On this page
A top-down revenue forecast estimates the size of a market and the company’s share of it. A bottom-up forecast builds revenue from quantities, prices, capacity or reporting segments. When the results differ, the useful answer is to expose the implied market share and operating assumptions, then investigate the gap before choosing a forecast.
What each route actually measures
The top-down route starts outside the company. An analyst may forecast an industry’s total revenue and then ask how much the company could capture. The market can be estimated from economic activity, industry growth, customer spending or other relevant sector measures. The final company revenue is the market size multiplied by an assumed market share. That share needs a defensible basis: competitive position, distribution, capacity and evidence of customer demand, rather than a convenient percentage chosen to make the model work.
The bottom-up route starts with the company’s operating units. For a retailer, it might use stores, transactions per store and average selling price; for a manufacturer, available production and expected utilisation can form the volume estimate. A company with several segments may need separate units and price assumptions for each. These inputs can be more concrete than a single growth rate, but their apparent precision does not make them reliable if the source data are incomplete or management’s expansion plan is uncertain.
CFA Institute’s Company Analysis: Forecasting overview says revenue can be projected through either top-down or bottom-up forecast objects and that different objects can reveal hidden assumptions or errors. It identifies market growth and market share as top-down drivers, and volume, price, segment, capacity and return-based measures as bottom-up drivers. Those are categories of model inputs, not a rule that one method always wins.
This is a company analysis task, not an instruction to trade. The CFA® programme page places the topic in the wider curriculum; the how it works page provides the study route. For a time-poor candidate, the highest-value practice is to make the two methods answer the same revenue question for the same period and company scope.
Build one illustrative forecast from both ends
Assume an invented regional market with £500 million of current annual sales. For the next year, suppose the analyst estimates market growth of 4%, so the market reaches £520 million. Suppose the company is expected to hold 10% of that market. The top-down company revenue forecast is therefore £52 million. All figures in this example are invented for arithmetic practice, not observed market data or a claim that these rates are typical.
The bottom-up view uses the same next-year period. Suppose the company expects 20 shops and forecasts 100,000 paid transactions per shop at an average realised price of £25 per transaction. The resulting forecast is £50 million. The transaction measure must match the price measure: if transactions include returns, cancellations or free services while average price excludes them, multiplication will be misleading. If a shop opens halfway through the year, treating it as a full-year shop would also overstate revenue.
The two models are £2 million apart. One might be tempted to take their midpoint. That does not explain why they differ. Divide the £50 million bottom-up forecast by the £520 million estimated market: it implies a share of about 9.62%, below the 10% used by the top-down model. The difference can be framed as a specific question: is the company likely to gain the sales needed to reach 10%, or is the operating plan closer to 9.62%? Rounding is shown for explanation; the exact quotient is about 0.09615.
The same bridge can run the other way. If 10% share is supported, the company needs £52 million of revenue. At £25 per transaction across 20 full-year shops, it needs 104,000 transactions per shop. That is 4,000 more than the bottom-up assumption. Now the disagreement has an operational interpretation. A candidate can ask whether capacity, customer traffic, opening hours and conversion support that extra volume. This is more informative than saying one forecast is “more conservative”.
Make the scopes and units comparable
Reconciliation fails when the models define the addressable market differently. A top-down market may include online sales, service revenue or competitors in a geography that the company’s bottom-up plan excludes. One method may use gross merchandise value while the other uses recognised revenue. A company with wholesale and retail channels might record different revenue per physical unit in each. Before changing a growth rate, put both forecasts on the same currency, accounting basis, time window and segment perimeter.
Then inspect price and volume separately. A higher average price might reflect an actual price rise, a richer product mix or fewer discounts. Those mechanisms have different consequences for demand and gross margin. Likewise, rising store count does not guarantee proportionate sales if new shops open late, attract fewer customers or take sales from existing shops. A model that hides all of these drivers in one “sales growth” cell is hard to test.
CFA Institute’s overview also distinguishes a forecast object from a forecast approach. A forecast object is what you project, such as market share or unit volume. An approach is how you obtain the assumption, such as extrapolating history, using a base rate, applying management guidance or making a documented judgement. A top-down market-share object can still use historical evidence; a bottom-up volume object can still include management’s planned expansion. Treating these pairs as synonyms makes it harder to diagnose where uncertainty entered the model.
Use the forecast gap to test the full financial model
Revenue does not stand alone. A sales forecast that implies rapid volume expansion may require inventory, working capital and investment in capacity. A price-led forecast may change margins differently from a volume-led forecast. CFA Institute notes that operating expense forecasts should be coherent with revenue forecasts, and that working capital and capital expenditure assumptions are linked to the activity expected. If one revenue route requires much more capacity than the existing asset base could provide, the model needs an explicit investment plan or a lower revenue estimate.
For a short horizon, detailed unit data may be relatively informative when the business reports them consistently. For a long horizon, it may be impossible to forecast every shop or customer cohort with useful precision, so market size and sustainable share can act as a constraint. Conversely, a broad industry growth assumption may miss a company-specific product launch or a shrinking store estate. The sensible mix depends on the business model and information available, not on a fixed hierarchy of methods.
Scenario analysis helps when the disagreement reflects genuine uncertainty rather than a spreadsheet error. In the invented example, one scenario might use 100,000 transactions per shop while another tests the 104,000 needed to support 10% market share. The analyst should also ask what changes in pricing or cost would accompany the higher traffic. The result is a range tied to identifiable drivers, rather than an arbitrary band around a point estimate.
A decision routine for the exam
- Identify the requested forecast period, revenue definition and company perimeter before calculating.
- For top-down, state market size, market growth and company share separately.
- For bottom-up, identify units, activity per unit and realised price, or the relevant segment-level equivalents.
- Convert one answer into the other method’s implied assumption, such as implied share or required transactions.
- Investigate material differences in scope, timing, capacity, prices and costs before accepting a number.
The common mistake is to assume that a detailed bottom-up spreadsheet must be right, or that an industry forecast is automatically safer because it starts with a large data set. Both can be internally consistent and still rely on a weak assumption. CFA Institute’s Practical Macro module also connects macro analysis with company earnings forecasts, which reinforces the value of checking company drivers against the broader setting. The exam skill is to locate the assumption that makes the two views disagree and explain what evidence would resolve it.
Sources
- Company Analysis: Forecasting | CFA Institute · cfainstitute.org · Retrieved
- Practical Macro | CFA Institute · cfainstitute.org · Retrieved



