Can You Deliver What You Are About to Sell?

Sales delivery capacity is the number of new customers your business can take on per month without the quality of delivery falling, and it is the constraint most go-to-market plans are written without ever checking, which is why so many companies hit their pipeline target and get worse at the same time.

The plan says twelve new clients a quarter. Nobody has established whether onboarding twelve clients a quarter is a thing this company can currently do.

The failure looks like success for about two months

This is what makes it hard to catch early.

Demand arrives. Deals close. The revenue chart does what everyone hoped. Then delivery starts running two weeks behind, which nobody flags because everyone is busy and being busy feels like winning. Onboarding gets compressed. The parts of the process that were quality control get skipped first, because they are the parts with no immediate customer-visible consequence.

By month four the churn shows up, and it gets attributed to fit, or to pricing, or to a competitor. It was capacity. The company sold more than it could deliver well, and the thing that broke was the least visible step in a process nobody had written down.

Capacity is a property of a process, not a headcount

The instinct is to answer the capacity question with a headcount calculation. Four delivery people, three clients each, twelve clients. This is almost always wrong, and wrong in the optimistic direction.

It is wrong because onboarding a client is not four people working in parallel. It is a sequence with handoffs in it, and a sequence is governed by its narrowest point, not by the sum of its parts. If the technical setup step runs through one person who also does support, your ceiling is set by that person's availability regardless of how many others you hire.

It is also wrong because the calculation counts work and ignores waiting. Most onboarding processes are mostly waiting: for the client to return a form, for an approval, for someone to come back from leave. Waiting does not scale down when you add people, and in most small companies it is the majority of elapsed time.

So the honest answer to "how many can we take on" requires knowing where the work actually goes and who it queues behind, which is a process question rather than a resourcing one.

Find the narrow point before you sell against it

The exercise is not complicated and it is rarely done.

Follow one real client from signed contract to properly live. Every step, every person, every system, and every wait. Not the version from the onboarding doc. The last real one, with the person who ran it.

Then ask the question that matters: which step has exactly one person who can do it? That step is your ceiling, and your ceiling is lower than your headcount suggests.

Then ask what happens at the ceiling. Some processes degrade gracefully under load and some fail suddenly. A queue that lengthens is survivable. A step where the single owner starts skipping the check is not, because the damage is invisible until a customer finds it.

Building a working model of how the delivery side actually runs, with the single points of failure and the timings on it, is what abi. FDE produces from interviews with the people doing the work. The map itself is free on any account. The interviews are the paid part.

Then let it change the plan

Here is the part that makes this a go-to-market question rather than an operations one.

A capacity number should change what you sell, not just how fast you sell it. If you can onboard six clients a month well and eleven badly, the plan does not say eleven with a note about hiring. It says six, plus a decision about which constraint to spend on first.

It should also change who you sell to. Different customer types consume wildly different amounts of delivery. If your process handles a certain shape of client in four days and another in three weeks, your ideal customer profile has an operational component that most positioning work never considers. That is a real input to targeting, and it comes from the delivery side of the business rather than the market side.

And it changes the promise. The commitments in your sales conversation, the timelines, the onboarding experience, the responsiveness, are all claims about a process. If the plan is built on claims the process cannot support, the plan is not ambitious, it is inaccurate.

A go-to-market plan that has been through this is a different document. Writing one that starts from what the business can actually do is the difference between a plan built on researched market reality and a list of targets, and the delivery ceiling belongs in that research alongside the market sizing.

The argument against waiting

The obvious objection: if you check capacity before selling, you never sell, because early on your capacity is one client and it is the founder.

That objection is correct for the first ten customers, and the whole point of the first ten is that you deliver them badly and expensively and learn what the process needs to be. Do not model capacity when you have not run the process enough times to have one.

The exercise becomes worth doing at the point where delivery is no longer entirely the founder, and where a target number exists that somebody expects to be hit. That is usually the same quarter, and it is usually the quarter after somebody first says the word scale in a meeting.

The uncomfortable version of the finding

Sometimes you run this and discover the ceiling is uncomfortably low, and that the fix is not a hire but a redesign of how the work moves.

That is a harder answer than hiring, and it lands badly in a growth conversation, because it means the constraint is something you built rather than something you can buy your way out of. It is also the more valuable finding, since a hire against a broken process buys you one multiple of a bad thing.

The two halves of this are not separate projects. What you can deliver sets what you can honestly promise, and what you promise sets what you have to be able to deliver. Companies that run those as two conversations end up with a plan the operation cannot meet and an operation optimising for a plan it was never shown.

One afternoon

Take your revenue target for the next quarter. Divide it into new customers. Then go and find out how long the last three onboardings actually took, end to end, including the waiting.

If the second number cannot produce the first, you have learned something about your plan before the market teaches it to you, which is the cheaper of the two ways to find out.