Why Most Cold Email Agencies Fail Their Clients
The cold email agency space has a reputation problem. Not because cold email does not work, but because most agencies delivering it do not understand what they are selling. They sell pipeline and deliver open rates. They promise meetings and produce spam complaints. They take fees for three months, produce nothing measurable, and leave the client with a damaged domain and a skepticism toward the channel that takes years to undo. If you are evaluating whether cold email fits your business in the first place, start there before hiring anyone.
We have seen this pattern hundreds of times. The clients who come to us with domain issues, burned infrastructure, or just zero results from a previous agency follow the same pattern almost every time. Here is what actually goes wrong.
Failure 1: Sending from the client's primary domain
This is the most common and most damaging mistake. An agency onboards a new client, skips the domain setup work, and starts sending from the client's primary business domain. The domain that receives inbound email. The domain tied to the company's Google Workspace. The domain that a spam complaint could blacklist.
Every professional cold email program should send from dedicated secondary domains. These are separate domains registered specifically for outbound use (for example, getcompanyname.com, outreach.companyname.com, or a slight brand variant). They are isolated from the primary domain so that any deliverability problem, any spam complaint, any blacklisting affects only the sending infrastructure and not the company's core communications.
Agencies that skip this step do so to save money and time. Setting up proper sending infrastructure takes 3 weeks minimum. Skipping it means faster onboarding, faster billing, and a faster path to disaster for the client.
Failure 2: No warmup, or fake warmup
A new domain has no sending history. Major email providers (Gmail, Outlook, Yahoo) use sending history as one of the primary signals for spam classification. A new domain that suddenly sends 500 emails on day one looks like a spam operation because spam operations typically use new domains and send at high volume immediately.
Proper warmup means running every domain and inbox through a 3-week warmup process before any live sending begins, using tools that simulate legitimate email engagement to build sending history with inbox providers.
What most agencies do instead: they skip warmup entirely, use an automated warmup service and count it as done after 7-10 days, or start client sending before the warmup completes. The result is deliverability failure within the first 3 weeks of the program, at which point the agency has already collected two months of fees.
Failure 3: Copy that reads like a template
Cold email copy is detectable. Experienced buyers can identify a mass cold email in the first two sentences. The tells are everywhere: generic openers that reference the company's website ("I came across your website and wanted to reach out..."), feature-first value propositions ("We help companies like yours improve their X by Y%..."), and social proof inserted before the ask is established.
Agencies that serve 50+ clients simultaneously cannot write genuinely personalized copy for each one. They use templates, swap in the company name, adjust two or three sentences, and call it done. The result is copy that gets deleted on sight.
High-performing cold email is specific. It references something real about the prospect's business, frames a problem they actually care about, and makes a narrow ask. This requires research, iteration based on reply data, and genuine expertise in the client's vertical. Most agencies have none of the three.
Failure 4: Treating lead lists as disposable
Lead quality directly controls deliverability. Sending to invalid email addresses, to contacts who have changed roles, or to lists that have not been verified recently generates bounces and spam complaints. Bounce rates above 3-5% are a serious deliverability signal. Gmail's warning threshold sits at 0.1%; sustained rates above 0.3% trigger filtering.
The cheapest agencies buy bulk lead lists from data brokers, do minimal verification, and send. The contacts may be 12-18 months stale. The bounce rate is high. The domain burns within 6-8 weeks.
Clique refreshes client lead lists every 3 days, stale data is one of the top reasons cold email programs underperform. All contacts are verified before entering sequences. This is not optional housekeeping; it is the primary mechanism for protecting inbox placement over the life of a campaign.
Failure 5: Measuring opens instead of pipeline
Open rates became unreliable as a performance metric after Apple's Mail Privacy Protection (MPP) launched with iOS 15 in 2021. MPP pre-loads email tracking pixels on behalf of users, creating artificial open inflation that can make a campaign appear to have a 60-80% open rate whether the email was actually read or not.
Agencies that report open rates as their primary success metric are either not aware of this (bad) or are using it to mask low reply rates (worse). The only metrics that matter in cold email are positive reply rate and pipeline generated. If an agency cannot show you both of these numbers clearly, you do not have enough information to evaluate whether the program is working. See the benchmark data to understand what a well-run program should actually produce.
Failure 6: No vertical specialization
Cold email copy and positioning are heavily vertical-dependent. What works for a SaaS company targeting VP of Engineering is different from what works for a staffing firm targeting VP of HR, which is different again from what works for a consulting firm targeting CFOs. The pain points, the language, the proof formats, the appropriate length and tone all differ.
Agencies that claim to serve any vertical are generalist shops. Their copy is generic because their playbooks are generic. Agencies with genuine vertical specialization have run dozens of programs in a specific industry, know which objections come up on reply calls, have tested dozens of subject lines and openers, and know which proof points move the needle. The reply rate gap between a generalist agency and a vertical specialist is typically 1-3 percentage points, which at scale is the difference between a dead program and a full pipeline.
What a serious agency actually looks like
Before signing with any cold email agency, ask these questions directly and expect specific answers:
- What is your average positive reply rate across clients in my vertical?
- Do you send from our primary domain or dedicated secondary domains?
- How long is your warmup period and how do you monitor it?
- How often are lead lists refreshed and verified?
- What deliverability monitoring do you do on active programs?
- What metrics do you report on monthly, and do we own the infrastructure if we leave?
Vague answers to any of these are a signal. A good agency has clear, specific answers to all of them because they run clean programs and can point to the data.
Quick answers
Ask for their average positive reply rate (not open rate) across clients in your vertical. Ask whether they send from your primary domain or dedicated secondary domains. Ask what their warmup protocol is. Agencies that cannot answer these specifically are generalist operations with minimal infrastructure expertise.
Positive reply rate (replies expressing interest, not all replies), meetings booked, and pipeline value generated. Open rates are not a reliable metric since Apple MPP inflates them artificially. Bounce rate and spam complaint rate should also be tracked as deliverability health indicators.
Start by asking for the deliverability data: bounce rate, spam complaint rate, and which inbox folders the emails are landing in (seed inbox data). If deliverability is healthy and replies are still low, the problem is copy or targeting. If deliverability is poor, the infrastructure was not set up correctly. In either case, ask the agency for a specific diagnosis and a concrete fix timeline. If they cannot provide one, that is your answer.