Every financial forecast starts with revenue, and most of its value depends on getting revenue and costs roughly right. A forecast that grows every line by the same percentage looks tidy but explains nothing. Good forecasting starts from history, builds on the real drivers of the business — customers, volumes, prices, capacity, cost behaviour — and is tested against common sense and market evidence.
What you will learn
- Use historical analysis as the starting point for forecasts
- Compare top-down and bottom-up revenue forecasting
- Build driver-based revenue and cost forecasts
- Apply sanity checks to keep forecasts realistic
Start with history
Before forecasting, understand the past three to five years:
- Growth rates of revenue by product, region and customer group.
- Margins and cost ratios from a common-size analysis — see horizontal and vertical analysis.
- Seasonality — which months or quarters are strongest.
- One-off items to exclude — unusual contracts, asset sales, restructuring costs.
- Relationships between drivers and results: how revenue responded to price changes, how costs moved with volume.
History is a guide, not a rule. Ask what will be different in future: new competitors, capacity limits, regulation, economic conditions.
Top-down and bottom-up forecasting
| Approach | Method | Strengths | Weaknesses |
|---|---|---|---|
| Top-down | Market size × expected market share | Links to the market; good for new ventures | Market data may be unreliable; share assumptions can be optimistic |
| Bottom-up | Build from units, customers, prices, capacity | Grounded in operations; easy to test | Can miss market constraints; detailed |
Using both and comparing the results is a strong discipline. If a bottom-up plan implies capturing 40% of the market within three years, the plan needs a very convincing explanation.
Driver-based revenue forecasts
Revenue can usually be broken into a few drivers:
- Volume × price — manufacturers and retailers.
- Customers × average revenue per customer — subscription and service businesses.
- Capacity × utilisation × price — hotels, airlines, hospitals, factories.
- Outlets × revenue per outlet — branch-based businesses.
- Loan book × yield — lenders.
Worked example: forecasting revenue and operating profit
A subscription business has 2,000 customers at an average of 1,500 a month. Assumptions: customers grow 10% a year, price rises 3% a year, variable costs are 40% of revenue, and fixed costs are 18,000,000 in year 1, rising 5% a year (illustrative figures).
| Year 1 | Year 2 | Year 3 | |
|---|---|---|---|
| Customers | 2,000 | 2,200 | 2,420 |
| Monthly price | 1,500 | 1,545 | 1,591 |
| Revenue | 36,000,000 | 40,788,000 | 46,212,804 |
| Variable costs (40%) | 14,400,000 | 16,315,200 | 18,485,122 |
| Fixed costs | 18,000,000 | 18,900,000 | 19,845,000 |
| Operating profit | 3,600,000 | 5,572,800 | 7,882,682 |
| Operating margin | 10.0% | 13.7% | 17.1% |
Margins rise because fixed costs grow more slowly than revenue — the effect of operating leverage. Is that realistic? Serving 21% more customers might need more staff or systems, so fixed costs may step up. Checking such implications is where forecasting judgement lies.
Forecasting expenses
Classify costs by how they behave — the same thinking used in cost and management accounting:
- Variable costs — forecast as a percentage of revenue or as cost per unit (materials, commissions, delivery).
- Fixed costs — forecast as amounts, adjusted for inflation and known changes (rent, salaried staff, insurance).
- Step costs — fixed within a range, then jump (a new shift, a new warehouse).
- Headcount-driven costs — number of staff × average cost, with planned hiring and pay increases.
- Discretionary costs — marketing, training — set by management decisions.
- Depreciation — from the fixed asset schedule, not a percentage of revenue; see debt and depreciation schedules.
External inputs
Revenue and costs depend on the economy: demand grows with incomes, costs with inflation and exchange rates. Link key assumptions to macroeconomic forecasts where relevant — see economic indicators. Elasticity estimates help translate price changes into volume changes, and imported input costs need an exchange rate assumption.
Sanity checks
- Growth versus history and peers — is growth far above anything achieved before or by competitors?
- Market share implied — does the forecast imply an unrealistic share?
- Capacity — can the business physically deliver the volumes?
- Margins — do they converge to levels sustainable in a competitive market?
- Working capital and investment — does growth come with the inventory, receivables and capital expenditure it needs? A forecast with rapid growth and no extra investment is inconsistent.
- Cash — does the business need funding to achieve the plan?
Presenting forecasts honestly
Forecasts are uncertain. Present a base case with clearly stated assumptions, along with downside and upside cases, and show which assumptions matter most through sensitivity analysis. Record the sources of assumptions so they can be reviewed later against actual results.
Common mistakes
- Growing every line by the same percentage.
- Ignoring seasonality in monthly forecasts.
- Assuming margins rise indefinitely as revenue grows.
- Forecasting growth without the investment and working capital it requires.
- Confusing targets with forecasts — a target is what management wants; a forecast is what is expected.
Key terms
- Driver — an underlying quantity that determines revenue or cost.
- Top-down forecast — revenue estimated from market size and share.
- Bottom-up forecast — revenue built from operational units such as customers and prices.
- Step cost — a cost fixed within a range of activity that jumps beyond it.
- Base case — the forecast using the most likely assumptions.
Frequently asked questions
How far ahead should revenue be forecast?
Operational budgets usually cover one year by month; strategic plans three to five years; project and valuation models the full investment horizon.
Should I use the historical average growth rate?
Use it as a reference, but adjust for known changes in the market, capacity and strategy.
How do I forecast a new business with no history?
Combine top-down market analysis with bottom-up operational drivers, use data from comparable businesses, and present wide scenarios.
Key takeaways
- Start from a careful analysis of history, excluding one-off items.
- Build revenue from drivers such as volume and price or customers and average revenue.
- Forecast costs according to their behaviour: variable, fixed, step and discretionary.
- Test forecasts against market size, capacity, margins, investment needs and cash.
Explore this in ARBIMETRA LAB
Apply these ideas to your own data. Files are analysed in your browser and are not uploaded.
- Time Series Decomposition: Trend, seasonal and residual parts.
- Moving Average: Simple and weighted moving averages.
- Autoregressive Model (AR): AR(p) estimation and forecasts.
- Moving Average Model (MA): MA(q) estimation and forecasts.
