How Many Mailboxes Does an Agency Need for Cold Email?

Most agencies decide how many mailboxes to buy by copying a number from a forum thread, and then spend the next quarter either burning domains or sitting on idle inboxes. The honest answer to how many mailboxes an agency needs starts at the other end. You pick the number of booked calls you want each month, work backwards through reply rates and sequence length, and the inbox count falls out of the arithmetic. Done that way, a typical web agency needs somewhere between 4 and 10 mailboxes, not the 50 that tool vendors like to show in screenshots.

The rest of this post walks through that calculation with real numbers you can swap for your own. It covers the daily send ceiling per inbox, how many inboxes one domain can carry, what to keep in reserve, and the signs that tell you the setup is too small or too big. None of it needs a spreadsheet more complicated than four multiplications.

Start from the calls you want, not the emails you can send

Sending capacity is the wrong starting point because it is almost unlimited. Any agency can buy 40 inboxes this afternoon, and the only thing that number tells you is how fast you can exhaust a lead list. The figure that pays the bills is booked discovery calls, so the calculation starts there and moves backwards one step at a time.

Take an agency that wants 10 booked calls a month from cold email. That is a modest target for a two-person studio and a realistic one for outreach that references each prospect's actual website. Every number below is a reasonable middle estimate rather than a best case, because planning on best cases is how inboxes get pushed past their limits in week three.

From booked calls back to positive replies

Not every interested reply becomes a call. Some people ask for a price by email and go quiet, some want to talk next quarter, and some were being polite. Across agency outreach, roughly half of positive replies turn into a booked call when someone answers them within a few hours. So 10 calls needs about 20 positive replies.

Speed matters here more than people expect. A reply answered the same morning books at close to that 50 percent rate, while one answered two days later often drops below 30 percent. If your team cannot answer replies quickly, plan on 30 positive replies for the same 10 calls and size everything up accordingly.

From positive replies back to leads contacted

A positive reply rate of 1.5 percent of leads contacted is a fair planning figure for well-targeted, personalised outreach to small businesses. Generic sends often sit closer to 0.5 percent, and outreach built around a real finding on the prospect's site can reach 3 percent or more. At 1.5 percent, 20 positive replies means contacting about 1,350 leads a month.

This is the number that most changes the answer. Doubling the reply rate halves every figure below it, including the mailbox count and the domain spend. That is why the time spent on targeting and on what the email actually says pays back more than any infrastructure decision in this post.

Turning leads into daily sends per inbox

Leads are not the same as emails. A sensible sequence sends three messages to each lead over about two weeks: the first email, a short follow-up four days later, and a final note a week after that. People who reply drop out, so the real average lands a little under three. Planning on 2.8 emails per lead gives about 3,800 sends a month for 1,350 leads.

Spread that across roughly 21 working days and the daily total comes to about 180 emails. Weekend sending is possible, but small business owners read differently on Saturdays and reply rates usually drop, so most agencies keep the volume on weekdays. The daily figure is what the mailbox count has to support.

The daily limit that keeps an inbox healthy

A warmed Google Workspace or Microsoft 365 inbox can comfortably send 30 to 40 cold emails a day, on top of the warmup traffic it keeps running. Mailbox providers technically allow far more, but those limits are for normal business mail, not cold outreach to strangers. Pushing a single inbox to 100 a day works for a week or two and then the open rate quietly collapses.

Use 30 a day as the planning number. It leaves room for a bad week, a list with more bounces than expected or a spam complaint without the whole inbox sliding into the spam folder. At 180 sends a day and 30 per inbox, the agency in this example needs 6 active mailboxes.

Why a new inbox cannot start at that number

None of those six inboxes can send 30 cold emails on day one. A fresh mailbox needs two to four weeks of gradual ramp-up before it carries real volume, starting with a handful of sends and climbing a few each day. A safe warmup schedule lays out what that ramp looks like day by day.

The practical effect is that the mailbox count has to be decided a month before the volume is needed. An agency that wants to start full sending on the first of next month should be buying domains and creating inboxes now, not on the day the campaign is ready. Treat the month of warmup as part of the cost of the campaign.

How many domains those mailboxes need

Every cold mailbox should sit on a secondary domain rather than the agency's main one. If a campaign goes badly and the domain's reputation drops, the damage stays on a domain used only for outreach, and the client-facing inbox keeps reaching people. Secondary domains usually look like the main brand with a small change, such as a different ending or a short word added before the name.

The common guidance is two or three inboxes per domain. Fewer wastes money on domains, while more concentrates risk, because a reputation problem on one domain then takes out a large share of your sending in one go. For six mailboxes, three domains with two inboxes each is the cleanest layout.

Setting each domain up so it survives

Each domain needs SPF, DKIM and DMARC records in place before the first email goes out, and a simple redirect from the bare domain to the agency's real website. That redirect matters because some prospects will type the domain into a browser, and an empty parked page looks like spam. Domains cost around $12 a year each, so the outlay here is small compared with the cost of an inbox.

Leave new domains for a week or two before attaching inboxes and starting warmup. A domain registered yesterday and sending email today is one of the easiest patterns for spam filters to spot. The same patience applies when adding capacity later.

The spare capacity worth paying for

Plan for one extra domain with two warmed inboxes that are not part of the daily send. They keep their warmup running and wait. When one of the working domains starts showing low open rates or more bounces than usual, its inboxes can be rested and the spares take over without a two-week gap in sending.

For the example agency that brings the total to 8 mailboxes on 4 domains. That is the full answer to the question for a team targeting 10 calls a month: six working inboxes, two on standby, and roughly four weeks of setup before the first full day of sending. Domains and inboxes for that layout cost far less each month than a single signed project brings in.

Signs the setup is the wrong size

The numbers on paper are a starting point, and the inboxes will tell you within a month whether the estimate was right. Three signals are worth checking every week: open rate by inbox, bounce rate by domain, and how many sends each inbox actually completed against its daily cap. A spreadsheet with one row per mailbox is enough to track all three.

An undersized setup shows itself as inboxes hitting their cap every day while the lead list grows faster than it empties. The tempting fix is to raise the per-inbox limit from 30 to 50. That nearly always ends with falling open rates a fortnight later, so add an inbox and a domain instead and accept the warmup delay.

An oversized setup is quieter but still costs money. Inboxes sending 10 a day because the lead list cannot feed them are paying a monthly seat each to do warmup traffic and little else. If more than a third of the working inboxes sit well under their cap for a month, rest two of them and put the money into better leads.

The one sign that should stop sending entirely is a sudden drop in opens across every inbox on one domain at once. That usually points to a reputation problem on the domain rather than with any single mailbox. Pause that domain, bring in the spares, and check its DNS records and recent bounce rate before it sends again.

Adjusting the maths when the inputs change

The calculation is only as good as the numbers going into it, and three of them move a lot between agencies. Changing any one of them changes the mailbox count, often by more than people expect, so it is worth rerunning the arithmetic when the business changes rather than buying more inboxes by habit. Team size is one, the call target is another, and the reply rate is the one that moves most.

When the team or the target grows

Doubling the call target doubles everything downstream, so 20 calls a month at the same 1.5 percent rate needs about 12 working inboxes on 6 domains, plus spares. Adding a second person to answer replies does not change the mailbox count, but it does protect the 50 percent booking rate that the calculation depends on. Growth in target and growth in team should be planned together.

When a second seller joins, give each person their own set of domains rather than sharing inboxes. Replies stay in one place per person, the reputation of each set can be tracked separately, and a problem with one seller's list does not drag down the other's sending. It costs one extra domain at most, which is cheap insurance.

When the reply rate is the real problem

If outreach is running at 0.5 percent positive replies, the same 10 calls needs about 4,000 leads and roughly 18 working inboxes. That is three times the infrastructure, warmup time and lead spend to reach the same result. More mailboxes will hide the problem for a while, but they will not fix it, and they multiply the damage when the domains start to tire.

The cheaper fix is almost always the message. Outreach that opens with something specific about the prospect's own site, such as a 6-second mobile load time or a contact form that fails on phones, earns replies at a multiple of generic templates. Prospecting from website flaws covers how to find those details at scale, and a bulk outreach workflow shows how to keep them personal when the list grows.

Before adding inboxes, check how much the email itself could carry. Run three of your current prospects through the free website analyser and see whether the findings would give each email a real opening line. If they would, raising the reply rate from 1.5 to 3 percent will halve the number of mailboxes this calculation asks for, which beats buying more of them.