Delivery Capacity Diligence: Testing a Services Target Against the Plan, Not the Trailing Year

Delivery capacity diligence is the pre-close test of whether the people, crews and partner arrangements behind a services target’s revenue can also deliver the plan the buyer is underwriting. It is the one commercial subject where a complete data room still cannot answer the question, because trailing records prove that capacity existed at last year’s volume and say nothing about who controls the capacity the plan would have to add.

This is the buyer’s test of a business it does not own yet. The operator’s own decision about how much delivery to buy rather than hire, and what bought capacity costs in margin and customer trust, is a different question answered from inside the company and after the close. The sibling test on the commercial records, the pipeline and the system behind the forecast, is CRM and pipeline data diligence. What a seller should have assembled before any of it begins is the commercial evidence pack. And the condition this test is trying to price in advance, demand arriving faster than the business can deliver it, is the one selling into a capacity ceiling describes from the operator’s chair once it is too late to price.

Why the trailing year cannot answer the question

Commercial diligence is built to verify numbers that exist. Revenue, margin, customer counts, win rates and channel performance are all claims about a period that has closed, and the work is to test whether each claim survives being recomputed from source. Delivery capacity does not behave like that. It is not a claim about the past at all. It is a constraint on a number that has not been produced yet.

The variable that makes capacity binding is the buyer’s own plan, and the plan is not in the data room. A target that comfortably delivered last year is being bought in order to deliver something else. S&P Global reported in 2026 that 71 percent of general partners and 53 percent of limited partners now prioritize operational value creation, which is another way of saying that very few plans are a promise to hold the business flat. The return is supposed to come from the increment between what the business delivered and what it is being asked to deliver, and capacity is the constraint on exactly that increment.

This is why capacity is the subject most often missed by a thorough process. Every other commercial question has a wrong answer that shows up as a discrepancy. Capacity has a wrong answer that shows up a year or more later as a ramp that did not happen, by which point it reads as an execution failure rather than an underwriting one. Nothing in the trailing file was false. The file was simply answering a question nobody had asked it.

Who controls the increment

The useful question is not who performs the work today. It is who decides whether more of it can be performed. Those are different people, and in a services business they often sit in different companies.

Capacity that is owned expands at the speed of a labor market. The constraint is recruiting, licensing where the trade requires it, and the gap between a hire starting and that hire being productive without supervision. It is slow, it is visible, and it is forecastable from the target’s own history.

Capacity that is contracted expands at the speed of a negotiation, and only as far as the partner’s own spare capacity allows. That spare capacity is the single input in the whole exercise that sits in a third party’s books. The target cannot evidence it and usually has never asked. The question worth putting to the target is what share of its partner’s total output it represents. If the share is large, the target is its partner’s concentration risk, and the partner’s pricing power rises at precisely the moment the plan calls for more volume.

Capacity bought on the spot market expands at the speed of price. It is always available, which is why it reassures a diligence team, and it is never available at the rate in the model. A plan that absorbs growth through spot capacity has not solved the constraint. It has converted a volume risk into a margin risk and moved it to a line nobody is watching.

Owned headcount expands at the speed of a labor market, a contracted partner expands at the speed of a negotiation, and spot capacity expands at the speed of price. All three are available. Only one of them is available at the rate the model assumes.

Utilization, and why a target with no slack reads as a good one

High utilization is read in a diligence summary as operational discipline, and it is. It is also the clearest available sign that the first increment of growth has nowhere to go. A business running close to full has already sold the headroom the plan intends to buy.

The number to ask for is not average utilization. An average is the one form in which this question cannot be answered, for reasons the next section covers. What is worth asking for is the distribution, and in particular the days on which a site turned work away, quoted beyond its normal schedule, or pushed a job to a partner because no crew was free.

Turned-away work is the most valuable capacity measure in a services business and the one least likely to be in the file, because nobody invoices it. It is also the only capacity measure that is simultaneously a demand measure. A site turning work away is telling the buyer two things at once: that the constraint is already binding, and that the market would have absorbed more. Those two facts point in opposite directions for the plan, and both of them matter.

Where the business has no record of turned-away work, the dispatch or scheduling system usually holds a usable proxy in the gap between when a job was requested and when it was scheduled. That gap widening through a year of flat headcount is a capacity ceiling being reached in slow motion.

Capacity is local even when revenue is reported at company level

Capacity cannot be averaged, and a platform-level utilization figure is an average of things that cannot substitute for each other. A business at 70 percent across eight sites can be at 95 percent in the two trades the plan intends to grow, and the slack sits in a trade the plan does not mention and a region the growth is not planned for.

A crew in one region is not capacity for a job in another. Travel time is a real constraint with a real cost, and a plan that quietly relies on moving crews between sites is a plan with a margin assumption inside it that nobody wrote down. The grain that answers the question is site, then job type or trade, then any license or certification the work legally requires.

That last grain is the one most often skipped and the one most likely to bind hardest. Where the work requires a credential, the number of people holding it is the ceiling, regardless of how many crews the business employs. A certification held by two people is not a credential. It is a capacity ceiling with two names on it, and the plan should be read against those two names rather than against the headcount schedule.

The same logic applies to anything in delivery that only one person can do. A buyer who has tested customer concentration and never tested delivery concentration has priced one side of the same risk.

Turning the growth plan into capacity the business does not have

The test is arithmetic, and it is short. Take the revenue plan by job type. Divide by the average job value for that type to get a job count. Divide that by the jobs a productive crew completes in a trading year to get the number of crews the plan requires. Subtract the crews the business has. The remainder is the plan’s real ask, expressed in the only unit that can actually be hired.

Then run the remainder against two things. The first is the model: does the cost build contain that many additions, scheduled before the revenue they are supposed to produce? The second is the target’s own record: has the business ever added productive crews at that rate in any year it has traded? If the plan requires more additions in its first year than the business has ever managed in its best one, that is not a stretch target. It is a different business, and the difference has to be funded and staffed rather than assumed.

Most models fail this test in the same way. Revenue grows as a rate and delivery cost grows as a percentage of revenue, which is arithmetically tidy and operationally silent. A model that grows revenue and grows delivery cost as a share of it has not planned capacity. It has assumed capacity, and the assumption is invisible because it never appears as a line anybody has to defend.

Three findings, and the one that moves timing rather than price

Capacity diligence produces three kinds of finding, and they are worth separating before any of them reaches a negotiation, because they are answered with different instruments.

The first is that the capacity behind trailing revenue is not what the seller described. A large bought share presented as owned delivery, or a partner the business could not replace, changes what the margin actually was. That is a price conversation, and it is the only one of the three that is.

The second is that the capacity exists but is not organized to expand. There is no recruiting function, no path from new hire to unsupervised work, no second holder of a required certification. The trailing number is sound and the growth mechanism does not exist. That is a plan conversation. It goes into the value creation plan with a named owner, a cost and a date, and it is usually cheaper to fix than it looks once somebody owns it.

The third is the common one and the one buyers handle worst. Everything the seller said is true, the capacity is real, the business can expand it, and the expansion takes longer than the model assumed. The revenue is not smaller. It arrives later. That is a timing conversation, and a discount is the wrong instrument for a date. Arguing it as a haircut is how a buyer loses credibility on the two findings that deserved one.

What the business cannot produce is evidence in its own right. A seller who cannot split delivered jobs into owned and bought at site grain is not withholding the number. The business does not keep it, which means nobody inside has been managing the constraint, which is both a finding and, for a buyer willing to install the measure, the cheapest value creation item in the file.

Frequently asked questions

What is delivery capacity diligence?

Delivery capacity diligence is the pre-close test of whether the people, crews and partner arrangements behind a services target’s revenue can also deliver the plan the buyer is underwriting. It is the one commercial subject where a complete data room still cannot answer the question, because trailing records prove that capacity existed at last year’s volume and say nothing about who controls the capacity the plan would have to add.

Why does trailing revenue not prove a services target can grow?

Revenue is evidence of capacity in the past tense. It confirms that the business delivered the volume it delivered, which is a statement about a year that is finished. The plan asks for volume the business has never produced, against a constraint that is not in the file because it has not yet been tested. A clean trailing record and an undeliverable plan sit together comfortably, because they are statements about different years.

What does it mean to ask who controls the capacity increment?

It separates who performs the work today from who decides whether more of it can be performed. Owned headcount expands at the speed of a labor market, a contracted partner expands at the speed of a negotiation and only as far as its own spare capacity, and spot capacity expands at the speed of price. All three are available to a buyer. Only one of them is available at the rate in the model.

Is a delivery capacity finding a price conversation or a plan conversation?

Usually neither. A misdescribed capacity base is a price conversation and a capacity that cannot expand is a plan conversation, but the common finding is that everything the seller said is true and the ramp is slower than underwritten. That is a timing conversation, and a discount is the wrong instrument for a date. A buyer who argues all three as a haircut loses credibility on the two that deserved it.

What should a buyer ask for to test delivery capacity before close?

The owned and bought split of delivered jobs by site and job type, the days a site turned work away or quoted beyond its normal schedule, the named holders of any license or certification the work requires, and the business’s own record of adding productive crews. A seller who cannot produce the owned and bought split at site grain does not run the business at that grain, which is itself a finding.

Have a revenue problem the board is asking about? Start a conversation.