I recently watched Safed Sagar, the name given to one of the Air Force operations during Kargil.
The sequence that stayed with me was the one on hitting enemy targets within a two-hundred-metre range, with no visibility of the target at all. It looks like pure skill, and a great deal of it is. But beyond the skill sits something else: defining the parameters that affect the trajectory, measuring them, and computing what they do to the release point. And doing all of it while the aircraft is travelling near the speed of sound.
Not slowly. Not on a second pass. At speed, blind, and accurate to two hundred metres.
Organisations are in the same position more often than they admit. Leaders frequently cannot see what is coming, and they still need to move fast. In that situation, the only thing that gives you confidence of landing close to where you intended is the parameters you have measured before release.
Most organisations have measured none.
The instruments are not what slow the aircraft down
This is the part I think gets misunderstood in business.
When people hear "measure the decision process", they hear weight. A form to fill in, a gate to pass, one more thing between having an idea and acting on it.
But an aircraft does not slow down to use its instruments. It goes fast because of them. Remove the atmospheric computation and you do not get a quicker attack run. You get one that has to fly lower, slower, and come around again. And still miss.
Take away the measurement and you do not get speed. You get movement without direction, which is a different thing entirely.
The same inversion applies to decisions, and almost every organisation has it backwards.
A company without decision infrastructure is not moving fast. It is re-deciding questions it already settled, because nobody can find the first answer. It is watching approvers hedge, because they cannot see what was granted in a similar case last quarter and do not want to be the outlier. It is granting a fourth exception to a rule that will cost three weeks of argument six months from now.
None of that shows up on any report. Which is precisely why it feels like speed.
What a trajectory actually depends on
A bomb released at altitude does not fall where it is pointed. It falls where the air puts it.
Temperature changes air density, which changes drag. Wind at release altitude is different from wind at eighteen thousand feet, which is different again near the ground, and the object passes through all of them. Humidity alters density. Pressure alters density. Every one of these bends the path in a knowable way, and knowable is the important word. None of it is guessed. It is measured, against records kept for years, precisely so that the crew does not have to guess in the seconds they have.
The crew does not choose the target. They do not choose the weather. What they control is the correction, and the correction is only as good as the parameters behind it.
Now ask the equivalent question about your own organisation.
When a manager approves a compensation exception on a Tuesday afternoon, what is the air doing?
Has that rule been bent before? How many times, and by whom? Did those people know about each other? Was there a pattern in who the exceptions were for? Is the rule itself still current, or was it written under conditions that no longer hold? How many decisions are standing on it right now, and would they all need revisiting if it changed?
Every one of those bends the trajectory of that Tuesday decision. Not one of them is measured in the overwhelming majority of organisations. The decision is released blind and everyone is faintly surprised where it lands.
The parameters of an organisational atmosphere
Once you start recording decisions properly, capturing what was decided, by whom, on what basis, and against which rule, a set of conditions becomes measurable that never was before. They group naturally, and each one bends the path in a specific way.
Density: how much weight is already in the air. How many decisions rest on a rule. How far a change to it would travel. A boundary definition with forty-seven promotions beneath it is a different object from one with three, and revising it is a different act. Air density determines how far a falling object travels; decision density determines how far a rule change travels.
Pressure: how hard the rules are being pushed. How often a rule is bent, as a rate rather than a raw count. What proportion of its scope it actually governs. And the number nobody expects: how many different people have granted exceptions to it, none of them aware of the others. Four exceptions granted by one person is a considered position. Four granted by four people who never spoke is a rule that stopped existing without anyone deciding it should.
Wind speed: how fast things are actually moving. How long decisions of this type take. How long each stage holds them. What proportion of the total delay each stage contributes. And how often a request comes back for rework, which conventional cycle-time measures cannot see at all. A request bouncing back twice looks like three short waits rather than one long journey.
Wind direction: which way the organisation is drifting. Where exceptions concentrate. If seventy percent of exceptions to a promotion rule are the same band transition, that is not a discipline problem across the company. That is one boundary that no longer matches reality, and everyone in the organisation already knows it. They have been acting on it quietly, one exception at a time, for eighteen months.
Visibility: how much of the atmosphere you can see at all. What proportion of decisions reference a rule that was actually recorded. Whether the reasoning behind a rule was ever written down. Whether the record can be proved to an outsider. This group governs all the others, because a computation with missing atmospheric data is not a computation. It is a guess with arithmetic attached.
Trend: whether conditions are changing. This period against the same period last year, never against last quarter, because appraisal season is not drift. Whether a rule has moved from healthy to being routinely worked around. And when the shift actually began, which is more useful than knowing it happened.
Consensus: agreement among those releasing. Whether approvers endorsed or approved with reservations. Whether two people reviewing in parallel ever disagree, because if they agree every time across fifty decisions, one of them is not adding a judgement.
None of this is exotic mathematics. Most of it is counting things and dividing. The instruments on that aircraft are not sophisticated because the physics is mysterious; they are sophisticated because somebody bothered to measure carefully, for years, until the patterns were reliable.
Why measurement, specifically
There is a difference between a forecast and a measurement, and it is worth being precise about, because it determines whether anyone trusts the system.
In India we have all made jokes about the meteorological department. An orange alert is issued, the day passes dry, and the credibility takes a small permanent dent. The forecast may have been perfectly correct, since a seventy percent chance of rain means no rain three times in ten, but a public that hears a warning and stays dry does not perform that arithmetic. It just stops believing.
A temperature reading has never had that problem. Nobody mocks a thermometer.
The ballistic computer on that aircraft is doing the second thing, not the first. It is not predicting whether the mission will succeed. It is measuring what the air is doing, right now, so the correction can be computed from facts.
That is the standard an organisational instrument should hold itself to.
"Eighteen exceptions this year, granted by four different approvers, seventy-two percent concentrated on one band transition" is a measurement. It cannot be wrong on Thursday. It can be checked, disputed, and traced back to specific records.
"This policy will fail next quarter" is a forecast, and every time it does not fail you spend a little of the trust you needed for the measurement.
The mathematics can support statements about what has historically happened. Of the rules that reached this state, roughly one in nine recovered without anyone deciding anything. That is genuinely useful, because it says waiting does not work. But it is a statement about a population, not a prophecy about one rule. The distinction is what separates an instrument from a fortune-teller.
The instrument
An organisation can measure all of this, and almost none do. Not because the mathematics is hard, but because the raw material does not exist. You cannot compute how often a rule has been bent if nobody recorded which rule each decision was following.
That is the gap SigmaGo is built to close.
It records a decision as it is made, not reconstructed afterwards. It links each decision to the rule it follows or bends, which is what makes every parameter above computable at all. And at the moment somebody is about to approve something, it puts the conditions in front of them.
Not a recommendation. A correction.
This bends the mid-year increment cap. It is the fourth exception this year, granted previously by three different approvers who did not know about each other. All four cited retention. Seventy-two percent of exceptions to this rule are the same band transition, and this request is one of them. Forty-seven decisions currently rest on this rule. It has not been reviewed in thirty-one months.
Six lines. Every one count from the record. The approve button stays live throughout.
The system does not choose the target and it does not release. It tells the person deciding what the air is doing.
Meritocracy is a measurement problem
There is a reason this matters beyond governance, and it may be the least obvious consequence of measuring at all.
Organisations describe themselves as meritocratic. Most mean it sincerely. But meritocracy is a claim about consistency: that similar cases are treated similarly, and that where they are not, there is a stated reason.
You cannot make that claim without measurement.
When four different managers each grant an exception without knowing about the others, the organisation has not made a decision about fairness. It has produced four separate acts of individual judgement that add up to a policy nobody wrote. Some people got the exception. Others, in similar situations, did not, because they asked a different manager, or asked in a different quarter.
That is not unfairness by intent. It is unfairness by invisibility, which is harder to see and harder to fix.
The moment exceptions are counted, the pattern becomes discussable. Seventy-two percent of these are the same band transition, and eleven of thirteen approvers gave the same reason. That is either a boundary that needs revising, or a population that has been quietly disadvantaged by a rule that no longer fits. Either way it is now a decision somebody can make deliberately.
Sustainability works the same way. An organisation survives its founders only if the reasoning behind its rules survives them too. A practice that outlives its reasoning is not continuity. It is inheritance, and inheritance is indistinguishable from good judgement right up until somebody asks why, and nobody can answer.
The thing worth taking from Safed Sagar
What struck me was not the courage, though there was plenty of it.
It was that precision at speed, without visibility, was not achieved by being careful. It was achieved by knowing exactly which parameters mattered, measuring them properly, and computing the correction in the seconds available.
Careful would have meant slower. Slower would have meant more exposure. The measurement is what made the speed survivable.
Every leader releases decisions into an atmosphere. Rules already bent four times. Frameworks nobody can justify any more. Approval stages that add three days and change nothing. Exceptions clustering on one group of people that nobody has counted.
Nobody measures any of it. Every decision is released blind, and organisations spend a great deal of time being surprised about where things landed.
You cannot go fast blind. You can only go fast measured.
