The duration at which a project costs least is set by an overhead recovery rate chosen inside finance rather than by anything about the work — and because the analysis is taught as a recovery move, enterprises discover their own cost-minimising duration only once they are already late.
A project falls behind. The team compresses it, adding shifts and resources to the activities controlling the date. The compressed version finishes sooner and costs less in total than the approved plan.
This is treated as a recovery success. It is proof that the baseline was never the cheapest way to do the work, and that nobody had looked. The enterprise did not save money by recovering; it stopped losing money it had been committed to losing since the plan was signed.
The same arithmetic was available at baseline and costs almost nothing to run. But it is filed under recovery, so an enterprise that never falls behind never learns what its cheapest duration was.
Underneath sits something less comfortable. The location of that minimum is not a fact about the work. Total cost is direct cost, which rises as the schedule is compressed, plus indirect cost, which accrues per unit of time and falls as the schedule shortens. The minimum sits where the marginal cost of buying one more day equals the indirect cost of running one more day. Change the indirect rate and the optimum moves, though nothing about the engineering, the labour or the supply chain has changed. That rate is set inside finance, on a basis the delivery organisation rarely sees derived and almost never challenges.
The Strategic Context
The number doing the deciding is an overhead recovery rate, built to absorb a pool of standing costs across the work expected to carry them: management, premises, systems, depreciation. It is a costing convention with a legitimate accounting purpose, not built to answer the question the delivery team asks of it. Two consequences follow.
The first is that a rate designed to spread a fixed pool does not behave like a cash saving. If a project finishes early and the standing costs continue, the pool does not shrink; it is recovered from whatever remains, raising the rate borne by other work. At project level this shows as a saving; at portfolio level it may be a transfer, and only the portfolio can tell the difference.
The second is that the rate determines the answer while sitting outside the accountability of whoever acts on it. A delivery director defending a compression decision is defending a conclusion whose most influential input they did not set.
The material that teaches this technique reflects the confusion faithfully. In one treatment the indirect rate exceeds every marginal compression cost, so the cheapest duration is simply the shortest achievable one. In another there is no indirect rate at all and compression is judged against contractual penalties, so that treatment contains no cost-minimising duration whatever. Same technique, same discipline, and whether an optimum exists at all depends on whether somebody supplied a rate.
What Leaders Take the Trade-Off to Be
Four readings are common, and each keeps the analysis in the wrong place in the lifecycle.
The first is that compression is a recovery tool. It is a pricing tool, answering what duration this work should be bought at — a question answerable before anyone is late.
The second is that the baseline duration is the cheap one and going faster must cost more. The baseline is whatever emerged from sequencing the work at normal resourcing, a process that never consults the cost of time. The plan the enterprise approves is the plan the network produced, not the one the economics preferred.
The third is that the analysis is technical rather than governed. The arithmetic is an afternoon's work for a competent scheduler; the consequential input is an allocation policy — a governance decision made in finance.
The fourth is that compression opportunities keep. The cheap moves — a longer lead time on a fabrication slot, a second crew booked early, an order placed before expediting is needed — expire as the project proceeds. By the time slippage is visible the enterprise is buying from a shorter menu at worse prices, which is why late compression looks expensive and confirms the belief that speed always costs more.
Reframing the Issue
The reframing is to treat duration as something the enterprise buys rather than something delivery reports. Every plan carries an implied price of time. If leadership does not state that price, the schedule states it by default, and the default is whatever fell out of the network. Stating it deliberately turns a scheduling artefact into a commercial decision: at this rate, this is the duration worth paying for.
It also exposes what the analysis is not. It is drawn on deterministic durations, so it prices one expected outcome rather than the spread around it. What a stated confidence level in a completion date is actually the probability of is not treated here; that is the subject of [Related article: What Exactly Was the Board Given the Probability Of?]. The curve says which duration to aim at, not how likely the enterprise is to land there. Nor is this a scope conversation: value management run during delivery is a different discipline with different governance, treated elsewhere in this series. The question here is what the same scope costs at different speeds.
Where the Minimum Actually Comes From
The rate does all the work
Compressing a schedule means buying days on the activities that control the date, cheapest first, and each successive day costs more than the last as the cheap options are used up. Against that rising curve runs a straight line: the indirect cost of each additional day the project exists.
The optimum is where the two meet. Everything about the work sits in the rising curve; the straight line is the rate. Set the rate high enough and the cheapest schedule is the shortest the work physically permits; set it low and the cheapest schedule is the one the network produced unaided. Same project, same team, same suppliers, two correct answers, decided by a number the delivery organisation did not choose.
Consider an offshore wind construction and commissioning programme, hypothetically. Vessel charter, the marine spread and weather standby are genuinely time-variable: every additional day is cash leaving the enterprise, at a rate close to observable. The same programme's share of corporate overhead is also charged per day, appears in the same line, and stops on no particular day if the project finishes early. Two different economic objects arrive as one number, and the optimum is computed from their sum.
An allocation is not a cash saving
The portfolio question is whether shortening this project removes cost from the enterprise or moves it. If the released resources are redeployed to work that would otherwise have been bought outside, or standing capacity is stood down, the saving is real. If neither happens, the enterprise has reduced one project's absorbed overhead and raised everyone else's, and the compression premium was a cash outflow bought with an accounting inflow.
That is not an argument against compressing but for running the curve twice: once at the rate that stops when the project stops, once at the accounting rate. The two optima are frequently different durations, and which to act on depends on whether the capacity can actually be released — a portfolio judgement no project can make for itself.
Consider a supermarket distribution network commissioning an automated facility while the legacy site runs in parallel, hypothetically. The rent, labour and transport penalty of the parallel run stop on cutover day: avoidable cash, and it belongs in the rate. The group's share of head-office cost does not stop, and including it makes early cutover look better than it is.
The technique arrives only after the options have expired
The teaching material is consistent here in a way worth noticing. Each treatment opens with a project already behind, and the reader is shown how to buy time back. The analysis that finds the cost-minimising duration is presented as the endgame of a recovery, never a step in approving a plan.
One case is more revealing still. A worked example of compression turns out not to describe compression at all: it reports that moving the work to an outsourced provider reduced both time and cost. Buying the same work faster and cheaper is not a time–cost trade-off. It is the discovery that the baseline was priced above the market rate, made accidentally during a recovery exercise and reported as though the technique produced it.
That is the pattern in miniature. Enterprises that run this analysis only under duress keep discovering, in the worst circumstances, facts about their own cost base that a planning-stage review would have surfaced calmly. What follows is usually a re-baseline, which removes the evidence that the original duration was never the cheapest. What re-baselining does to the record of prior commitments is examined in [Related article: Re-Baselining Erases the Record of Every Promise You Broke]; the loss here is financial.
Decision Framework: The Indirect-Rate Provenance Check
No time–cost analysis should be accepted, at baseline or in recovery, without a check on the rate that determines its answer. The indirect-rate provenance check is six questions and one arithmetic step.
- Origin. Who set the rate, on what basis, and when was it last recalculated? An inherited rate is a decision nobody has made recently.
- Purpose. Was it built to recover a fixed pool across expected volume, or measured as the daily cost of running this project? Only the second is a cost of duration.
- Decomposition. Split it into the portion that stops when the project stops — site establishment, hired plant, temporary supervision, parallel running, financing on drawn capital — and the portion that continues regardless.
- Redeployment. If this project ends a month earlier, what happens to the released capacity? If the honest answer is that it waits for the next project, the absorbed portion is not a saving.
- Dual computation. Run the curve at the avoidable rate and at the full accounting rate, and record both optimum durations. Where they differ materially, escalate the difference to the portfolio: it is a question about capacity, not about this project.
- Timing. Confirm the curve was computed at baseline approval, not first drawn during recovery. If not, the schedule was approved without a price on time.
The arithmetic step: for the proposed duration, state the marginal cost of the next day of compression against the avoidable indirect cost of the next day of duration. If the first is lower, the enterprise is choosing to be slower than its own economics justify, and should say so deliberately. The adoption test is one question to any compression case: which rate did this use, and who owns it? A case that cannot answer was calculated, not analysed.
From Strategy to Execution
Immediate. Take the three largest projects in delivery and ask for the indirect rate embedded in each business case, with its decomposition. Expect one number, undecomposed, of uncertain age. That absence is the finding.
Medium term. Make the dual computation a condition of investment approval, so no capital release occurs without a stated cost-minimising duration alongside the proposed one and a reason for any gap. Require the same check whenever the rate changes, since a shift in overhead policy silently revalues every schedule in the portfolio. Note too that compression operates only on the activities controlling the date; the slack on everything else is a separate resource with its own ownership problem, examined in [Related article: Float Is the Only Resource With No Owner], which governs that reserve where this article prices the duration.
Long term. Build the finance and delivery interface that makes the rate contestable. The capability worth having is not a better scheduling model but a shared, current, decomposed view of what a day of project time costs — and the expectation that the number is produced before the plan is approved, not after it has failed.
Signals to Monitor
The strongest signal is a compression business case that appears only after a slippage report. Where the analysis is never seen at baseline, the enterprise is approving durations it has not priced.
Watch for a single undecomposed rate applied across dissimilar projects; for compression decisions whose stated savings are entirely absorbed overhead; and for recovery exercises that come in cheaper than the original plan, which are evidence of a baseline that was never optimal. Watch too for procurement findings surfacing during recovery: a supplier faster and cheaper than the in-house assumption is a pricing discovery, not a compression result, and should have been found at tender.
Questions for the Leadership Team
- For our three largest projects in delivery, what indirect rate per day was used in the approved business case, who set it, and when was it last recalculated?
- What proportion of that rate stops if the project finishes early, and what proportion continues, recovered from other work?
- For how many current projects has a cost-minimising duration ever been computed, and at what point in the lifecycle?
- When we last compressed a schedule, was the saving cash we stopped spending, or overhead reallocated to other projects?
- If a major project finished a month early, what would happen to the capacity it releases?
- When our overhead recovery rate last changed, which project schedules were re-examined?
Each requires evidence held in two functions that rarely compare notes. That is the point of asking.
Closing Perspective
The enterprise is already making this decision. Every approved plan contains an implied price of time, and every schedule never tested against a rate is a bet that the duration the network happened to produce is also the duration that costs least. That bet is occasionally right and never examined.
What leadership carries is not a scheduling responsibility. It is ownership of a number that decides how fast the organisation builds things, currently lodged in a costing convention no delivery decision-maker is invited to interrogate. An enterprise that decomposes that number, contests it and computes the curve before approval rather than after slippage will not always choose to go faster. It will, for the first time, know what it is choosing.
About EraNorth Insights
EraNorth Insights publishes practical analysis on strategy, projects, operations, transformation and decision intelligence for professional and organisational use. About EraNorth.