Capacity planning: size outbound to the calendar you can serve

The most common outbound mistake is planning forward from list size instead of backward from the number of conversations the team can actually run, follow up on, and close. Forward planning produces a number that sounds ambitious and cannot be served: forty meetings a month landing on two reps who can each handle eight well. The result is not forty opportunities. It is sixteen decent conversations and twenty-four accounts that will never take a second meeting from you.

Work backward

Start at the end of the funnel and multiply upward. Each step uses your own numbers where you have them and conservative assumptions where you do not.

1. Qualified conversations per rep per week that get proper follow-up. Proper means research before, notes after, and a next step executed within two days. For most B2B teams selling a considered purchase, that is six to ten per rep per week, not twenty.
2. Apply your qualified-conversation rate to get the positive replies required. If two thirds of positive replies clear criteria, you need 1.5 positive replies for every conversation.
3. Apply your positive reply rate to get the weekly touch volume required.
4. Size sending infrastructure to that volume with conservative per-inbox limits.

The arithmetic, drawn

Run this with your own inputs before committing to a program size. The number that comes out of step five is the one that determines cost, ramp time, and whether the plan is achievable at all — and it is almost always larger than the intuitive guess.

Infrastructure is the hard ceiling

Sending domains need warming, and per-inbox daily volume has to stay conservative to protect reputation. That ceiling is real, and pushing through it is how programs damage the asset they depend on.

A workable configuration looks like this:

Add capacity by adding properly warmed infrastructure, not by raising limits. This is the single most expensive lesson in outbound, because the damage is invisible for two weeks and then takes two months to repair.

Plan the ramp

A realistic ramp is weeks, not days:

• Weeks 1–3: Domain and inbox provisioning, DNS records, warming. No campaign sends. List building and verification run in parallel.
• Week 4: Controlled launch on one narrow segment at a fraction of target volume. Watch bounce rate and placement daily.
• Weeks 5–6: Reply handling under real load. This is where the response path gets stress-tested and where most process gaps surface.
• Weeks 7+: Volume increases in steps, only once deliverability and reply quality hold steady at the current level.

Compressing this schedule does not accelerate results. It moves the failure earlier.

The failure mode nobody plans for

Capacity is not only sales capacity. Three other constraints bind before the sending limit does:

Reply-handling capacity. Someone has to read, judge, and respond to every reply within the hour. At 1200 touches a week that is a real workload, and it is the first thing to break under volume.

List supply. A narrow ICP has a finite number of accounts. Burning through it in eight weeks leaves you with nothing to send to in month three and a database of contacts who have already declined once.

Offer capacity. If the thing you are selling requires a specialist to scope, and there is one specialist, the program is sized to that person regardless of how many meetings the reps can hold.

Write down which of the four constraints binds first. That is the number the whole program is planned against; everything else has slack.

Reviewing the plan

Capacity assumptions should be revisited monthly against actuals — specifically the positive reply rate and the qualified-conversation rate, since those two multiply through the entire calculation. A positive reply rate that comes in at 1.2% instead of the assumed 2% does not mean the program failed. It means the required touch volume is 67% higher than planned, and the honest response is either to add infrastructure on a proper warming schedule or to lower the meeting target. Quietly raising per-inbox send limits to close the gap is the one option that makes everything worse.