How much does the answer move when one assumption changes?

Here is the lesson that separates people who use DCFs from people who are used by them. A DCF's output looks solid, but it rests on a few assumptions — and nudging any of them can swing the value dramatically. Let's watch it happen.

One lever at a time

Take the model you assembled last unit — intrinsic value around €7.00 per share. Now change one input and hold everything else fixed:

• Discount rate. Drop WACC from 8% to 7% and the value might jump to roughly €8.50; raise it to 9% and it might fall to about €6.00. A single percentage point — well within the margin of error on WACC — moves the answer by a fifth or more.
• Perpetual growth. Nudge terminal growth from 2.5% to 3.5% and the value can leap toward €9 or beyond; drop it to 1.5% and it sags toward €6. That tiny (discount − growth) gap does the damage.
• Forecast growth. Change five-year growth from 8% to 5% and the near-term cash — and everything built on it — shrinks accordingly.

Same company, same day, wildly different "intrinsic values." Nothing about the business changed — only your beliefs about it.

Why the far assumptions bite hardest

The discount rate and terminal growth are the most dangerous levers because they act on the terminal value, which is usually the biggest slice of the total. A small twist there is magnified across the largest, most distant, least knowable chunk of the valuation. The near-term forecast, which you understand best, matters least; the far future, which you understand least, matters most. That's the cruel geometry of a DCF.

The honest response

This isn't a flaw to hide — it's information. The right reaction to "my value swings from €6 to €9" is not to pick €7.00 and pretend, but to report the range and name the assumptions driving it. A DCF that admits its own fragility is worth ten that project false confidence.

Try it now

1. Take your assembled DCF and change only the discount rate by one point up, then one point down. Record the two new values — how wide is the swing?
2. Now do the same with terminal growth (±1 point). Which lever moved the answer more, and why? Both act on the terminal value, which is the largest and least knowable slice.
3. Check that one point of WACC is even inside your margin of error. The safe rate under it is not a constant:

Read how far the floor itself has moved over five years. If the input travels further than the sensitivity you just measured, quoting the output to two decimals is theatre. 4. Now the comparison that matters. The market's own number is its market capitalisation:

Is the gap between your value and the price larger or smaller than the swing you recorded in steps 1 and 2? Often it is smaller — which means your model, honestly run, cannot tell the price from fair. That is a humbling and useful thing to notice.