Ask a CEO what bad decisions cost the company and you get a story: the hire that did not work, the market that was not there. Ask what they cost per year and you get a shrug, because nobody keeps that ledger. I built the dice-tax calculator to make the shrug into a number, and this guide is the calculator written out, with the sources beside each line, so you can see where the number comes from and argue with it.
The name is deliberate. Deciding on instinct alone is rolling dice with loaded odds, and the tax is what the odds cost you over a year.
The three lines of the bill
The cost of deciding badly does not arrive as one invoice. It arrives as three, and only the last is visible.
The first is the cascade: people executing the wrong things because a decision was late, unclear or wrong. Michael Mankins and Eric Garton measured it in Time, Talent, Energy (2017) as organizational drag, and found the average company loses about 21 percent of its productive capacity to it, against 13 percent for the top quartile. Not all of that is decisions. Mankins and Steele's earlier work (2005) on why strategies underdeliver traced the loss through a waterfall of causes, and roughly a third of it sits on decision-related failures: slow decisions, poor decisions, decisions not communicated. The calculator uses 37 percent as the share of drag that decisions explain, and it lets you change it.
The second is the everyday calls that fail: the vendor switches, the pricing tweaks, the hires, the discounts granted on a Friday afternoon. Individually small, collectively the bulk of what a company decides. Paul Nutt's twenty-year study (1999) found that decisions made with a single option in view failed about 44 percent of the time, and that weighing real alternatives brought that to about 30 percent. Those two rates are the difference between the two bills.
The third is the big bets that fail, the calls that shape the company: a market, a product line, a second office. Same failure rates, far larger stakes.
The arithmetic
For each line, the calculator multiplies a volume by a stake by a failure rate by a loss. A 40-person company making 15 everyday calls a week and one big bet a month, at a fully loaded cost of about $80,000 per employee, everyday stakes of $2,000 and big-bet stakes of $50,000, with half the stake lost when a decision fails, produces two bills.
Deciding blind, at the average company's drag and the single-option failure rate, the year costs on the order of $690,000. With a structured process, at the top-quartile drag and the multi-option failure rate, it costs on the order of $460,000. The difference is about $230,000 a year for a company of that size, and it scales with headcount and with how many decisions you make.
Every one of those inputs is a knob on the calculator. If your everyday stakes are $500 and not $2,000, change it. If you think decisions explain a tenth of drag rather than a third, change that too. What matters is that the arithmetic is visible, and that the coefficients come from published studies rather than from a marketing department.
The two numbers I would not change
The failure rates and the drag figures are the spine of the bill, and they come from unusually solid work. Nutt followed four hundred real decisions for two years each; nobody has done anything like it since. Mankins and Garton's drag index comes from Bain's client base across industries. These are cohort differences: companies that decided one way versus companies that decided another. They are not the effect of any tool, and the calculator says so on the page.
Two other findings explain why the gap between the bills is not smaller. Russo and Schoemaker (1992) found that managers who say they are 90 percent sure are right about half the time. Kahneman and colleagues (2016) found that professionals judging identical cases at one insurer differed by a median of 55 percent, where the executives had guessed 10. Overconfidence and noise are why single-option decisions fail at 44 percent rather than at a rate anyone would accept if they saw it written down.
What a structured process is worth
The second bill is the one you can move toward, and the levers are the ones the research names: a real alternative before every decision that matters, the team's independent read before the decision-maker's view is known, conditions written so they can be checked, and a look-back. None of them cost much. Nutt's alternatives finding is an afternoon of work per big bet.
The calculator is careful to price the structured process as a process, not as a purchase. The moss-coloured box at the bottom shows the difference between the bills; what you do with it is your call.
Where Midfire fits
The dice-tax calculator is Midfire's, and the second bill describes how Midfire runs a decision: the alternatives, the independent input, the conditions, the look-back. The product's price is a line in the bill, and at any company size in the calculator's range the difference between the bills is many times that line. Put your own numbers in before you believe any of ours.
The honest caveat
A calculator like this is an argument. It says: if your company decides the way the average company decides, and if the published failure rates apply to you, this is what it costs, and this is what the successful half's habits would be worth. The only way to know your actual number is to keep the ledger, decision by decision, with a look-back on each. Which is, as it happens, what the second bill is asking you to start doing.