How a Board Actually Reads a Go/No-Go Memo (And What Most Decks Get Wrong)

A board doesn't read a 40-page feasibility deck. Someone reads it for them, extracts three numbers and a recommendation, and that's what actually gets discussed in the room. Most of the deck exists to survive a challenge after the fact, not to inform the decision itself.

What the room actually asks

Sit in enough expansion approval meetings and the questions converge on the same handful, regardless of sector:

1. What's the number, and what's it made of?

Not "is the market attractive" — a specific score, and enough visibility into its inputs that someone can push back on one without discarding the whole conclusion. A demand score that's really one analyst's confidence dressed up as a number gets torn apart the moment someone with local knowledge disagrees with a single input. A demand score built from named drivers — sector mix, capacity bands, delegate rate trajectory — survives that challenge, because disagreement narrows to the specific input in question rather than the whole verdict.

2. What would have to be true for this to fail?

Every credible recommendation names its risks before someone else does. The weakest memos bury risk in a generic "market risk" line item. The strongest name the two or three things that would actually flip the verdict — a competitor announcement, a delegate-rate compression, a lease-rate shift — and say what would need to happen to trigger a re-look.

3. Why this city and not the other four on the list?

A recommendation that can't be compared against alternatives on the same basis isn't a recommendation — it's an opinion with a market attached. This is the argument for scoring every candidate city against one consistent model: not because any single score is infallible, but because the relative ranking across a shortlist is far more defensible than any individual number in isolation.

4. Who verified this, and how?

An AI-assisted or consultant-built analysis lives or dies on whether its claims are checked, not just generated. A board doesn't need to see the sourcing for every sentence — it needs confidence that someone (or something) went back over the first draft specifically hunting for what wasn't actually verified, and flagged it rather than smoothing it over.

5. The one nobody asks out loud

The fifth question rarely gets said directly, but sits behind every board discussion of an expansion memo: has anyone in this room actually been proven wrong by a recommendation like this before, and if so, what changed? A memo that ignores its own track record — that presents this quarter's shortlist with the same confident tone as every previous quarter's, regardless of how those played out — reads as untested rather than confident. The strongest recommendations acknowledge, briefly, what a prior verdict got right or wrong and why this one is built differently as a result.

This is also where a consistent, comparable scoring model earns its keep in a way a bespoke feasibility study can't. If every prior recommendation was built by a different analyst using a different process, there's no way to actually audit whether the method is improving — every miss gets explained away as market-specific bad luck. If every recommendation runs through the same model, a pattern of misses in a particular sector, region or delegate-capacity band becomes visible, and the model itself can be corrected rather than each individual verdict being separately re-litigated after the fact.

What this means for how a memo should be built

The instinct is to make the deck longer and more comprehensive — more comparables, more appendices, more caveats. That's backwards. A memo that survives the room is built the other way: locked, defensible numbers first, a named and ranked set of risks second, and prose that exists only to explain the numbers — never to substitute for them.

The uncomfortable version of this: if your expansion memo would fall apart under the five questions above, the problem usually isn't the writing. It's that the underlying analysis wasn't built to survive them in the first place — and no amount of narrative polish fixes that after the fact.

The takeaway

Before the next memo goes to committee, run it against those five questions yourself. If any of them exposes a number you can't defend or a risk you haven't named, that's the actual work — not another pass on the executive summary.