The plan was beautiful. It always is. Someone had built it in the good software, the one that draws the dependencies as arrows and sets the milestones on their dates like trains that have never once been late. The room looked at it and felt the particular calm a well made plan produces, which is the feeling of a problem already solved. Eleven weeks. Four workstreams. A launch date with a little flag on it. Nobody in the room quite believed the eleven weeks, but nobody said twelve either, and so eleven became the number the company organized itself around and would later be genuinely surprised by.
Ask where the eleven came from and you get an answer that is honest and also a fiction. It came from adding up the steps. Someone estimated each task on the assumption that the task would go the way tasks go when you picture them, which is to say cleanly, in sequence, performed by a person who has nothing else to do, who does not fall ill, on a team that does not lose the one engineer who understood the integration, inside a company that will not decide three weeks from now that something else is suddenly the priority. The estimate is not a prediction. It is a best case wearing a prediction's clothes.
The plan priced a team that does not exist.
This is the quiet mechanics of almost every plan a company makes. It is built from the inside, task by task, each one priced at the speed it would run if the world cooperated. What it leaves out is the outside view, the deflating and statistically correct question nobody asks in the room: how long did the last five things that looked like this actually take. Not how long were they meant to take. How long did they take. The answer is always longer, the gap is always roughly the same size, and the company that has shipped forty projects late by a third walks into its forty-first certain this one will be different, because this time it has accounted for everything. It has accounted for everything except the one thing that made it late forty times, which is that accounting for everything is not possible, and the plan that claims to have done it is the least trustworthy kind.
The obvious objection is that experienced teams know this, so they pad. They take the honest estimate, add a buffer, and call the buffer realism. But a pad is not the outside view. It is the inside view with a tip added, and it fails for two reasons that compound. The first is that the buffer is set by the same optimism that set the estimate, so it is always too small, a gesture toward uncertainty rather than a measurement of it. The second is that a visible buffer does not survive a room that wants an earlier date. Someone asks whether the team is sandbagging, the reserve gets read as slack rather than as insurance, and it is negotiated away in the meeting by people who will not be the ones explaining the slip. The buffer that would have absorbed the friction is spent before the work begins, traded for a number that looks better on a slide.
Be precise about what the gap actually is, because this is where the usual story goes wrong. The distance between the plan and reality is almost never a competence gap. The people were not worse than expected. The work was not harder than it looked. What the plan omitted was not skill. It was friction, and capacity, and the ordinary entropy of a real organization: the approval that sat in an inbox for four days, the two people quietly shared across three projects, the context rebuilt every Monday because the work kept getting interrupted, the handover that went to someone who had never seen the system. None of that is a failure of talent. All of it is invisible on a Gantt chart, because a Gantt chart has no row for the week your best person spends covering for the one who left.
It is worst where the plan is about people rather than product. A capability build, a hiring ramp, a new team brought to delivery standard: these get planned the same confident way, as if readiness were a date you reach by adding up training weeks. The plan says the sixty new hires will be productive by March, because onboarding runs eight weeks and sixty minus attrition still clears the number. It prices the hires who stay, who learn at the assumed rate, supported by managers with nothing more urgent than developing them. The real ramp loses people in week three, learns at the speed the floor actually permits, and competes for the attention of managers running a live operation. The gap between the staffing plan and the staffed reality is not a people problem you can coach away. It is the same optimism, applied to the one resource that has feelings and the option to resign.
Which is the uncomfortable thing the plan really is. It is a portrait of the team you wish you had, doing the work in the conditions you wish you were in. Every box assumes the strongest version of the person: rested, undistracted, fully briefed, working the problem in a straight line from start to finish. Real work is not done by that person. It is done by the actual one, on an actual Tuesday, with four other things on the list and a manager already asking about a fifth. The plan priced the first person. The company is staffed by the second. And the distance between them is not a rounding error you can sprint to close at the end. It is the most reliable feature of the whole exercise, and it was designed in before anyone wrote a line of code.
The slip did not happen in the last fortnight. It happened the day the plan was approved, when a best case was written down as a commitment and everyone agreed to treat a hope as a date.
So the failure is not that the estimate was wrong. Estimates are always wrong. The failure is subtler and more human: the plan was never a forecast at all. It was a statement of intent dressed as arithmetic. When a team commits to eleven weeks it is not predicting the future, it is announcing the version of itself it intends to be, and half hoping that declaring the intention firmly enough will keep the friction away. It never does. But the intention feels like rigor, and rigor is what the room was assembled to produce, so the plan is approved on the strength of how disciplined it looks rather than how likely it is. The more detailed the plan, the more this happens. Precision reads as realism. It is usually the reverse: the longer the list of tasks, the more places there are for the optimism to hide, one confident little number at a time.
And because the shortfall surfaces at the end, it gets blamed on the end. The launch slips and the story the company tells is about the final two weeks: the integration that ran hard, the sign off that came late, the scope that crept. All true, and all beside the point. The slip did not happen in the last fortnight. It happened the day the plan was approved. The closing weeks are simply where the bill, which was always coming, finally arrives. Fixing the last two weeks changes nothing, because the error was never there. It was in the first hour, in the room, in the arithmetic nobody questioned because it added up.
What rebuilding looks like.
So what would you change. Not the people, and not the discipline, since the plan had plenty of that. You would change what the plan is permitted to assume. You would make every estimate carry the outside view beside the inside one: not only how long should this take, but how long did the ones like it actually take, with the gap written in ink rather than apologized for later. You would stop pricing the ideal team and start pricing the real one, the team that is shared and interrupted and occasionally short a key person, because that is the team that will do the work. And you would treat slack not as a sign of weak planning but as the one honest line in the document, the place where you concede that friction is a certainty and reserve for it on purpose instead of being ambushed by it on schedule.
The hard part is that no planning room can see its own optimism, because the room is where the optimism is manufactured. Everyone in it is rested, briefed and looking at a clean document, which is the exact condition under which people most overestimate themselves. The plan flatters the planners by showing them at their best, then sends them off to be their actual selves somewhere the plan cannot follow. That gap is what SSUNDAR built the crisis simulation to make visible. It puts leaders under the conditions a plan pretends away: competing demands, missing information, decisions stacking faster than they can be cleanly made, the friction arriving early and all at once. On paper, in the planning room, every leader looks deliberate and every timeline looks achievable. Under load you find out which parts of the plan were a forecast and which parts were a wish, and you find it out in an afternoon rather than eleven weeks too late.
Your plan was never a forecast of the work. It was a portrait of the team you were counting on being, and the work got done by the team you actually were.