Every investment decision rests on a forecast — of cash flows, interest rates, margins or exit values. The problem is rarely the lack of a forecast. It is the false precision of a single number.
A valuation of 124.6 million looks exact, yet it can move by a third when two assumptions change. Decision-makers need to understand that range, not just the midpoint.
Start with the value drivers, not the spreadsheet
Identify the few variables that really move the result: revenue growth, margins, capital expenditure, working capital, the discount rate and the terminal value. In most DCF models the terminal value accounts for a large share of total value, which makes the long-term growth rate and the discount rate critical assumptions.
Build three coherent scenarios
A scenario is a consistent story, not a random mix of optimistic and pessimistic inputs.
- Base case — the most likely path, based on current plans, market data and consensus where available.
- Upside case — what has to go right: faster growth, better pricing or lower financing costs.
- Downside case — a plausible stress: weaker demand, margin pressure, higher rates or a delayed exit.
Link macroeconomic assumptions to company drivers explicitly. If the downside assumes higher interest rates, the discount rate, financing costs and exit multiples should all reflect it.
Add sensitivities and probabilities
Sensitivity tables show how value changes when one or two inputs move — typically the discount rate against the terminal growth rate. Probability-weighting the scenarios produces an expected value, but always report the range as well: an investment committee needs to see the downside, not only the average.
Stress-test what would break the investment
Reverse the question: which combination of assumptions makes the investment miss its return target or breach a covenant or investment limit? Knowing the break-even point is often more useful than another decimal place in the base case.
Make the model auditable
- Separate inputs, calculations and outputs.
- Document the source and date of every assumption.
- Version the model and record who changed what, and why.
- Automate data updates with SQL or Python instead of copy and paste.
- Compare forecasts with actual results after each reporting period.
Rule of thumb: if you cannot explain in one sentence why the upside and downside differ from the base case, the scenarios are not ready for an investment committee.
From one-off model to forecasting process
Forecasting pays off when it becomes routine: updated regularly, compared with actuals and improved over time. That is what an automated valuation engine delivers — the same logic every quarter, with a clear audit trail.
This article is for information purposes only and does not constitute investment advice. Figures are illustrative.