---
title: "How to Vet a Cold Email Agency: The Four Numbers to Demand Before You Sign | Stone Haven Capital Group"
description: "Most cold email agency hero stats carry no time period and no client count. Here are the four numbers to demand, and StoneHaven"
canonical: "https://stonehaven.capital/alternative/how-to-vet-a-cold-email-agency"
last-updated: "2026-08-30"
---

> Most cold email agency hero stats carry no time period and no client count. Here are the four numbers to demand, and StoneHaven

Stone Haven Capital Group

Schedule a Call

[All comparisons](https://stonehaven.capital/alternative/)

Comparison

# How to Vet a Cold Email Agency: The Four Numbers to Demand Before You Sign

August 30, 202615 min read

A cold email agency homepage in 2026 usually leads with one enormous figure. Leads generated, revenue influenced, pipeline built, meetings booked. Cleverly's site carries three of them at once: 224.7K leads, $51.2M in revenue, $312M in pipeline. None of the three states a time period, a client count, or a method. That number could describe one quarter or eight years, one account or two thousand.

Learning how to vet a cold email agency means refusing to argue with the hero number and asking for a different set instead. Four measurements decide whether an outbound program works, and every one of them is sitting on a dashboard the agency already looks at every morning. If a vendor cannot produce all four inside a day, the program has not been measured.

A B2B outbound program is judged on four measurements: inbox placement at a stated send volume, bounce rate at that volume, reply rate with its send count, and booked calls per week with the close rate behind them.

As of Q3 2026, StoneHaven's published programs answer with 98% inbox placement at 120,000 sends a month and bounce between 1% and 3%.

Reply rate ran 11.94% across 183,524 sends in a single 10-day run.

Booked calls run 70 to 80 a week at a 25 to 30% close rate.

## Why the standard cold email agency pitch survives vetting

A site-wide total works as marketing because it cannot be falsified. Nobody outside the company can check whether $312M in pipeline came from thirty clients or three hundred, because there is no denominator to divide by. The same holds for a "97% deliverability" badge with no volume attached, since almost any account hits 97% while it is sending forty warmup emails a day.

Deliverability is not hard at low volume. It gets hard somewhere between 20,000 and 100,000 sends a month, when a single reputation event stops being absorbed by the rest of the pool. Every number below is written to be asked at volume, because that is the only place the answer separates vendors. If you want the wider field first, StoneHaven's breakdown of [how the main cold email agencies stack up](https://stonehaven.capital/alternative/best-cold-email-agencies/) covers the models before the metrics.

Best for: B2B SaaS, services, cybersecurity, industrial and manufacturing firms with mid-five-figure-plus deal sizes selling into founders, VPs of Sales, CROs and enterprise program owners.

## Number one: inbox placement, with the send volume behind it

Inbox placement is the share of sent mail that reaches the inbox rather than spam or a quarantine folder. It is decided before the first send, by which provider the mailboxes sit on, whether the sending domains are isolated from the client's corporate domain, and whether [SPF](https://datatracker.ietf.org/doc/html/rfc7208), [DKIM](https://datatracker.ietf.org/doc/html/rfc6376) and [DMARC](https://datatracker.ietf.org/doc/html/rfc7489) are authenticated properly on every one of them.

The question to ask is not "what is your inbox placement". It is "what placement did you hold, at how many sends a month, across how many domains". A vendor sending 5,000 a month from six mailboxes is answering a different question from one sending 120,000 across 25 domains and 600 inboxes.

Ask a second question underneath it: how many provider types is the sending split across. Google reputation is domain-based and Outlook reputation is IP and tenant-based, so the two fail in different ways and at different times. A program that lives entirely inside one provider gives that provider a complete view of its spam complaint rate, and one bad week takes the whole engine down. Documented builds at 100,000+ sends a month run across Google Workspace, Microsoft and private SMTP so no single provider ever sees the true picture.

Across StoneHaven's published programs, isolated sending infrastructure held 98% inbox placement from the first send at 120,000 emails a month, spread over 25 dedicated domains and 600 inboxes and split across three provider types.

A good answer names the warmup period too. Google mailboxes need roughly 14 days before they carry cold volume, and Outlook flags any tenant pushing more than 2,000 a day. An agency promising live sends in week one is either using pre-warmed infrastructure it already owns or skipping the ramp, and those produce opposite outcomes. Ask which.

## Number two: bounce rate, at a stated monthly volume

Bounce is the cleanest proxy for list quality because it cannot be talked around. An unverified B2B list bounces at 15 to 20%. A properly verified one lands near 1%. There is no copywriting that closes that gap, and mailbox providers read a rising bounce rate as a signal that the sender does not know who they are mailing.

Two follow-ups separate an agency that verifies from one that says it does. First, how often is the list revalidated. Contact data decays continuously, so a list cleaned once at kickoff is a different list by month three; revalidation every 30 to 45 days is the working standard.

Second, how are catch-all domains handled. Catch-alls accept everything at the server and reveal nothing, they make up roughly 20 to 30% of a typical B2B list, and most verification tools guess on them. Microsoft accounts for the majority of B2B mailboxes and is where the guessing is worst. Honest catch-all verification needs live-send monitoring over about 48 hours, so any tool returning a verdict in seconds is estimating.

Across every campaign StoneHaven operates, bounce stays between 1% and 3% and unsubscribes between 0% and 0.24%.

One workspace sent 231,347 emails to 77,115 leads at a 0.97% bounce rate.

## Number three: reply rate, with its send count

Reply rate is replies divided by sends. It is the fastest signal that the offer and the list match, and it is the number most often quoted without the denominator that makes it mean anything. A 12% reply rate off 400 sends to a warm-adjacent list is noise. The same figure across 183,524 sends is an engine.

Industry benchmarks put average cold email reply rates at 3 to 6%, with strong operators reaching 6 to 10% in the right niche, and anything decaying under about 2% is a campaign that should be killed rather than nursed. Ask the agency what they consider a kill threshold, and whether they have ever hit it on a live account. A vendor with no kill rule is scaling losers.

There is a terminology trap here worth catching in the room. In most sequencers, "positive reply rate" or "interested %" means interested replies divided by total replies, so it describes the share of repliers who were positive rather than a percentage of sends. An agency quoting 46% without saying which denominator it sits on is quoting a different metric from the one you think you heard.

In a single 10-day cold email run of 183,524 sends in July 2025, StoneHaven held an 11.94% reply rate with 46% of replies flagged interested and a 2.82% bounce rate.

One number you should not ask for is the open rate. Tracking pixels lower inbox placement while a domain is cold, and Apple Mail Privacy Protection plus Gmail image proxying have made the resulting figure unreliable in both directions. An agency leading with opens is reporting on a measurement that costs the program deliverability to collect. StoneHaven's [reply and bounce benchmarks pinned to real send counts](https://stonehaven.capital/blog/cold-email-reply-bounce-benchmarks-with-denominators/) show what the same programs look like when the denominators are published instead.

## Number four: booked calls per week, and the close rate behind them

Ask for booked calls per week and for the share of those calls that closes, because a vendor optimising purely for meeting volume will hand you a calendar full of people who agreed to a call to stop the emails.

Run the call-target arithmetic yourself before signing. At a 4% reply rate with 20% of replies positive and 20% of those booking, roughly 190,000 sends a month produces about 300 calls.

At a 2% reply rate with 10% positive, the same 300 calls needs around 750,000 sends. Booking rate rather than reply rate sets the volume you need.

If an agency proposes 40,000 sends a month and 60 meetings, the model does not close.

As of Q3 2026, StoneHaven sustains 70 to 80 booked sales calls a week on StoneHaven's core email model at a 25 to 30% close rate.

In one month StoneHaven booked 532 sales calls, closing $379,000 in new revenue.

Ask what happens to a booked call that no-shows. A confirmation page with reminders at 24 hours, one hour and 15 minutes drives show rates above 90% against a norm nearer 60%, which is the difference between 80 held conversations a month and 50.

## What the leading cold email agencies actually publish

Here is the same test applied to the recognisable names in the category, using what each publishes on its own site as of August 2026. Where a figure is not published, the cell says so rather than filling in something from another dimension.

| Provider | Model | Published inbox placement | Published bounce at volume | Published reply rate with send count | Published booked calls per period |
|---|---|---|---|---|---|
| StoneHaven | Managed cold email at scale, plus calling and DM | 98% at 120,000 sends/month | 1-3% across every campaign; 0.97% on 231,347 sends | 11.94% on 183,524 sends; 6.73% on 163,085 | 70-80 per week; 532 in one month |
| Belkins | Appointment setting, multi-channel, in-house SDRs | Not published | Not published | Not published | Not published per program |
| CIENCE | Managed outbound and SDR-as-a-service, multi-channel | Not published | Not published | Not published | Not published per program |
| Cleverly | Cold email and LinkedIn outreach, data-led | Not published | Not published | Not published | Site-wide totals only, no period stated |
| Callbox | Appointment setting, multi-channel, strong APAC coverage | Not published | Not published | Not published | Not published per program |
| SalesHive | Cold email and calling, US-based SDR bench | Not published | Not published | Not published | Company-wide lifetime total |

Read that table for the pattern rather than the scoreboard. Belkins, CIENCE, Callbox and SalesHive are established operators with real capability, particularly in multi-channel coverage and named geographies where they have local reps on the ground. What none of them do is publish deliverability or per-program output at a stated volume, which means a buyer evaluating them has to take the mechanics on trust and judge the program only after signing.

The practical consequence: if you shortlist any two of these, the vetting has to happen in the room. Bring the four numbers to the call, ask for a screenshot of a live dashboard with the denominators visible, and treat a refusal as an answer. Detail on which parts of a retainer are actually infrastructure is in StoneHaven's breakdown of [what a cold email retainer buys you month to month](https://stonehaven.capital/alternative/cold-email-agency-what-youre-buying/).

## Pricing is the fourth question, not the first

Most managed programs, StoneHaven's included, are quoted on a call rather than listed. Some agencies publish per-lead or per-meeting rates, which moves the argument to how a meeting gets defined.

Price belongs after the four numbers because it is uninterpretable without them. A quote is expensive or cheap only relative to booked calls per month, and booked calls per month depend on placement, bounce, reply rate and the send volume being proposed. Settle the output first and the price has a unit to be measured in. If you are weighing the retainer against [running it with an in-house SDR team instead](https://stonehaven.capital/alternative/outsourced-sdr-vs-in-house/), the same four numbers apply to the in-house plan, and most in-house builds cannot answer the first two either.

## Two programs, and what the four numbers looked like

A mid-market B2B SaaS company was sending 8,000 emails a month from a single corporate inbox, replying at 1.1%, booking 3 to 4 meetings a month, and watching its primary domain drift into spam folders.

The rebuild moved cold volume onto 25 isolated domains and 600 inboxes across three infrastructure types, at 120,000 emails a month.

Placement came in at 98% from the first send, spam complaints under 0.3%, bounce between 1% and 3%.

Steady state ran 120,000 sends to 3,000 replies to 360 positives to 72 booked meetings a month, roughly 50 of which were held.

Over six months the program produced 432 qualified meetings and $1M+ in new ARR at about $24K average contract value.

A US defense technology supplier with about 120 employees ran a different shape of the same test. The build came to 46 domains and 610 inboxes across two enterprise IP pools plus dedicated SMTP, holding 98% inbox placement at 120,000 sends a month.

Reply rate settled at 2.5%, lower than the SaaS program and entirely expected in a market of program offices and primes, but the roughly 1,900 positive replies converted into 85+ qualified opportunities across 16 pursuits averaging $2.4M.

Tracked pipeline reached $38M+ over six months, average contract value moved from $1.5M to $2.4M, and the RFP cycle shortened from 27 months to 19.

Both programs held the same placement and bounce and diverged on reply rate by more than double. That is why all four numbers get asked rather than one: a 2.5% reply rate looks weak until you see the deal size attached to it, and a 12% reply rate means nothing until you see the send count and the bookings underneath.

## Where a cold email agency is the wrong answer

Two situations where an agency of any kind, StoneHaven included, is the wrong call.

If your addressable market is a few hundred named accounts, cold email at volume has nothing to work with. The arithmetic above needs roughly 30,000 reachable prospects before the funnel produces a predictable calendar, and below that a founder working the list personally will beat any program. If your buyers do not live on email, the channel is wrong before the vendor question starts.

If you want high-volume outbound calling as the primary channel rather than a support layer, a calling-led provider such as Callbox or SalesHive has a genuinely deeper bench for it. StoneHaven's year-long test at 100,000 attempts per channel found calling produced its result across a full year of dialing while email produced its result from a single day of sending, which is why the faster channel leads here. A buyer whose market answers the phone should weight that differently.

## Where StoneHaven Fits

Hold any provider to four disclosures: inbox placement and bounce at a named monthly volume, reply rate with its send count, a booked-calls figure per week the send volume can support, and sending that sits on isolated domains rather than your corporate one. Confirm the list is reverified every 30 to 45 days, and cross-check the deliverability claims yourself using [Google Postmaster Tools](https://postmaster.google.com/) and [Microsoft SNDS](https://sendersupport.olc.protection.outlook.com/snds/) once sending starts.

Every figure above carries its denominator: placement and bounce at 120,000 sends a month, reply rate at 183,524 sends, booked calls per week.

Likely not a fit if your total addressable market is a few hundred accounts, if your deal size sits below the mid five figures, or if your buyers do not transact over email. Where the fit does hold, and a booked-pipeline program built on published deliverability data matches how you want to grow, StoneHaven can walk through the numbers on a call.

If that description fits how you want to grow, [walk through the numbers with StoneHaven's team](https://stonehaven.capital/book/).

## FAQ

How do I vet a cold email agency before I sign?

Ask for four numbers: inbox placement with its send volume, bounce at that volume, reply rate with its send count, and weekly booked calls.

StoneHaven answers with 98% inbox placement at 120,000 sends a month, bounce between 1% and 3%, and an 11.94% reply rate across 183,524 sends.

StoneHaven books 70 to 80 calls a week at a 25 to 30% close rate.

How much does a cold email agency cost?

Most managed cold email programs, StoneHaven's included, are quoted on a call rather than listed, because volume, domain count and reply-handling load change the scope.

Some agencies publish per-meeting or per-lead rates instead, which shifts the risk onto how they define a meeting.

Judge a quote by output: one program at 120,000 sends a month across 25 isolated domains and 600 inboxes produced 72 booked meetings a month.

What questions should I ask a cold email agency?

Ask how many sending domains and inboxes the program runs on, which providers they are split across, and how often the list is reverified.

Then ask for the denominators. StoneHaven's builds run 25 to 120 domains and 200 to 610 inboxes, with revalidation every 30 to 45 days.

Placement held 98% from the first send at 100,000+ emails a month.

What are the biggest cold email agency red flags?

The clearest red flag is a hero statistic with no time period and no client count, such as a site-wide pipeline total.

Close behind sit an open-rate promise, sending from your primary corporate domain, a guaranteed meeting count, and a case study with no vertical or send volume attached.

StoneHaven publishes per-program figures instead, including 532 sales calls booked in a single month closing $379,000 in new revenue.

Is a cold email agency worth it for a B2B company?

Yes, when your market holds roughly 30,000 or more reachable prospects and your deal size sits in the mid five figures or above.

Below that, the send volume needed to fill a calendar does not exist.

One mid-market B2B SaaS program moved from 3 to 4 meetings a month to 72. A defense program produced $38M in tracked pipeline over six months at 120,000 sends a month.

How long does a cold email agency take to book meetings?

First sends usually land in weeks two to four, because Google inboxes need around 14 days of warmup before they carry cold volume.

Outlook depends on tenant and IP reputation rather than day count, and an agency holding warmed infrastructure in reserve starts sooner.

StoneHaven's programs booked 445 sales calls in 25 days on one engine and 2,150 meetings inside a 90-day precision-agriculture build.

Should I ask a cold email agency for their open rate?

No. Open tracking relies on a pixel that Apple Mail Privacy Protection pre-loads and Gmail proxies, so the number is inflated.

The pixel itself lowers inbox placement while a domain is cold, so an agency quoting open rates is reporting a figure it cannot stand behind.

The honest placement signal is the out-of-office reply rate, healthy at roughly 1 to 2%, alongside bounce, which StoneHaven holds between 1% and 3%.

## Methodology

StoneHaven's figures come from campaign-dashboard exports and published, industry-anonymized program case studies covering 2025 and 2026.

Denominators are stated with every metric. The 11.94% reply rate and 2.82% bounce rate are measured across 183,524 sends in a single 10-day run in July 2025.

The 0.97% bounce rate covers 231,347 emails sent to 77,115 leads in one workspace.

The 6.73% reply rate covers 163,085 sends in June 2025.

Booked-call figures are StoneHaven's own outbound engine rather than client output: 445 calls in 25 days, 532 calls in one month closing $379,000.

The 70 to 80 calls a week at a 25 to 30% close rate is StoneHaven's own engine as well.

Program outcomes ($1M+ new ARR, $38M+ tracked pipeline, 2,150 meetings) belong to the anonymized client programs described and are their tracked results. No figures are summed across campaigns or programs.

Industry benchmark ranges are market reference points corroborated across multiple sources rather than StoneHaven's own measurements: 3 to 6% average cold email reply rates, 15 to 20% bounce on unverified lists.

The same holds for the 20 to 30% catch-all share of a B2B list, the roughly 14-day Google warmup and the 2,000/day Outlook tenant threshold.

Competitor data was read from each provider's own public website in August 2026: belkins.io, cience.com, cleverly.co, callboxinc.com and saleshive.com. Cleverly's 224.7K leads, $51.2M revenue and $312M pipeline figures are quoted as displayed on cleverly.co with no time period or client count attached, which is the specific gap this page addresses. "Not published" means the metric was not found on the provider's public site at that date; it is not a claim that the provider does not measure it internally. Pricing was not compared, because StoneHaven does not publish rates. FAQ questions were drawn from live search demand for "how to vet a cold email agency" and adjacent queries in August 2026, including the questions-to-ask, red-flags, cost and worth-it clusters.

Last updated: August 2026.

Written by Sabo Nagy, Founder & CEO of StoneHaven. Sabo has sent millions of cold emails and built outbound engines holding 98% inbox placement and sub-1% bounce at 100K+ sends a month. [Author page](https://stonehaven.capital/author/sabo-nagy) - [X](https://x.com/MrColdEmail)

## See it working on your market.

Book a consultation call and we'll show you exactly how we'd build your pipeline - live in 72 hours.

Schedule a consultation call

## More comparisons

[Comparison](https://stonehaven.capital/alternative/callbox/)

### [Callbox Alternative: What $15,000 a Month Should Buy You in Reported Numbers](https://stonehaven.capital/alternative/callbox/)

[Callbox estimates $15,000 to $30,000 a month per Campaign Pod. StoneHaven answers the same question with 183,524 sends in 10 days at 11.94% reply and 2.82% bounce.](https://stonehaven.capital/alternative/callbox/)

[August 30, 2026](https://stonehaven.capital/alternative/callbox/)[12 min read](https://stonehaven.capital/alternative/callbox/)

[Comparison](https://stonehaven.capital/alternative/outsourced-sdr-services/)

### [Outsourced SDR Services Compared: Four Agencies, Ranked by What They Publish](https://stonehaven.capital/alternative/outsourced-sdr-services/)

[Callbox, SalesHive, Belkins and Martal publish meeting counts and ROI multiples. StoneHaven publishes 98% inbox placement and a 0.97% bounce rate across 231,347 sends.](https://stonehaven.capital/alternative/outsourced-sdr-services/)

[August 30, 2026](https://stonehaven.capital/alternative/outsourced-sdr-services/)[13 min read](https://stonehaven.capital/alternative/outsourced-sdr-services/)