The Forecast: Estimating Future Cash Flows

We are about to get into the heavy lifting. This is where investing stops being about reading the news and starts being about using math and logic to build your own forecast.

We are going to touch upon a tool called DCF (Discounted Cash Flow), but don't let the letters scare you. The concept is actually quite old-fashioned and logical: How much money will this business generate in the future, and how much is that money worth today?

If you can answer this question clearly, you have the power to value any company in the world. If you can't, you are just guessing.

The Foundation: Starting with Today's Reality

Before you look at a fancy financial model or a spreadsheet full of numbers, you must return to the basics. You must look at what the company is doing right now.

This is your Earnings Anchor.

You cannot build a forecast for the year 2030 on a foundation that doesn't exist today. If a company is losing money right now and burning cash, your forecast has a mountain of hurdles to clear just to break even.

Step 1: The Base Case

Start from free cash flow — operating cash flow minus capital expenditure — taken from the published cash flow statement.

Use free cash flow rather than net income. They are not interchangeable: net income includes non-cash charges like depreciation and excludes the capex a business needs to keep running, and the gap between the two can be large and persistent. Every annual report contains a cash flow statement, so there is no need to approximate.

Normalise the base year. This is the step most often skipped. If last year included a large one-off — a disposal, a legal settlement, an unusually light capex year — that distortion propagates through every forecast year and through the terminal value, where it gets multiplied.

Worked example. A company's free cash flow over five years: £42m, £38m, £71m, £45m, £47m.

The £71m year included a property sale. Taking it as the base would bake a one-off into perpetuity. A more defensible base is around £45m — the trend excluding the exceptional year.

That single choice changes the final valuation by roughly 50%, before you have made a single assumption about the future.

Step 2: The Base Rate

Next, you need to look at Revenue. Revenue is the fuel. If revenue doesn't grow, profit rarely grows for long.

• The Starting Point: How much money is coming in the door today?
• The Trend: Has revenue been growing at 10%, 50%, or shrinking?
• The Industry: Is the entire industry growing, or is it a shrinking pie?

If the industry is shrinking, a great company might still fail. This is why you must start with industry context, not just the company's stock ticker.

Estimating Growth: The "Magic Number"

This is the step most people get wrong. They look at a stock that doubled in price last year and assume it will double again next year.

Stop right there. That is guessing, not forecasting. True forecasting is about probability and sustainability.

You are looking for the Compound Annual Growth Rate (CAGR). You are trying to estimate the average rate at which the business can grow, year after year, for the foreseeable future.

Where Does Growth Come From?

To estimate a realistic growth rate, you must look at three specific drivers:

1. Market size. A company cannot outgrow its market indefinitely. If the addressable market grows 4% a year, sustained 20% growth requires taking share from someone — so identify who, and whether they will let you.
2. Market share. Is the company gaining, holding or losing? Share gains can outpace the market for a while, but rarely for a decade, and competitors respond.
3. Unit economics. Is each additional unit of revenue as profitable as the last? Growth bought by discounting is not the same as growth from demand, and it shows up as expanding revenue with flat or falling margins.

The size effect. 20% growth on £100m of revenue means finding £20m. The same rate on £10bn means finding £2bn — an entire mid-cap's worth of new business, every year. Large companies grow more slowly for arithmetic reasons, not managerial ones.

The End of the Road: Terminal Value

Here is the tricky part. A company doesn't disappear after 5 or 10 years. It keeps going.

If you project cash flows for only 5 years, you are missing the rest of the company's life. This is why we use a concept called Terminal Value.

Terminal value is essentially a shortcut. It is an estimate of what the business is worth at the end of your forecast period, assuming it will continue to grow at a stable rate forever after that.

The Perpetuity Method

The standard approach — the Gordon growth formula:

Terminal value = Final year's free cash flow × (1 + g) ÷ (r − g)

where g is the perpetual growth rate and r is your discount rate.

Worked example. Year 10 free cash flow of £80m, terminal growth of 2.5%, discount rate of 9%:

TV = £80m × 1.025 ÷ (0.09 − 0.025) = £82m ÷ 0.065 = £1,262m

Then — and this is the step people miss — discount it back to today. That £1,262m sits at the end of year 10, not now:

£1,262m ÷ 1.09¹⁰ = £533m

Forgetting that division roughly doubles your valuation. It is the single most common error in a beginner's model.

Two rules the terminal growth rate must obey

1. It must be below long-run economic growth. 2–3% is normal. A company growing perpetually faster than the economy eventually becomes larger than the economy, which cannot happen.
2. It must be below the discount rate. If g ≥ r, the denominator (r − g) becomes zero or negative and the formula returns an infinite or negative value. The model doesn't warn you — it just produces nonsense.

Check how much rests on it

Always calculate the split:

Terminal value as % of total = discounted TV ÷ total valuation

If 85% of your valuation sits in terminal value, then 85% of your answer rests on a guess about the distant future, and the ten years of careful forecasting are decoration. A ratio of 60–75% is typical and acceptable. Much above that, extend the forecast period — it forces you to think explicitly about years you were otherwise waving through.

Growth Must Fade

The single most common modelling error is applying one growth rate across the whole forecast period.

No company sustains high growth indefinitely. Success attracts competition, large numbers get harder to grow, and markets saturate. A model assuming 25% growth for ten years is assuming a company outruns those forces for a decade.

Fade the rate instead:

The shape matters more than the individual figures. A model that fades will produce a defensible answer even with imperfect numbers; a model that doesn't will produce an indefensible one however carefully you picked the rate.

The Common Mistake: The "Just Add More" Trap

Beginners often make this fatal error: They keep adding "optimistic" assumptions until the numbers look good.

The pattern looks like: "£10m this year, £15m next year, £20m the year after — and by year 10 it will be £100m." Each step sounds reasonable in isolation; compounded, they assume a tenfold increase.

Why this is dangerous: it is trivially easy to reach any valuation you like by adjusting inputs, and the model gives you no resistance while you do it. If you find yourself raising the growth rate until the answer matches the current share price, you have stopped valuing the company and started justifying a decision.

Good forecasting is boring. It relies on the historical record, the growth rate of the market the company sells into, and assumptions you would be comfortable defending to someone who disagreed with you.

Forecast a range, not a path

The honest way to handle this is three cases rather than one:

That spread is the honest output of a DCF. If the share trades at 250p, it is below even the bear case, which is genuinely interesting. At 470p it sits inside the range and the model tells you very little — which is itself worth knowing, and far more useful than a single confident number.

Putting It All Together: The Logic

When you build a DCF forecast, you are not trying to predict the future. You are trying to find the fair value of the business.

1. Base — normalised free cash flow today.
2. Growth — a rate you can justify, fading over time.
3. Horizon — five to ten years of explicit forecasts.
4. Terminal — the perpetuity, discounted back to today.
5. Discount — the rate that converts all of it to present value (next lesson).

Compare the result to the market price — but compare a range to it, and treat a narrow gap as no signal at all. The DCF calculator will run the arithmetic once you have the inputs.

Summary

• Normalise the base. Start from free cash flow, adjusted so the base year is representative rather than exceptional.
• Be Boring: Growth rates should be sustainable, not heroic. As companies get bigger, they grow slower.
• Terminal value usually dominates. Discount it back to today, keep terminal growth below both economic growth and the discount rate, and check what proportion of your total it represents.
• Fade the growth. A single growth rate applied for a decade is the commonest way to produce a nonsense valuation.
• Forecast = Decision Tool, Not a Crystal Ball: You are not trying to be right. You are trying to see if the stock price makes sense based on the math of the business.

Disclaimer: This lesson is for educational purposes only and does not constitute financial advice. Forecasting involves significant uncertainty and error.