How to Forecast Cash Flows in a DCF

Forecasting cash flows in a DCF means projecting revenue forward, applying an operating margin to get EBIT, then subtracting cash taxes, adding back depreciation, and subtracting capital expenditures and the change in net working capital to reach unlevered free cash flow for each year of the explicit forecast period. Every input traces back to revenue, since margins, capex, and working capital are all typically modeled as a percentage of it. Wall Street Prep frames the goal simply: forecasts "should ideally come from a 3-statement model" so revenue, the income statement, and the balance sheet move together instead of being guessed independently. Get this step wrong and every number downstream, from enterprise value to the implied share price, inherits the error.

TL;DR

Revenue drives the forecast; Street of Walls' worked model uses a flat 10 percent annual growth rate tied to historical trends.
SG&A held at 14.5 percent of revenue and D&A at 5.1 percent of sales are the concrete margin assumptions from that same model.
Capex should stay ahead of D&A as a percent of revenue when the business is growing, per Mergers & Inquisitions.
Net working capital is usually forecast as a percent of sales, since most operating working capital items move with revenue.
The forecast period is typically 5 years (CFI) up to 10 years, long enough for the business to reach a steady state.

What does it mean to forecast cash flows in a DCF?

Forecasting cash flows means turning a handful of assumptions, mainly revenue growth, margins, capex, and working capital, into a year-by-year unlevered free cash flow figure for the explicit forecast period. This is the step that sits between "walk me through a DCF" at a conceptual level and an actual spreadsheet. You are not forecasting free cash flow directly; you are forecasting the drivers that produce it, because free cash flow itself has no independent momentum. Corporate Finance Institute's DCF guide puts the forecast period at "approximately five years" for most businesses, with longer windows reserved for resource-intensive industries. The output of this step feeds directly into the unlevered free cash flow figure that gets discounted at WACC in the rest of the model.

How do you forecast revenue growth?

You forecast revenue first because nearly every other line item in the model is built as a percentage of it. There are two standard approaches. A growth-based forecast applies a single year-over-year growth rate and works best for stable, mature businesses. A driver-based forecast breaks revenue into components like units, price, and customer count, then builds each up separately, which is more granular but requires more data. Street of Walls' worked DCF model uses the simpler growth-based approach: it projects "revenue at an annual growth rate of 10%, which is in-line with historical growth rates," and recommends checking that assumption against management guidance, sell-side estimates, and the company's own trailing performance.

$Revenue_t = Revenue_{t-1} \times (1 + g)$

Here g is the assumed annual growth rate. Applying a flat 10 percent to a starting revenue base of 500 million dollars produces 550 million in year one, 605 million in year two, and 665.5 million in year three. The single biggest mistake at this step is picking a growth rate with no anchor. Interviewers expect you to justify it against a real reference point, whether that is the company's own three-year historical CAGR or the growth rate of its end market.

How do you forecast operating margin and EBIT?

You forecast operating margin by holding key expense lines constant, or gradually improving them, as a percentage of the revenue you just projected, then subtract those expenses from revenue to reach EBIT. Street of Walls' model keeps SG&A "constant as a percentage of Revenue (14.5%)" and lets COGS decline slightly as a share of revenue to reflect economies of scale as the business grows. The mechanical build is straightforward once the percentages are set.

$EBIT_t = Revenue_t - COGS_t - SG\&A_t$

Using the 550 million dollar year-one revenue figure and the 14.5 percent SG&A assumption, SG&A itself works out to about 79.75 million dollars for that year. If COGS is assumed at 60 percent of revenue, COGS is 330 million, which leaves EBIT of roughly 140.25 million dollars, an EBIT margin near 25.5 percent. The point of holding margins as ratios rather than flat dollar figures is that they scale automatically as revenue moves, which is exactly what a real operating business does.

How do you forecast depreciation and capex?

You forecast depreciation and amortization (D&A) and capital expenditures (capex) as percentages of revenue too, but the relationship between the two percentages matters more than either number alone. Street of Walls' worked model sets D&A at "5.1% of Sales." Mergers & Inquisitions frames the capex side of the relationship directly: "CapEx as a percentage of revenue stays ahead of D&A as a percentage of revenue in each year because [the company's] cash flows are growing," which keeps net property, plant, and equipment rising to support that growth. For a mature, stable company the two percentages should converge over time; for a growing one, capex should sit above D&A.

Forecast driver	Typical basis	Example assumption
Revenue growth	Historical CAGR or driver-based build	10 percent per year (Street of Walls)
SG&A	Percent of revenue, held constant	14.5 percent of revenue
D&A	Percent of sales	5.1 percent of sales
Capex	Percent of revenue, above D&A when growing	Slightly above the D&A percentage
Net working capital	Percent of sales, or turnover ratios	Change modeled as percent of revenue growth

Applying 5.1 percent of the 550 million dollar year-one revenue gives D&A of about 28.05 million dollars. If capex is assumed at 6 percent of revenue to stay ahead of D&A, capex comes out to 33 million dollars for the same year, a gap of roughly 4.95 million that reflects real net investment in the asset base.

How do you forecast net working capital?

You forecast the change in net working capital (NWC) as a percentage of revenue or revenue growth, because most operating working capital items, accounts receivable, inventory, and accounts payable, move with sales rather than independently. Street of Walls notes that "most OWC items are driven by Sales of the company," so a common shortcut is to hold NWC as a fixed percentage of revenue and let the change flow from the change in revenue itself. Mergers & Inquisitions adds a practical caveat: for many companies working capital changes are "simple timing differences" and not a major driver of the valuation, so it is fine to keep this assumption simple unless the business has unusual receivables or inventory dynamics.

$\Delta NWC_t = NWC\%_{\ of\ revenue} \times (Revenue_t - Revenue_{t-1})$

If NWC is assumed at 15 percent of revenue, the increase in revenue from 500 million to 550 million dollars, 50 million, implies an NWC increase of 7.5 million dollars that year. That 7.5 million is cash tied up in operations rather than available to investors, so it gets subtracted in the free cash flow build.

How do you put it all together into free cash flow?

You put the pieces together by taking EBIT, taxing it, adding back D&A, then subtracting capex and the change in NWC, which is the same unlevered free cash flow formula used in walk me through a DCF.

$FCF_t = EBIT_t \times (1 - T) + D\&A_t - CapEx_t - \Delta NWC_t$

Continuing the year-one numbers at a 25 percent tax rate: EBIT of 140.25 million after-tax becomes 105.19 million (140.25 x 0.75), plus D&A of 28.05 million, minus capex of 33 million, minus the NWC increase of 7.5 million, lands at approximately 92.74 million dollars of unlevered free cash flow for that year. Repeat the same chain, revenue growth, margin, D&A, capex, NWC, for each year of the explicit forecast period, and the result is the stream of cash flows that gets discounted at WACC before a terminal value captures everything beyond it.

Frequently Asked Questions

What is the most common mistake when forecasting DCF cash flows?

Projecting revenue growth without a matching increase in capex or working capital. If the model assumes 10 percent revenue growth but holds capex or NWC flat, it implies the business grows for free, which does not happen in the real world. Every dollar of growth requires some reinvestment, so capex and working capital should scale with revenue, not sit static.

Should you forecast capex and D&A to be equal?

Only for a mature, low-growth company. For a growing business, capex should run ahead of D&A as a percentage of revenue, since the company is adding net new assets, not just replacing depreciated ones. Mergers & Inquisitions notes that capex staying above D&A is exactly what keeps net PP&E rising to support growth; the two percentages tend to converge only once growth slows.

How long should the DCF forecast period be?

Typically five years, per Corporate Finance Institute's default guidance, extending to ten years for companies that need more time to reach a stable, mature growth rate. Resource-intensive industries sometimes use even longer horizons tied to the life of the underlying asset. The forecast should run long enough that the business is growing at a steady, sustainable rate by the final year, since that steady state is the assumption baked into the terminal value.

Is a growth-based or driver-based revenue forecast better?

It depends on the business and the data available. A growth-based forecast, a single year-over-year growth rate, is simpler and works well for stable, mature companies. A driver-based forecast, building revenue from units, price, customers, or market share, is more granular and more defensible but requires more detailed data. Interviewers generally accept a growth-based forecast in a timed setting as long as the growth rate is justified against a real reference point.

Why do you use percent-of-revenue assumptions instead of flat dollar amounts?

Because most cost lines and working capital items scale with the size of the business, not on a fixed schedule. Holding SG&A, D&A, capex, and NWC as percentages of revenue means the whole model automatically adjusts when the revenue assumption changes, which keeps the forecast internally consistent. A flat dollar forecast would break that link and require manually re-deriving every line every time the growth assumption changes.

What happens if working capital is ignored in the forecast?

Ignoring the change in net working capital overstates free cash flow, because cash tied up funding receivables and inventory growth would never get subtracted. For a fast-growing company this can be a meaningful overstatement, since working capital needs typically grow alongside revenue. Even when the effect is modest, per Mergers & Inquisitions "simple timing differences," leaving it out entirely is a mechanical error interviewers will catch.

Sources

Wall Street Prep, "DCF Model Training: 6 Steps to Building a DCF Model in Excel": https://www.wallstreetprep.com/knowledge/dcf-model-training-6-steps-building-dcf-model-excel/ (checked July 2026)
Mergers & Inquisitions, "DCF Model: Full Guide, Excel Templates, and Video Tutorial": https://mergersandinquisitions.com/dcf-model/ (checked July 2026)
Street of Walls, "Discounted Cash Flow Analysis": https://www.streetofwalls.com/finance-training-courses/investment-banking-technical-training/discounted-cash-flow-analysis/ (checked July 2026)
Corporate Finance Institute, "DCF Model Training: The Ultimate Free Guide to DCF Models": https://corporatefinanceinstitute.com/resources/financial-modeling/dcf-model-training-free-guide/ (checked July 2026)