Domain acquisition, renewals, records, authentication, redirects, and replacement inventory.
The accounts that create theoretical capacity, plus setup and account-management overhead.
Sending platform, enrichment, verification, list sourcing, suppression, and storage.
Health review, ramp management, reply routing, incident recovery, and operator time.
The easiest cold email budget to misunderstand is the one that multiplies mailbox count by mailbox price and stops.
That number is useful, but it is not the cost of the outbound system. It excludes the domains carrying those accounts, the sequencer sending campaigns, the data required to reach relevant people, verification losses, monitoring, replacements, and the human work that prevents the system from damaging itself.
First-party example: one Ink Persuasion planning model uses 249 Google Workspace mailboxes at an account cost of $3.50 per mailbox per month. That equals $871.50 per month for the mailbox-license layer only. It is not a universal Google list price and it does not include domains or the rest of the stack.
Start with a cost model, not a tool list
mailbox licenses + amortized domain cost + sequencer + lead sourcing and enrichment + email verification + warmup and monitoring + replacement reserve + operator and QA time = monthly outbound infrastructure cost
Some vendors bundle several lines. That changes the invoice, not the economics. Keep the layers visible so you can compare providers, identify waste, and understand what disappears if a subscription is removed.
The 249-mailbox worked example
249 x $3.50 = $871.50 mailbox cost/month 249 x 20 campaign emails/day = 4,980 theoretical campaign emails/day 4,980 x 20 sending days = 99,600 theoretical campaign emails/month $871.50 / 99,600 = $0.00875 mailbox-license cost per theoretical send
The denominator matters. If 15 percent of accounts are ramping, paused, broken, or held in reserve, do not divide by 99,600. Use the actual campaign sends that passed platform and delivery controls.
Why domains belong on a separate line
A domain is normally purchased or renewed on a different billing cycle from the mailboxes created under it. Domain cost should be amortized into the monthly model rather than folded invisibly into a mailbox estimate.
Also budget operational work around domains: DNS configuration, SPF and DKIM, DMARC rollout, redirects, tracking choices, ownership records, renewal control, monitoring, and replacement inventory. A domain that expires or loses its authentication state can interrupt many mailboxes at once.
Do not use one blended per-mailbox figure unless the vendor contract explicitly bundles and defines every domain cost. If it does, keep a separate internal count anyway so replacement and renewal exposure remain visible.
Use a conservative operating ceiling
Ink Persuasion's default operating ceiling is 20 campaign emails per mailbox per day. Long-established mailboxes with clean history can sometimes operate at 25 or 30 in our experience, but age alone does not make a mailbox safe. That range is an operator rule, not an endorsement or sending limit from Google.
Google's sender guidance emphasizes authentication, low spam rates, consistent sending, gradual volume increases, and monitoring delivery responses. It does not validate a universal cold-email threshold. Recipient systems, message quality, list quality, complaint behavior, domain reputation, and campaign patterns still matter.
The 30-day mailbox readiness plan explains why a mailbox should ramp gradually even after warmup. The mailbox recovery and retirement guide explains when preserving one account costs more than replacing it.
Model actual capacity, not maximum capacity
active mailboxes x approved daily campaign limit x sending days x availability rate = usable monthly campaign capacity
The availability rate should account for ramping accounts, temporary pauses, health incidents, authentication changes, provider throttling, manual holds, and reserve capacity. Use observed availability from the last month when possible.
Also distinguish campaign emails from warmup or other account activity. A dashboard may show total account traffic that is higher than the number of prospecting messages you intended to send.
Add the variable data costs
Mailbox and domain costs remain relatively predictable. Data costs move with list quality and campaign design:
- Company and contact sourcing
- Role, geography, and company filtering
- Email enrichment
- Mailbox verification
- Catch-all handling
- Duplicate and suppression checks
- Security-provider exclusions
- Re-sourcing failed or ambiguous rows
A cheap raw list can become expensive after low usable yield. The campaign runway model converts send targets into lead demand, while the two-gate verification workflow explains why not every purchased row belongs in the campaign.
Include replacement reserve
A production system needs spare capacity. Budget a reserve for mailboxes and domains that are warming, under review, or ready to replace failed infrastructure. Without a reserve, every incident becomes a choice between missing the send plan and pushing unhealthy accounts harder.
Keep replacement economics explicit:
planned active capacity x expected monthly replacement rate x setup and ramp cost per replacement = monthly replacement reserve
Do not invent a universal replacement rate. Measure it by provider, domain cohort, campaign type, and age.
Price the operator layer
Infrastructure is not self-managing. Someone must review authentication, mailbox health, campaign limits, list imports, exclusions, bounce patterns, replies, sender assignments, DNS changes, and failed automations. If that labor is not in the model, the campaign appears cheaper than it is.
Track operator time by repeatable work category. It becomes easier to decide what to automate and what must remain under human review.
Use two outcome costs
cost per usable campaign send = total monthly infrastructure cost / actual campaign sends that passed controls cost per qualified positive reply = total monthly infrastructure cost / verified qualified positive replies
The first metric exposes operational waste. The second connects infrastructure to commercial output without calling every reply a success. Keep automatic replies, out-of-office messages, bounces, and unqualified responses out of the denominator.
Frequently asked questions
How much does cold email infrastructure cost?
It depends on mailbox and domain count, providers, sequencer, data, verification, monitoring, replacement, and operator labor. A mailbox-only total is not the full infrastructure cost.
Does $3.50 per mailbox include domains?
Not in the Ink Persuasion example. The 249 by $3.50 calculation is mailbox licenses only. Domains and every other layer are modeled separately.
How many emails should one mailbox send per day?
Ink Persuasion normally caps campaign email at 20 per mailbox per day. Older, proven mailboxes may sometimes run at 25 to 30 under close monitoring. This is an operating rule, not a guarantee or Google guideline.
What is the best way to compare providers?
Normalize every proposal into the same cost stack and compare actual usable capacity, contract limits, setup work, replacement terms, monitoring, and support.
Run the cold email cost model every month
A cost model becomes misleading when it stays frozen while mailbox health, usable data yield, replacement volume, reply quality, or provider terms change. Reconcile planned capacity with the month that actually happened.
- Mailbox availability: compare licensed mailboxes with the number that were genuinely campaign-ready.
- Usable sending capacity: remove ramping, paused, damaged, and replacement mailboxes before calculating capacity.
- Data yield: reconcile sourced leads with verified, deduplicated, campaign-eligible contacts.
- Operating cost: include monitoring, replacement, exception handling, and human QA rather than reporting software alone.
- Commercial output: divide the full monthly cost by verified qualified positive replies, not blended responses.
That monthly review turns a static cold email cost calculator into an operating control. It shows whether the system became more efficient or merely sent more volume.
Sources
Want something like this done for your outbound system?