The core modeling conventions and why they matter
| Convention | Rule | Why |
|---|---|---|
| Input separation | All assumptions in dedicated input cells (blue font) | Change one number, model updates everywhere |
| No hardcoded numbers | Every number in a formula links to an input or prior calc | Prevents ‘magic numbers’ that break auditing |
| Consistent sign convention | Revenue positive, expenses negative (or vice versa) | Prevents sign errors that flip conclusions |
| Error checks | Balance sheet balances, cash flow reconciles | Catches formula errors before they compound |
| Scenario toggle | Switch between base/bull/bear with one input | Easy stress-testing |
Breaking the debt-interest circular reference
The most common modeling error: circular references between debt interest and cash flow. Interest depends on debt balance, which depends on cash flow, which depends on interest. Use an iterative solver or break the circularity with prior-period debt.
Reconcile net income to free cash flow
Review a company’s financials in Fundamentals. Try to reconcile: does Net Income + D&A − CapEx − ΔWC approximately equal FCF? This is the basic check that catches most data errors.
What an unbalanced balance sheet signals
Your model’s balance sheet doesn’t balance — assets exceed liabilities + equity by $12M. What should you do?
Why right answers from wrong formulas are dangerous
Every assumption must trace to a driver
You've built a 200-tab Excel LBO model. Senior PM questions your Year 3 revenue: 'Why 8.5% growth vs your peer 6.2% guidance?' You reply 'The model says so.' What went wrong?
Check your understanding
Sit with the ideas.
You receive a colleague's DCF model. Revenue in Year 3 is hardcoded (not linked to the growth rate assumption). The balance sheet does not balance by $3M. There is no sensitivity table. The terminal growth rate is 4% (above long-term GDP growth). How many red flags are there, and which is most dangerous?
Why: