Presenting to a room that will ask what you would cut

The case is good. The mechanism holds, the cost is credible, the measurement is obtainable and the falsifier is real. You finish, and somebody asks what should come off the plan to make room for it.

If you have no answer, the most likely outcome is not rejection. It is deferral — a note to revisit next planning cycle, which in an organisation with a clock on it means the case dies quietly of old age. This lesson is about arriving with the answer.

Why the question gets asked

It is not hostility, and reading it as hostility is the most common mistake. The question is a test of one thing: whether you understand that capacity is a queue rather than a pool.

A pool has slack in it. A queue does not — work enters it in an order and the order is what people negotiated. Someone who believes they are asking for money out of a pool proposes additions. Someone who knows they are asking for a place in a queue proposes trades. The second person is easier to fund, and not because they are more agreeable. They are cheaper to evaluate, because they have already done half of the evaluator’s job.

The backdrop is worth holding in view, from the fund-and-the-clock lesson. Bain, a consultancy that sells advisory work to private-equity firms and is therefore describing its own market, reports holding periods at exit that “now hover at around seven years” against an earlier average of five to six, and deals that once “required only 5% annual growth in… EBITDA” now needing roughly 10% to 12%. Read from the other side of the table, that is an environment in which an additive request with no trade attached looks like a request from someone who has not noticed the constraint everybody else in the room works inside every day.

Three things a proposal can displace

This course’s own breakdown, not anyone else’s. Naming which one you are consuming is most of the answer, because they are refilled on completely different timescales.

• Engineering capacity. Engineer-weeks that were going to be spent elsewhere. The most visible resource and the one people think of first. It is also the most fungible: a quarter of capacity freed somewhere is a quarter of capacity gained here.
• Operator and reviewer attention. The people who would have to look at your thing — a central team, a security review, an architect, a portfolio company’s own engineering lead who has to agree to integrate. This is the resource most often consumed silently and least often written into a cost line, and unlike engineer-weeks it does not scale by hiring quickly.
• Calendar position. The slot on the plan. If your work has to happen before something else can, you are not consuming capacity so much as reordering commitments — some of which were promised to a customer, and some of which were merely chosen internally. The difference between those two matters enormously and is usually knowable by asking.

Four answers, worst to best

“Nothing — this is additive”

This is almost never true, and everyone in the room knows the conditions under which it could be. It reads as not having looked, which retroactively weakens the cost field on your page: if you did not investigate what this displaces, what else did you not investigate?

“That is for leadership to decide”

True, and useless. Prioritisation is genuinely not your decision. But the question was not asking you to decide. It was asking whether you have information the decider does not, and handing it back unanswered wastes the one part of the trade-off you were uniquely positioned to assess — the relative value of two things you actually understand.

“I would cut project X”

Better, and risky in a way worth naming. Nominating someone else’s work for the chopping block is a political act as well as an analytical one, and if you are new, or from a central team arriving into a portfolio company, you will not have the standing to make it land. It can also be simply wrong, because you do not know which commitments are contractual.

“I would trade it against my own work first”

This is the answer. Start with your own list. If your team has three things queued and you are asking to add a fourth, say which of the three you would drop, and why the drop costs less than the addition gains.

It works for a reason that has nothing to do with humility. Trading against your own queue is the only version of the answer where you have complete information, where the estimate is yours to defend, and where nobody can accuse you of solving your problem with someone else’s budget. It also demonstrates the thing the question is actually probing: that you hold your own work to the same bar you are asking them to apply to it.

Deferral is a real answer, if you price it

Not every trade is a cut. Often the honest proposal is a reordering: this goes first, that goes a quarter later, nothing is cancelled. That is a legitimate and much easier ask, and it becomes dishonest at exactly one point — when you present it as free.

Deferral has a cost and you should be the one to name it. A quarter of delay might mean a customer commitment slips, a competitor ships first, a renewal conversation happens without the feature, or simply that the deferred work now lands closer to the end of the hold period, where there is less time for it to compound. Say which one applies. A reordering with its cost stated is a proposal; a reordering with its cost omitted is the same additive request wearing a different hat.

The homework this actually requires

You cannot answer the question from inside your own team’s backlog. The preparation is unglamorous and takes an afternoon:

• Read the current plan. Not the roadmap slide — the list of what is actually being worked on and by whom.
• For each item near yours in the queue, find out whether it is committed to a customer, required by a regulator or an auditor, or internally chosen. Those three categories have completely different cut costs and the difference is almost always knowable by asking one person.
• Find out who has to review your work as well as build it. The reviewer queue is the one that surprises people, because it does not appear on any plan and it is where central-team proposals stall.

If you have not done that work, the honest answer is still available and still much better than bluffing: “I do not know the plan well enough to name the trade yet. Here are the three things I would need to know, and I can have an answer by Friday.” That is a person managing their own uncertainty out loud, which is the behaviour this entire course is training.

What this does not cover

Your three cases are now complete documents: hypothesis, cost with a displacement line, mechanism, measurement, falsifier. Everything structural is in place, and one thing is not.

Nothing so far has stopped you from writing a case that is well-formed and still wrong — one where the measurement moves, the number looks good, and your change had nothing to do with it. The last lesson in this course is about that gap, why it is the failure mode the whole course was built to prevent, and what it costs the second time somebody checks.