Building a Simple Valuation Model: A Walkthrough
3 steps · one page
In short
A valuation model is a structured way of writing down what you believe about a business and seeing what those beliefs imply.
What this is. A mechanical walkthrough of how a valuation model is assembled, using Wexford Instruments and the Pillar 22 parameters. It ends in a range and a statement of what the range does not tell you. It is not a method for deciding what to buy, and it produces no view on the illustrative $12.00 price.
That is a modest description and it is the accurate one. The model does not discover value; it computes the consequences of assumptions — and the discipline of building one is valuable mostly because it forces the assumptions to be explicit and mutually consistent.
The seven steps
1. Assemble the history. Several years of the three statements, because a single year cannot show a trend. This is where Pillar 24 does its work.
2. Identify the drivers. Not every line needs forecasting. For most businesses four do most of the work: revenue growth, operating margin, capital expenditure, and working-capital intensity. Everything else can usually be tied to one of those.
3. Forecast the drivers, with reasons. Each assumption should have a sentence attached explaining why — and an assumption that cannot be justified in a sentence is one you have not really made.
4. Build free cash flow from the drivers. Revenue, then margin gives operating profit, then tax, then add back non-cash charges, then subtract capital expenditure and the working-capital movement. The point of building it from drivers rather than forecasting the total is that the components can be checked individually.
5. Discount, and set a terminal value. Using the cost of equity and terminal growth, remembering that the terminal value will be most of the answer.
6. Triangulate. Check the output against what multiples imply and against the implied growth in the current price. Agreement between methods is a sign the arithmetic is sound; it is not evidence the answer is right, since all the methods share the same inputs.
7. Present a range, never a point. Using coherent scenarios.
Worked example
Worked example: the model, assembled (canonical figures, USD millions). Step 1 — history. Revenue 880.0 → 1,000.0; operating margin 8.5% → 10.0%; capital expenditure 78.0 against depreciation of 42.0; working capital consumed 14.0; free cash flow 20.3. Step 2 — drivers. Revenue growth, operating margin, capex intensity at 7.8% of revenue, and working-capital intensity at roughly 1.4% of revenue. Step 3 — assumptions, each with a reason. Growth moderating from 13.6% because the current rate reflects an investment cycle that will not repeat indefinitely; margin holding at 10.0% because the expense base scaled exactly with revenue last year, showing no operating leverage to extrapolate; capex falling toward depreciation as the programme completes; working capital tracking revenue. Steps 4 and 5 — the base case. Free cash flow of 20.3 growing at 8% for five years, terminal growth 2%, discounted at 9.0%: explicit period 98.7, terminal 282.5, total 381.2. Step 6 — triangulate. Reverse the market price instead: $12.00 implies perpetual growth of 7.19% on reported free cash flow of 20.3, 4.12% on free cash flow at maintenance capital expenditure (56.3), and about 3.6% if earnings of 62.3 are capitalised instead (a P/E of 19.26 at a 9.0% cost of equity). The reversals disagree with each other by more than three percentage points depending on which cash flow is capitalised — which is itself the finding: the disagreement between the market and the model is a disagreement about how much of today's spending is investment rather than cost. Step 7 — the range. Coherent scenarios give 247.7 to 614.9. And what the model has produced. A range of 247.7 to 614.9 against a market capitalisation of 1,200.0 — and the honest report is that this model cannot bridge that gap. Either the driver assumptions are too conservative — particularly the capex assumption, since free cash flow at maintenance levels would be 56.3 rather than 20.3 — or the market expects something the model does not contain. The model has identified exactly where the disagreement sits. It has not resolved it, and it is not capable of resolving it. (Canonical figures; independently verified. Implied growth solves price = cash flow × (1 + g) ÷ (r − g) at r = 9.0%.)
Frequently asked
7 questions
What does a valuation model actually do?
It computes the consequences of your assumptions. It doesn't discover value — the discipline is valuable mainly because it forces assumptions to be explicit and mutually consistent.
Which drivers matter most?
For most businesses four do most of the work: revenue growth, operating margin, capital expenditure, and working-capital intensity. Most other lines can be tied to one of those.
Why build free cash flow from drivers rather than forecasting it directly?
Because the components can be checked individually. A total forecast hides which belief is doing the work; a driver-built forecast exposes it.
How do I know if an assumption is any good?
If you can't justify it in a sentence, you haven't really made it. Every assumption should carry its reason — that's what makes the model reviewable, by you or anyone else.
What does triangulation prove?
That the arithmetic is sound. It isn't evidence the answer is right, since the methods share inputs — on the illustration, the reversals of the price disagree with each other by more than three points depending on which cash flow is capitalised, which locates the disagreement rather than settling it.
What did the walkthrough conclude?
A range of 247.7 to 614.9 against a market capitalisation of 1,200.0 — and that the model can't bridge the gap. Either the driver assumptions are too conservative, particularly on capex, or the market expects something the model doesn't contain.
Is that a failure of the model?
No — it's the model working. It identified precisely where the disagreement sits, which is what this kind of work is for. Resolving it requires judgement about the business rather than more arithmetic.
References
- Investor.gov (SEC) — How to Read Financial Statements —
- Investor.gov (SEC) — Investor Bulletin: Performance Claims (a valuation figure needs its assumptions stated to be checkable) —
Educational and informational only — not investment advice, a recommendation, or an offer to buy or sell any security. Investing involves risk, including the possible loss of principal. Worked examples use fictional companies and figures.