One Currency. One Moment. Many Futures.

When the Strait of Hormuz went into flux, planners across a dozen industries had hours to decide: reroute, re-source, pre-build, or hold. A shipment three weeks out was suddenly a question. So was the capacity booked to convert it, and the promise date already sitting in a customer's system.

Every one of those decisions was good or bad on the day it was made, with the information available that day, across the futures plausible that day. What the strait did afterward changed the outcomes, not the quality of the decisions. A good decision can produce a bad result. A bad decision can get lucky. The outcome is the world's grade, not the decider's. Last week's edition made that case in full: vyan.ai/resources/blog/judging-decisions-by-how-they-turned-out

Which leaves the question worth pressing. If not the outcome, then what? If we cannot grade decisions by how they turned out, what exactly are we supposed to grade?

The question deserves a complete answer, because the graders most companies reach for instead are worse. Outcome-grading at least measures something real, just too late. The substitutes measure the wrong thing entirely, and do it with confidence.

The graders we use instead

Plan adherence measures obedience, not judgment. A planner who follows a stale plan into a wall scores a hundred percent. A planner who breaks from the plan and saves the quarter spends the review meeting explaining herself. Adherence says nothing about whether what the plan said was worth doing. By the time a plan is three weeks old, it usually was not.

Forecast accuracy grades an input, not a decision. You can own the most accurate forecast in your industry and still destroy value with it, because a forecast says nothing about how much inventory to position against it, what capacity to reserve, or what being wrong in each direction costs.

The forecast is the weather report. The decision is what you pack. Grading planners on forecast accuracy is grading travelers on meteorology.

Functional KPIs are the subtlest trap, because each one is real and each one is being hit. Service, inventory turns, utilization, freight: every one scored in its own units, with no exchange rate between them. So every function closes the quarter green while the enterprise closes it red, and no one made a bad decision by any scoreboard they could see. That failure got a full edition three weeks ago: vyan.ai/resources/blog/every-function-delivered-enterprise-value-didnt

The common flaw is the same underneath. A decision spends capital to move the business across time and under uncertainty, and none of these graders can see capital, time, or uncertainty. Fix all three at once, or you have only moved the blind spot.

One currency

Start with the unit. Grade every decision in a single decision currency that works across every function, and make that currency charge for the capital the decision ties up. The capital charge is what lets a percentage, a stock figure, and a risk exposure be added up and compared, and what stops a function from looking brilliant while quietly borrowing the balance sheet. Economic Value Added prices this exactly: profit, after the cost of the capital used to earn it. The June edition makes the full case; this piece takes it from there.

Because settling the currency answers only a third of the question. A grade has three properties: a unit, a timing, and an evidence base. The unit is the easy one. The two that outcome-grading gets wrong are when the grade arrives, and how many futures it is computed against.

The moment

A scoreboard you can only read after one future unfolds is a scoreboard for historians. The decider needs the grade on the day the decision is made, because that is the only day it can change anything.

Reported annually, EVA is a verdict. Computed at the moment of choice, it is an instrument. Same number, different moment, entirely different power.

And the moment is not a metaphor. Every operating decision has a point of commitment: the purchase order is cut, the capacity reserved, the promise date given, the truck loaded. Before that point, the grade can still steer the decision. After it, the capital is committed, the option has expired, and any grade that arrives is a post-mortem in the costume of management.

Most companies compute their economics on the wrong side of that line: quarterly, in arrears, in finance's systems rather than the planner's. Everything argued here comes down to moving one computation across it.

The futures

Computing a decision's EVA against one assumed future is just outcome-grading in advance. You have swapped the world's dice roll for your own, and yours is not better informed.

The honest computation runs the decision across several hundred plausible futures, varying demand, supply, lead times, prices, and disruption. What comes back is a distribution: an expected value, and a tail. Run the alternative across the same futures, and you hold two distributions side by side, the better decision visible before the world chooses which future to deliver.

Make that concrete. Suppose the choice is whether to pre-position a buffer of a critical component ahead of a shipping route that might seize up. The buffer costs a certain quarter-million to hold. Graded against the single most likely future, where the route stays open, it is pure waste: minus a quarter-million. That is the number that gets a planner overruled.

Now grade it honestly. Across several hundred futures, the route stays open in most and seizes in a meaningful minority. In the open futures the buffer costs its quarter-million. In the seize futures, the company without it misses weeks of shipments, breaks promise dates, and forfeits margin in the millions. Weighted together, the buffer's expected EVA is positive and its tail dramatically tighter. The distribution says buy the insurance. The single-future grade said fire the planner who did.

A fair question follows: where do several hundred futures come from? Not a strategy offsite. Not an analyst inventing scenarios. They come from the company's own transaction history, the honest record of how demand actually varies, how suppliers actually deliver, and how lead times actually drift, as opposed to what the master data politely assumes. Resample that record, stretch it within the ranges it has really exhibited, add the disruptions worth insuring against, and the result is futures that are each plausible and collectively truthful. The futures are not speculation. They are the company's own past, allowed to speak in more than one voice.

The usual objection here is data: ours is not clean enough. It is the wrong worry. You are not predicting which future arrives. You are describing the range the business has actually shown, and transaction history describes that range whether or not the master data is tidy. The messiness is precisely the signal, because that is the uncertainty the decision has to survive.

Return to the strait. The reroute decisions that looked expensive while the waterway stayed open were, across the futures visible at the time, the ones that protected the EVA distribution. The strait staying open did not make them wrong. It made them look unlucky, a different thing entirely. And a company that punishes unlucky-but-right will teach its planners to stop protecting the tail. That lesson gets very expensive exactly once.

So here is the whole answer in one line. Grade the decision in one currency that charges for capital, at the moment it is made, across the futures that were plausible at that moment. Anything less is the wrong unit, the wrong time, or the wrong number of worlds.

Can a company actually run this way?

The fair objection is that this reads like a decision theorist's fantasy: elegant on the page, impossible on a Tuesday. It is not. A generation of serious operators has run on economic profit, and companies do today. The practice is proven, just unevenly distributed. What is rare is sustaining it from the operating seat, at operating tempo, for decades. That is the case study this newsletter is presenting.

A large consumer goods enterprise ran its operating decisions on EVA for the better part of two decades. Not reported it. Ran on it: pricing, inventory, capacity, portfolio, the daily and weekly choices most companies settle with dashboards and seniority. The number was the language decisions were argued in, before the capital moved.

On July 14 I am hosting a fireside with the person who spent a career inside that system: Dr. Rakesh Sinha, 39 years at Godrej, now a VYAN advisor. The session is called The Number That Ran Godrej: what it takes to make one number the operating language of a company, what breaks, what resists, and what becomes possible. If the question at the top of this piece is one you have asked, this is the hour that answers it. Register here: vyan.ai/webinars/eva-fireside-rakesh-sinha

One closing thought. Judging decisions by their outcomes at least has the decency to wait for reality. Judging them by adherence, accuracy, or functional dashboards does not even do that. The honest alternative is not a softer grade. It is a harder one, issued earlier, in a currency the whole enterprise can read, against every future that could have arrived.