Executive summary: Cost of delay is what a decision costs you per unit of time while it remains unmade. It converts deliberation into a running meter, which is the only framing that reliably moves capital approvals, because it stops the conversation being about whether to spend and starts it being about what waiting costs. This guide covers how to calculate it, how to sequence projects using cost of delay divided by duration, and why it beats return on investment as a decision frame.
- What is cost of delay?
- Why does it work better than an ROI case?
- How do you calculate cost of delay?
- How do you sequence projects with CD3?
- Which decisions does it change?
- How do you make it visible enough to act on?
- What are the most common cost of delay mistakes?
- Cost of delay: operator FAQ
- About the Stagnation Assassin
What is cost of delay?
Cost of delay is the money a decision costs you for every day, week, or month it stays unmade. It reframes deliberation as an active expense rather than a neutral pause. In a constrained operation it is usually calculable with real precision, because the loss is simply the throughput your constraint fails to produce while you decide.
The concept is straightforward and almost nobody applies it, because organizational culture treats analysis as free. A meeting to review the proposal again costs nothing on any budget line. Another two weeks of due diligence appears on no report. The expense is entirely invisible, which is precisely why it grows so large.
Here is the reframe that changes behavior. Every day you spend deciding whether to fix your constraint is a day your constraint runs at its current rate. The output you did not produce today is gone. It is not deferred to next month, not recoverable in a catch-up push. Constraint capacity is perishable inventory, and every hour of it expires unused.
I have led transformations at Berkshire Hathaway, Illinois Tool Works, and Whirlpool, and cost of delay is the single most reliable tool I have for breaking approval paralysis. Not because the arithmetic is clever, but because it changes what the room is arguing about. Without it, the discussion is about whether the spend is justified. With it, the discussion becomes about how much the discussion itself is costing, and that discussion tends to end quickly.
Why does it work better than an ROI case?
An ROI case invites comparison against other uses of the money, which is an argument that can run indefinitely. Cost of delay attaches a price to the argument itself. It shifts the burden from justifying action to justifying inaction, and it puts a running number on the alternative everyone was treating as free.
Watch how the two framings play out in a room.
The ROI framing. This project returns 340 percent over three years. Finance asks how that compares to the other capital requests. Operations asks about the assumptions. Someone suggests a sensitivity analysis. The proposal goes back for refinement and returns in six weeks. Nothing about this process is unreasonable, and it costs a fortune.
The cost of delay framing. Our constraint currently runs at 78 percent utilization. Getting it to 88 percent adds roughly 23 units a day. Every day we do not act, that output does not happen. Here is what a week of further review costs. Now the six-week refinement cycle has a price tag attached, and it is usually larger than the investment being debated.
That is the whole mechanism. You have not changed the economics of the project at all. You have changed which option carries the burden of justification. Delay stops being the safe default and becomes a choice someone has to defend.
There is a fairness point worth acknowledging here, because this technique can be misused. Cost of delay is a legitimate argument when the analysis genuinely has diminishing returns, which in constraint work it usually does, since 80 percent confidence is enough to start exploiting. It is not a legitimate argument for skipping due diligence on decisions that are irreversible or that carry real downside risk. Use it to break paralysis on reversible operational moves. Do not use it to bulldoze governance on permanent commitments.
How do you calculate cost of delay?
Take the throughput improvement the decision would deliver, express it per unit of time, and value it at throughput per unit rather than at revenue or accounting margin. In constrained operations this is unusually precise, because the improvement is measured in constraint hours and constraint hours have a computable value.
Work through the structure on a constraint improvement.
Step one, quantify the improvement. Suppose eliminating changeover and downtime losses would move the constraint from 78 percent to 88 percent utilization, adding 23 units of daily output.
Step two, value each unit correctly. Use throughput value, meaning selling price minus truly variable cost. Not revenue, which overstates it, and not fully loaded margin, which understates it badly because most loaded costs do not change when you produce one more unit. Say throughput value is 400 dollars per unit.
Step three, compute the daily meter. 23 units at 400 dollars is 9,200 dollars per day of forgone throughput. That is the number. Every single day the decision sits unmade, the business gives up 9,200 dollars it will never recover.
Step four, scale it to the decision timeline. A five-day week costs 46,000 dollars. A month of further review costs roughly 184,000 dollars. Two months of deliberation costs about 368,000 dollars, which in most cases exceeds the investment under discussion by a wide margin.
A constraint improvement worth 23 additional units a day at $400 of throughput per unit costs $9,200 for every day it goes unapproved. Two months of deliberation costs roughly $368,000, which routinely exceeds the investment being debated. The analysis was never free. It simply never appeared on any budget line.
Two technical cautions. First, this arithmetic only holds where the improvement genuinely lifts system output, which means it must be at the constraint. Running the same calculation on a non-constraint improvement produces a large, impressive, and entirely fictional number, because the system output would not have changed. Second, use conservative estimates. A cost of delay figure that gets challenged and collapses damages your credibility on every subsequent proposal. Understate it and let the number still be devastating.
How do you sequence projects with CD3?
Rank projects by cost of delay divided by duration, known as CD3. A short project with moderate delay cost often outranks a long project with high delay cost, because finishing it fast stops its meter and releases the team sooner. Sequencing by CD3 rather than by size measurably reduces total delay cost across a portfolio.
This is where cost of delay stops being a persuasion device and becomes a genuine prioritization method. Most organizations sequence improvement work by perceived importance, which usually means largest benefit first. That is intuitive and it is not optimal, because it ignores how long each project blocks the others.
Take four projects competing for the same team.
- Project A: cost of delay 50,000 dollars per week, duration 2 weeks
- Project B: cost of delay 30,000 dollars per week, duration 1 week
- Project C: cost of delay 20,000 dollars per week, duration 8 weeks
- Project D: cost of delay 10,000 dollars per week, duration 1 week
Sequenced by size, you run A, B, C, D. Total delay cost across the portfolio comes to 530,000 dollars. Sequenced by CD3, dividing each cost of delay by its duration, the order becomes B, A, D, C, and total delay cost falls to 460,000 dollars. Same projects, same team, same durations. Seventy thousand dollars saved purely by ordering.
The counterintuitive result is that Project C, with a substantial 20,000 dollars per week of delay cost, belongs last. Its eight week duration means running it early blocks everything behind it for two months, and the accumulated delay on those blocked projects exceeds what you save by starting C sooner. Long projects are expensive in a way that has nothing to do with their budget.
Which decisions does it change?
Four categories most often: capital approvals that are stuck in review cycles, improvement project sequencing, whether to expedite a constraint repair outside normal procurement, and whether to make an interim fix now rather than wait for the permanent solution. All four share a structure where waiting appears free and is not.
Capital approvals stuck in review
The classic case. A proposal cycles through refinement while the constraint keeps running at its current rate. Attaching the daily meter to the review process usually compresses a multi-month cycle into weeks, because nobody wants to be the reason the meter kept running.
Improvement project sequencing
Covered above through CD3. The gain here is permanent rather than one-time, because it changes how every future portfolio gets ordered. It also gives you a defensible basis for saying no, which is what makes a kill list survivable politically.
Expediting a constraint repair
A part for the constraint has a four week standard lead time and a two week expedited option costing 3,000 dollars more. Standard procurement policy says take the cheaper option. Cost of delay says two additional weeks of constraint downtime costs vastly more than 3,000 dollars, so the expedite is obviously correct. Write this rule into policy so it does not require a debate each time.
Interim versus permanent fixes
Teams often wait for the proper solution rather than implementing a partial fix now. Cost of delay makes the arithmetic explicit: if an interim fix captures 60 percent of the benefit and can be deployed in two weeks against six months for the permanent one, the interim fix usually wins by a wide margin even accounting for doing the work twice.
How do you make it visible enough to act on?
Put the running number where decision-makers see it repeatedly: on the proposal itself, on the project board, and in the meeting where the decision keeps getting deferred. Cost of delay only changes behavior when it is visible at the moment of deferral, not when it is buried in an appendix nobody reads.
Three practices make it stick.
Put the meter on the front page. Every improvement proposal should carry its cost of delay per week in the header, next to the investment figure. Seeing 46,000 dollars per week beside a 160,000 dollar request reframes the entire document before anyone reaches the analysis.
Show accumulated delay, not just the rate. A running total of what deferral has already cost is more forceful than a rate. If a proposal has been in review eight weeks, the header should say so and show the accumulated figure. This is uncomfortable, which is the point.
Report it after the fact. When a decision finally lands, report what the delay cost. Not to assign blame, which will kill the practice immediately, but to build institutional awareness that deliberation has a price. Organizations learn this by seeing it repeatedly, and the learning shows up as faster decisions on the next proposal.
The behavioral point underneath all three is that urgency is manufactured by visibility. People do not defer decisions because they are careless. They defer because deferral appears costless, and appearing costless is a property of measurement, not of reality. Change the measurement and the behavior follows without anyone needing to be exhorted.
Sequencing four projects by size rather than by cost of delay divided by duration cost one portfolio $530,000 against $460,000 for the CD3 order. The $70,000 gap required no extra resources and no faster execution. It came entirely from the order the work was done in.
What are the most common cost of delay mistakes?
Four recur: calculating it on non-constraint improvements where system output would not have changed, using revenue or loaded margin instead of throughput value, weaponizing it to bypass genuine due diligence, and computing it once for a proposal instead of maintaining it as a running figure.
Mistake 1: applying it to non-constraints
The most common error and the one that destroys credibility fastest. A team calculates an impressive cost of delay for an improvement on a process with spare capacity. The number is large and meaningless, because system throughput was never limited there. Someone eventually notices, and every future cost of delay figure gets discounted.
Mistake 2: valuing units wrongly
Using revenue per unit inflates the figure dramatically and invites a justified challenge. Using fully loaded accounting margin deflates it, because most loaded costs do not vary with an additional unit. Use throughput value, price minus truly variable cost, and be ready to show the derivation.
Mistake 3: using it to bulldoze governance
Cost of delay is a legitimate argument against unnecessary deliberation on reversible operational decisions. It is not a licence to skip diligence on irreversible commitments or decisions with real downside exposure. Deploy it that way once and finance will correctly treat it as a manipulation technique thereafter.
Mistake 4: treating it as a one-time calculation
Teams compute it for the approval meeting and never update it. But cost of delay changes as the constraint moves, as demand shifts, and as throughput value changes with mix. A figure calculated six months ago against a constraint that has since migrated elsewhere is not merely stale, it is pointing at the wrong process entirely.
My own mistake was overusing it. Having found a tool that reliably broke approval paralysis, I attached cost of delay figures to nearly everything, including proposals where the underlying case was genuinely weak. It worked for a while and then it stopped working, because the leadership team learned that a cost of delay number in my deck did not reliably indicate a good decision. Reserve it for cases where the constraint link is real and the analysis genuinely has diminishing returns. Its power comes from being credible, and credibility does not survive routine use.
Cost of delay: operator FAQ
What is cost of delay?
The money a decision costs for every day, week, or month it remains unmade. In a constrained operation it equals the throughput your constraint fails to produce while you deliberate, valued at price minus truly variable cost. It reframes analysis time as an active expense rather than a neutral pause.
How do you calculate cost of delay?
Quantify the throughput improvement the decision would deliver, value each unit at price minus truly variable cost, and express the result per day or week. For example, an improvement worth 23 additional units daily at $400 of throughput per unit costs $9,200 every day the decision goes unmade.
What is CD3 in project prioritization?
CD3 means cost of delay divided by duration. Ranking projects this way rather than by size reduces total delay cost across a portfolio, because a short project with moderate delay cost often outranks a long project with high delay cost. Long projects block everything behind them, which is a cost independent of their budget.
Why is cost of delay more persuasive than ROI?
An ROI case invites open-ended comparison against other uses of capital. Cost of delay attaches a price to the deliberation itself, shifting the burden from justifying action to justifying inaction. It does not change the project economics, only which option has to be defended, and delay stops being the free default.
About the Stagnation Assassin
Todd Hagopian is a Fortune 500 transformation executive who has generated $3B+ in shareholder value across Berkshire Hathaway, Illinois Tool Works, Whirlpool, and JBT Marel, where he serves as VP of Global Product Strategy. Known as The Stagnation Assassin, he is the author of two published books: The Unfair Advantage: Weaponizing the Hypomanic Toolbox and Stagnation Assassin: The Anti-Consultant Manifesto. His blog is published in 15+ languages and read by operators worldwide. Bring him to your stage via the speaking page or connect with him on LinkedIn.
Next step: put a meter on the decision
The proposal sitting in your review cycle has a price per week, and nobody has calculated it. Book a 20 minute session and I will help you put a defensible number on what your open decisions are costing, then show you how to sequence the work so the meter stops fastest. Start here.