A sensitivity table is one of the most trusted parts of a DCF model, mostly because it looks so objective. Rows and columns, a clean grid of numbers, no opinions attached. That's exactly what makes it dangerous when it's built wrong. Nothing throws an error. Nothing looks broken. The table just quietly tells you something untrue, and everyone in the room nods along because grids of numbers feel like facts.
Here are seven mistakes that show up constantly in DCF sensitivity analysis, none of which will trigger a warning in Excel, and all of which can meaningfully distort the story your valuation is telling.
Quick Answer: What's the single biggest mistake to watch for?
Treating input variables as if they move independently of each other. In reality, a lower growth assumption usually comes with tighter margins and higher capital needs too, not just lower revenue in isolation. A sensitivity table that only flexes one variable at a time, without accounting for how real-world downside or upside cases actually move together, tends to understate how much a valuation can really swing.
On this page
- Mistake 1: Treating Correlated Variables as Independent
- Mistake 2: Mismatching Cash Flow Type and Discount Rate
- Mistake 3: Ranges That Are Too Narrow or Too Wide
- Mistake 4: Building Sensitivities on a Base Case That Isn't Clean Yet
- Mistake 5: Reading the Grid as a Probability Map
- Mistake 6: Ignoring How Much Terminal Value Is Driving Everything
- Mistake 7: Burying Stakeholders in Too Many Tables
- A Quick Pre-Presentation Quality Scan
- Building It Right in Excel
- Key Takeaways
- Frequently Asked Questions (FAQs)
Mistake 1: Treating Correlated Variables as Independent
Most sensitivity tables flex one or two variables while holding everything else constant. That's fine as a mechanical exercise, but it quietly assumes those variables don't affect each other in real life, which usually isn't true. Revenue growth, operating margin, and capital expenditure tend to move together in a real downturn or a real growth spurt. Inflation and interest rates typically move together too. If your "bear case" only lowers revenue growth while leaving margins and WACC untouched, you're not actually modeling a realistic bear case. You're modeling a much milder scenario dressed up to look like a stress test.
Fix: When you build a downside or upside case, deliberately move the variables that would plausibly move together in that scenario, not just the one variable that's easiest to isolate in a two-way table.
Mistake 2: Mismatching Cash Flow Type and Discount Rate
Unlevered free cash flow needs to be discounted at the Weighted Average Cost of Capital (WACC). Levered free cash flow needs the Cost of Equity instead. Mixing these up is a surprisingly common error, and it doesn't produce an Excel error message. It just quietly produces a wrong valuation that still looks perfectly normal.
Once that mismatch exists in the base case, every cell in your sensitivity grid inherits the same error, scaled across every combination of inputs you test. The table will look thorough and precise. It'll also be wrong from top to bottom.
Fix: Before building any sensitivity table, confirm your cash flow definition and discount rate are aligned. This check takes thirty seconds and prevents an error that can otherwise hide in plain sight through dozens of grid cells.
Mistake 3: Ranges That Are Too Narrow or Too Wide
A sensitivity table is only useful if the range of values it tests reflects genuine uncertainty. Pick a range that's too narrow, maybe WACC from 9% to 10%, and the table will imply your valuation is far more precise and stable than it actually is. Pick a range that's too wide, maybe 4% to 20%, and the table stops being informative and starts looking like a random number generator, which tends to get the whole table dismissed by whoever's reading it.
Fix: Anchor your range to something defensible, like a reasonable band around your company's actual cost of capital, or a spread that reflects genuine disagreement among comparable analysts, not just round numbers that happened to be easy to type into a Data Table.
Mistake 4: Building Sensitivities on a Base Case That Isn't Clean Yet
Sensitivity tables amplify whatever is already in your base case model, including its mistakes. If your base case still has a timing error, a double-counted cash flow, or an inconsistent tax assumption, running a sensitivity analysis on top of it doesn't reveal that problem. It just multiplies it across every scenario in the grid, making the error harder to spot rather than easier.
Fix: Run a basic quality check on your base case before building sensitivities on top of it. Confirm the timing of cash flows is consistent, taxes are applied correctly, reinvestment assumptions make sense, and nothing is being double-counted between historical and projected periods.
Mistake 5: Reading the Grid as a Probability Map
This one isn't really a modeling mistake, it's a mistake in how the table gets interpreted afterward. A sensitivity table is a structured "what if" map, not a probability forecast. Every cell in the grid represents one hypothetical combination of inputs, not an equally likely outcome.
Teams sometimes treat the center of the table as "most likely" and the edges as "unlikely" without ever actually assigning real probabilities to any of it. That's a comfortable assumption, but it's rarely something the model itself supports.
Fix: When presenting a sensitivity table, be explicit about what it is and isn't showing. It shows how the valuation responds to different assumptions. It doesn't tell you which assumption is actually most likely to happen.
Mistake 6: Ignoring How Much Terminal Value Is Driving Everything
| Terminal value % of total DCF value | What it typically means |
|---|---|
| 60% to 75% | Fairly normal for most companies, since a large share of value legitimately sits far in the future. |
| Above ~85% | Worth a closer look. Often signals overly aggressive long-term growth or margin assumptions baked into the terminal period. |
When terminal value makes up the majority of your total valuation, and it usually does, your discount rate and terminal growth rate assumptions are doing most of the heavy lifting in the whole model. A sensitivity table that doesn't put real focus on these two variables is missing the part of the model that actually matters most.
Fix: Make discount rate versus terminal growth rate your primary two-way sensitivity table, not an afterthought buried behind other variables. These two inputs together typically explain the majority of how much your valuation moves.
Mistake 7: Burying Stakeholders in Too Many Tables
It's tempting to show off analytical thoroughness by presenting five or six sensitivity tables covering every possible variable combination. In practice, this usually backfires. Readers can't hold that many grids in their head at once, the actual takeaway gets lost, and the presentation starts to feel like it's hiding the answer rather than clarifying it.
Fix: Keep it to one primary two-way table (typically WACC versus terminal growth), one sensitivity focused on your key operating driver (like revenue growth or margin), and a short written explanation of what's actually driving the valuation. That combination covers almost everything a stakeholder needs without overwhelming them.
A Quick Pre-Presentation Quality Scan
Before you present a DCF sensitivity table to anyone, run through this list. It takes a few minutes and catches most of the mistakes above.
- Does the cash flow type match the discount rate (unlevered with WACC, levered with Cost of Equity)?
- Is the base case free of timing errors, double-counting, and inconsistent tax treatment?
- Do the ranges you're testing reflect genuine, defensible uncertainty, not arbitrary round numbers?
- If you're showing a bear or bull case, do multiple related variables move together, not just one in isolation?
- Is your primary table WACC versus terminal growth rate, given how much of total value usually sits in the terminal period?
- Are you presenting one clear table and a short explanation, rather than burying the finding under several tables?
- Have you compared the DCF output against a comparable-company check, rather than relying on sensitivity analysis alone to validate the model?
Building It Right in Excel
Mechanically, a DCF sensitivity table is just a two-way Data Table in Excel, with WACC down one axis and terminal growth rate across the other, and enterprise value as the output formula in the corner. If you need a refresher on the exact steps for building one, along with one-way tables and Scenario Manager for coherent bear and bull cases, see our full walkthrough on sensitivity and scenario analysis in Excel.
If your formulas in the sensitivity grid aren't updating when you change an input, that's usually a calculation mode issue rather than a modeling one. Our guide on why Excel formulas stop updating covers the most common causes, including a setting specific to Data Tables that catches people off guard.
Key Takeaways
- A sensitivity table can look completely clean and professional while still being built on a flawed foundation. None of these seven mistakes throw an Excel error.
- The single biggest issue is treating correlated variables as if they move independently, which understates how much a valuation actually shifts in a realistic scenario.
- WACC and terminal growth rate deserve to be your primary sensitivity pairing in most models, since terminal value usually drives the majority of total valuation.
- A sensitivity table shows how your model responds to different assumptions. It doesn't tell you which assumption is most likely, and it doesn't replace comparing your output to market prices or comparable companies.
- Fewer, cleaner tables with a clear narrative communicate more than a wall of grids ever will.
Frequently Asked Questions (FAQs)
What's the most common mistake in a DCF sensitivity table?
Treating each input variable as if it moves independently, when in reality several key drivers, like growth, margins, and capital needs, tend to move together. A sensitivity table that only flexes one variable at a time can understate how much a valuation actually shifts under a realistic downside or upside case.
Which two variables should a DCF sensitivity table usually focus on?
Discount rate (WACC) and terminal growth rate are the most common two-way sensitivity pairing, since together they usually drive the majority of the swing in enterprise value, especially in models where terminal value makes up a large share of total value.
How wide should the ranges be in a sensitivity table?
Wide enough to reflect genuine uncertainty, but not so wide that the table implies outcomes nobody actually believes are plausible. Ranges that are too narrow hide real risk, and ranges that are too wide make the table look alarmist and get dismissed by readers.
Is a high terminal value percentage always a red flag in a DCF?
Not automatically. A terminal value between roughly 60% and 75% of total DCF value is considered fairly normal, since most of a company's value legitimately sits far in the future. A terminal value above about 85% of total value is worth double-checking, since it often signals overly aggressive long-term growth or margin assumptions.
Should sensitivity analysis replace comparing a DCF to market prices or comparable companies?
No. Sensitivity analysis tells you how your own model responds to changing assumptions, but it doesn't tell you whether your assumptions are reasonable in the first place. Comparing your DCF output to comparable company multiples or the market price is a separate, complementary check that a sensitivity table can't substitute for.
Related Articles
External References
- Wall Street Prep. "Common Errors in DCF Models." wallstreetprep.com
- Phoenix Strategy Group. "5 Common Errors in DCF Models and How to Fix Them." phoenixstrategy.group
- ModelReef. "DCF Sensitivity Analysis: Two-Way Tables (WACC vs. Terminal Growth)." modelreef.io
0 Comments