on reply data, not gut feel. For European companies running outbound into the US market, there are additional layers around sender timezone, cultural tone, and legal compliance under CAN-SPAM that compound the operational complexity. If that's your situation, the European cold email agency guide covers the specifics.
If you're running a B2B ecommerce model and want to push wholesale buyers to a webshop using cold outbound, the deliverability requirements are identical but the offer structure is different. A US promotional products brand we work with sends targeted discount codes via cold email to B2B buyers, driving them to a dedicated webshop landing page. The same bounce-rate discipline applies, but the reply mechanic is a click-to-purchase rather than a booked call, which changes how you measure success. More on that model at B2B ecommerce cold email.
If you want to understand how deliverability fits into the full program, the outbound lead generation agency pillar has that context.
Most founders we talk to at this stage have one of two problems: either they've had a bad experience with an agency that burned their domains, or they're starting from scratch and don't want to make that mistake. If that's where you are, book a discovery call and we'll go through your current setup or brief together before you commit to anything.
A quick decision framework for evaluating any cold email deliverability agency
Use this as your filter before any final conversation:
Ask for the positive reply rate across their last 3 active programs. A real number, not a case study headline. Below 1% on a targeted B2B program means something is broken.
Ask what they do when a sending domain's bounce rate crosses 2%. If the answer doesn't include domain rotation and a root cause review, that's a gap.
Ask to see a sample inbox placement test report. If they can't produce one, they aren't running them.
Ask who specifically manages deliverability monitoring and how often. A name and a cadence, not a job title.
Ask whether they track open rates as a performance metric. The correct answer is no, and here's why.
Five questions. If an agency stumbles on more than one, the infrastructure conversation isn't going to get better once you're a paying client.
FAQ
What does a cold email deliverability agency actually do?
It manages the sending infrastructure, domain health, bounce rates, and inbox placement that determine whether your outbound emails reach the inbox at all. This includes domain provisioning, DNS configuration, warm-up protocols, ongoing placement testing, and rotating infrastructure when health signals degrade.
How is deliverability different from copy or targeting?
Deliverability determines whether the email arrives. Copy and targeting determine whether it gets a reply. Both matter. Fixing deliverability without fixing offer or list quality will produce cleaner bounce rates and still zero replies.
Why don't you track open rates?
Apple Mail Privacy Protection, live since 2021, fires tracking pixels on open regardless of whether the email was actually opened. Open-rate data is inflated and unreliable as a result. We track positive reply rate and bounce rate instead, which are actual signals of real engagement and deliverability health.
How long before a new domain is ready to send at volume?
4 to 6 weeks of warm-up before carrying real outbound volume. Rushing this is one of the most common ways a cold email program starts dead.
What's a healthy bounce rate for cold email outbound?
Below 2% per sending domain is the standard threshold. We target below 1.5% in practice. Above 2% and you are actively damaging sender reputation. Above 5% and most sending platforms will pause your account automatically.
What's a realistic positive reply rate for a B2B cold email program?
1 to 4% of contacted accounts, depending on ICP specificity and offer clarity. A tightly scoped audience with a direct offer lands at the higher end. A broad list with a generic pitch rarely clears 0.5%.
The bottom line on choosing a cold email deliverability agency
The infrastructure question and the performance question are not separate. An agency that can tell you its clients' positive reply rates, show you a placement test report, and name the person monitoring your sending domains every week is doing the job. One that leads with open rates and a dashboard login is selling you a tool dressed up as a service.
If you want to audit your current sending setup or start a program built to run clean from day one, book a discovery call and we'll tell you exactly what we'd change and why, based on what we're seeing across the programs we run right now.
