The questions investors ask
Some questions are predictable and should be pre-built into the model. "What happens if growth is 20% lower?" "What if churn is 2 points higher?" "What if we raise less than target?" Each of these should be a toggle or scenario that can be shown immediately.
Other questions are unpredictable and require model flexibility. "What if we delay the London launch by 6 months?" "What if we raise prices 15% on new customers?" These require adjusting specific assumptions in real time. A well-structured model handles this smoothly.
The worst response is "let me build that for you and get back to you in a week." It signals that the model is not genuinely flexible. By the time the answer arrives, the moment has passed. Investors expect models that can answer most questions within the meeting.
How to build for flexibility
Separate assumptions from outputs. Every lever should live on a single assumptions page. The P&L, cash flow, and balance sheet are all calculations from those assumptions. When an investor asks a question, you change one or two inputs on the assumptions page and the outputs update automatically.
Avoid hardcoded values in output pages. Every number in the P&L should be a formula referring back to assumptions. If Q3 revenue is typed directly, it will not update when assumptions change. The model looks flexible but is not actually.
Build scenarios as toggles. Base case, upside, downside. The toggle switches between assumption sets, and all the outputs recalculate. This lets you show three different scenarios within seconds rather than maintaining three different files.
The right level of detail
A model that has 500 rows of detail is often less useful than one with 100 well-chosen rows. Investors do not care about the granular detail; they care about the drivers. Revenue by product line, expenses by function, headcount by team - these are the right granularity for most investor conversations.
Too much detail also makes the model fragile. Every additional row is another place errors can hide. Every additional assumption is another place the logic can break. Ruthlessly simplify to the drivers that actually matter.
If investors ask for more detail during diligence, it is easy to add. Starting with too much detail and trying to simplify later is harder. Design for the 80% use case and add complexity only when required.
Presenting the model
Walk through the model live when time allows. A 15-minute model walkthrough where you show the structure, explain the assumptions, and flex a couple of scenarios builds more confidence than a polished slide deck.
Pre-prepare the key scenarios you expect investors to ask about. Know what happens at 20% growth deceleration, at higher churn, at slower hiring. Having the answers ready means you can respond without re-running the model, which signals strong command of the business.
Know the limits of the model. If an investor asks about a scenario the model cannot handle (say, a major new business line), acknowledge it directly. "The current model does not handle that cleanly - let me think through how to answer that." This is more credible than trying to force a bad answer.