This guide covers how to structure one correctly — the tab layout, the key formulas, and where automation saves time on the maintenance work that compounds over months.
The Tab Structure That Actually Works
A model built for investor review typically needs five tabs in this order.
Assumptions is where every input lives — growth rates, headcount ramp, churn, pricing, COGS as a percentage of revenue. Nothing gets hardcoded in the P&L or balance sheet. One cell changes here; everything downstream updates. If your model has hardcoded values scattered across tabs, someone will miss one during an update and produce wrong output.
P&L runs monthly for years 1–2, quarterly for years 3–5. It pulls from Assumptions using named ranges and calculates gross profit, EBITDA, operating income, and net income. The monthly columns matter: investors want to see how burn evolves month by month, not just annual totals.
Balance Sheet ties to the P&L via retained earnings. Ending retained earnings on the income statement must equal the equity section of the balance sheet — if they don't reconcile, something is wrong, and you want to catch it before the investor does.
Cash Flow uses the indirect method: start with net income, add back non-cash items (D&A, stock compensation), adjust for working capital movements, then show investing and financing activities separately. This is where models usually break — working capital schedules get detached from the actual balance sheet movements.
Scenarios holds the sensitivity analysis: a base/bear/bull toggle and two-variable tables showing how valuation or runway changes across combinations of revenue growth and margin assumptions. Investors ask about downside scenarios in nearly every first meeting.
Building the P&L
Start with revenue. For a SaaS company with monthly cohort data:
=Assumptions!B5 * (1 + Assumptions!B6)
Where B5 is prior period ARR and B6 is the monthly growth rate. Use named ranges once your model grows past a few hundred rows — MoM_Growth_Rate is easier to audit than Assumptions!B6 when you're tracing a formula six months later.
COGS as a percentage of revenue:
=Revenue * Assumptions!COGS_Pct
Operating expenses break into categories: headcount (from a separate hiring schedule), software, marketing spend, and G&A. Headcount is the most work to build correctly and the most-scrutinized line in an early-stage model, so it deserves its own tab.
The Hiring Plan Schedule
Create a separate Headcount tab with one row per role: role title, department, start month, monthly salary, and a benefits load factor. For US-based employees, a load factor of 1.20–1.25x is standard — that covers employer payroll taxes (FICA runs 7.65%), health insurance, and 401(k) match. International employees differ materially: UK employer National Insurance contributions are 13.8%, Singapore CPF employer contributions are 17% for employees under 55.
In the P&L, sum salaries for active headcount using SUMPRODUCT:
=SUMPRODUCT(
(Headcount!C2:C50 <= Current_Month) *
Headcount!D2:D50 *
Headcount!E2:E50
)
This calculates total monthly salary cost for roles whose start month falls on or before the current period. It handles staggered hiring ramps without requiring a separate formula column per role — which breaks every time you add a row.
Wiring the Balance Sheet
The balance sheet is where most models break down. The most common mistake is calculating ending cash as a plug — whatever number makes assets equal liabilities plus equity. The correct approach:
- Cash flow statement calculates net cash change for the period
- Ending cash = prior period ending cash + net cash change
- Balance sheet pulls ending cash from the cash flow statement
Put a tie-out check somewhere visible on the balance sheet:
=IF(
ABS(Total_Assets - (Total_Liabilities + Total_Equity)) < 0.01,
"✓ Ties",
"⚠ Off by " & TEXT(ABS(Total_Assets - (Total_Liabilities + Total_Equity)), "$#,##0")
)
A tolerance of $0.01 handles rounding from percentage-based calculations across 60 monthly columns.
Sensitivity Analysis
As of March 2026, Google Sheets has no native equivalent to Excel's Data Table function in What-If Analysis. According to Google's Sheets documentation, the TABLE() function is not supported. The standard workaround is a CHOOSE() function tied to a scenario selector dropdown:
=CHOOSE(Scenario_Selector, Bear_Revenue, Base_Revenue, Bull_Revenue)
Where Scenario_Selector is a data-validated dropdown cell containing 1, 2, or 3. All downstream formulas reference this function rather than hardcoded values, so switching scenarios updates the entire model instantly.
For a true two-variable sensitivity table — say, revenue growth rate vs. gross margin — you'll need to either build it manually (filling each cell with a reference to overridden assumptions) or write an Apps Script function to sweep values and populate the output range. Neither is elegant, but the manual approach is easier to audit.
Circular References and Why They Break Models
Google Sheets disables iterative calculation by default. If your interest expense references ending debt balance, which references net cash flow, which references interest expense — you'll get a circular reference error rather than a calculated result.
The most common fix: break the circularity by calculating interest on the prior period's debt balance rather than the ending balance. This introduces a one-period lag that's negligible for annual models and acceptable for monthly ones.
If you need true iterative calculation, enable it in Settings > Calculation > Iterative Calculation, with a maximum of 30–50 iterations. Enable it deliberately and document it in the assumptions tab — otherwise the next person touching the model will be confused about why certain cells behave differently.
Google Sheets caps each spreadsheet at 10 million cells. A 5-year monthly model with 15 tabs and 500 rows per tab is roughly 450,000 cells — well within limits. Performance issues are more likely to come from volatile functions like NOW() or INDIRECT() triggering recalculation on every edit than from raw cell count.
Automating the Maintenance Work
Building the initial structure takes 3–4 hours for someone who's done it before. The ongoing work — updating actuals each month, extending projections when assumptions change, reformatting output for a board deck, rebuilding sensitivity tables when the key drivers shift — is what actually accumulates.
Most of that work follows predictable patterns: extend date headers by one column, copy formula patterns down new rows, replace projected values with actuals in the right cells. It's low cognitive effort but high time cost, especially in models with 20+ tabs and linked sheets.
ModelMonkey handles this category of work within Google Sheets. Rather than manually extending 12 months of headcount projections when three roles are added, you describe the change and it updates the affected sections directly. It reads your existing model structure — your named ranges, your tab layout, your formula patterns — rather than generating a generic template overlay.
It's not useful for the initial structural decisions, which require judgment about how to model your specific business. It's useful for the repetitive maintenance that follows.
What Investors Actually Review
Due diligence on a financial model focuses on three things: how revenue growth is calculated and what assumptions drive it, the path to profitability and how sensitive it is to key assumptions, and cash runway — how many months until the next raise, and what that assumes about burn.
A well-structured model answers all three without needing a guided tour. If an investor has to ask "where does this growth rate come from," the assumptions tab isn't pulling its weight.
The model should also survive basic stress-testing: what happens to runway if growth comes in 20% below plan? If gross margin compresses by 5 points? If headcount ramp slips by a quarter? Investors will run these scenarios themselves; having them pre-built signals that the founder has thought through the downside.