Forecasting & Earnings Models
Revenue Forecasting
Revenue forecasting is the foundation of any earnings model. All other financial projections — costs, margins, earnings, and cash flow — are typically derived from or anchored to revenue estimates. The accuracy of a research analyst's revenue forecast directly determines the quality of the entire financial model. There are two primary approaches to revenue forecasting: top-down and bottom-up.
Top-Down Revenue Forecasting
The top-down approach starts with macroeconomic or industry-level data and works down to the company level. The analyst begins with total addressable market (TAM) size, applies an expected market growth rate, estimates the company's market share, and derives revenue. This approach is useful for understanding a company's position within its industry and for validating bottom-up estimates.
Example: The U.S. cybersecurity market is estimated at $80 billion, growing 12% annually. Company X has a 5% market share. Top-down revenue estimate: $80B x 1.12 x 5% = $4.48 billion. If the company is gaining market share, the analyst might assume 5.5% market share, yielding $4.93 billion.
Bottom-Up Revenue Forecasting
The bottom-up approach builds revenue from the company's individual products, segments, geographies, or customer cohorts. This method is more granular and typically more accurate because it captures the specific drivers of the company's business. Common bottom-up frameworks include:
- Price x Volume: For companies selling discrete products. Revenue = units sold x average selling price (ASP). The analyst forecasts unit growth and pricing trends separately.
- Subscribers x ARPU: For subscription businesses (SaaS, telecom, media). Revenue = number of subscribers x average revenue per user. The analyst forecasts subscriber additions, churn, and ARPU trends.
- Same-store sales + new stores: For retail businesses. Revenue = existing store revenue growth (same-store sales) + revenue from new store openings. The analyst forecasts comparable store growth and the pace of new store rollouts.
- Backlog/bookings conversion: For project-based businesses (defense, construction, enterprise software). Revenue is derived from existing backlog and expected new order intake, adjusted for conversion rates and timing.
Exam Tip
The Series 86 exam tests your ability to identify appropriate revenue drivers for different business models. For a SaaS company, the key drivers are subscriber count, churn rate, and ARPU. For a manufacturer, focus on unit volumes and average selling prices. For a bank, model net interest income (NII) based on loan balances, interest rates, and net interest margin. Always decompose revenue into its fundamental drivers rather than simply projecting a top-line growth rate.
Expense and Margin Forecasting
Once revenue is projected, the analyst builds out the expense structure to derive profitability estimates. Expenses can be forecast using several approaches, and the best method depends on the nature of each cost category.
Percentage of Revenue Method
Many expenses are forecast as a percentage of revenue based on historical trends and expected changes. COGS, SG&A, and R&D are commonly modeled this way. The analyst examines historical expense ratios, identifies trends (improving or deteriorating margins), considers management guidance on cost initiatives, and projects forward ratios accordingly.
Fixed vs. Variable Cost Analysis
Understanding the fixed-variable cost mix is essential for modeling operating leverage. Variable costs (raw materials, sales commissions, shipping) scale proportionally with revenue. Fixed costs (rent, base salaries, depreciation on existing assets) remain relatively constant regardless of revenue levels. Companies with high fixed costs exhibit high operating leverage — small changes in revenue lead to amplified changes in operating income.
Segment-Level Modeling
For diversified companies, modeling at the segment level provides more accurate results because different business segments may have different growth rates, margin profiles, and cost structures. Each segment should be modeled independently and then aggregated to produce consolidated financial projections. Corporate overhead and intercompany eliminations are applied at the consolidated level.
Example
Operating Leverage Model: A company has $500M revenue, $200M in variable costs (40% of revenue), and $200M in fixed costs. Operating income = $100M (20% margin). If revenue grows 10% to $550M: variable costs = $220M (40%), fixed costs remain $200M, operating income = $130M (23.6% margin). A 10% revenue increase drove a 30% increase in operating income. This illustrates operating leverage — understanding this concept is essential for accurate earnings modeling.
Building an Integrated Earnings Model
A complete earnings model integrates the income statement, balance sheet, and cash flow statement into a unified, internally consistent framework. The three statements are linked through several connections:
- Net income from the income statement flows to retained earnings on the balance sheet and is the starting point for operating cash flow
- Capital expenditures increase PP&E on the balance sheet and appear in investing cash flow
- Depreciation reduces PP&E on the balance sheet, is expensed on the income statement, and is added back in operating cash flow
- Debt issuance/repayment changes the debt balance on the balance sheet and appears in financing cash flow
- Working capital changes (receivables, inventory, payables) link balance sheet changes to operating cash flow adjustments
- Interest expense on the income statement is calculated from the debt balance on the balance sheet
A properly integrated model will automatically balance (assets = liabilities + equity), and any changes to revenue assumptions will cascade through all three statements. Building and maintaining integrated models requires discipline and attention to detail, but it produces the most reliable projections and allows for comprehensive scenario analysis.
| Forecast Approach | Best Used For | Advantages | Limitations |
|---|---|---|---|
| Top-Down | Market sizing, industry context | Captures macro trends, good for validation | Less granular, may miss company-specific drivers |
| Bottom-Up | Company-specific projections | Granular, ties to specific business drivers | Data-intensive, may miss macro shifts |
| % of Revenue | Expense forecasting | Simple, leverages historical relationships | Assumes stable cost structure |
| Fixed/Variable | Operating leverage analysis | Captures leverage effects accurately | Requires detailed cost breakdown |
Scenario and Sensitivity Analysis
Scenario analysis involves defining distinct cases (base, bull, bear) with different sets of coherent assumptions. Each scenario tells a different "story" about how the future might unfold:
- Bull case: Favorable assumptions — strong revenue growth, margin expansion, successful new product launches, favorable industry dynamics. May assume best-case management execution and tailwinds from macro trends.
- Base case: The analyst's most likely scenario, reflecting balanced assumptions about growth, margins, and competitive dynamics. Should represent the analyst's highest-conviction estimate.
- Bear case: Adverse assumptions — revenue misses, margin compression, competitive losses, macro headwinds. Represents a realistic downside scenario (not the absolute worst case, which would be bankruptcy).
Sensitivity analysis varies individual assumptions while holding others constant to isolate the impact of each variable on the output. Common sensitivity variables include revenue growth rate, operating margin, WACC, terminal growth rate, tax rate, and capex assumptions. Results are typically presented in two-variable sensitivity tables (also called "data tables").
Key Takeaway
The difference between scenario analysis and sensitivity analysis is important: Scenario analysis changes multiple assumptions simultaneously to create coherent narratives (e.g., a recession scenario with lower revenue AND lower margins AND higher default rates). Sensitivity analysis changes one or two variables at a time while holding everything else constant to isolate individual impacts. Both are essential for communicating the range of outcomes and the key drivers of value.
Consensus Estimates and Earnings Surprises
The consensus estimate is the average (or median) of all published analyst estimates for a particular financial metric, most commonly EPS and revenue. Consensus estimates are compiled by data providers such as Bloomberg, FactSet, and Refinitiv and serve as the market's expectation against which actual results are compared.
An earnings surprise occurs when a company reports results that differ materially from the consensus estimate. Positive surprises (beats) typically drive the stock price higher, while negative surprises (misses) typically drive it lower. However, the magnitude and direction of the stock price reaction depends not just on the surprise relative to consensus, but also on several other factors including the quality and sustainability of the beat or miss, management's forward guidance relative to expectations, the "whisper number" (unofficial expectations among buy-side investors that may differ from published consensus), and the broader market and sector sentiment.
Research analysts seek to develop estimates that are more accurate than the consensus, creating an informational advantage for their clients. When an analyst's estimate differs significantly from consensus, they should clearly articulate the reasons for the divergence and the specific assumptions that drive the difference. This differentiated view is the primary source of an analyst's value to institutional investors.
Warning
Beware of "earnings management" — the practice of companies manipulating reported earnings to meet or slightly beat consensus estimates. Common techniques include adjusting revenue recognition timing, releasing or building reserves, changing depreciation estimates, or timing asset sales to generate one-time gains. Analysts should look beyond reported EPS and examine the quality and sustainability of earnings beats.
Check Your Understanding
Test your knowledge of forecasting and earnings models.
1. For a subscription-based SaaS company, the most appropriate bottom-up revenue model uses:
2. A company has $400M revenue, $160M variable costs, and $140M fixed costs. If revenue increases 15%, what is the new operating margin?
3. The difference between scenario analysis and sensitivity analysis is:
4. In an integrated three-statement model, net income connects to the balance sheet through:
5. An analyst's primary value to institutional investors comes from: