A project that returns $40,000 on a $30,000 investment can still destroy value. If that $40,000 arrives five years from now and your money could have earned 10 percent a year elsewhere, the future payoff is worth only about $24,800 in today's dollars — you paid $30,000 for something worth $24,800. The project looks profitable and quietly makes you poorer.
That gap between "the cash adds up" and "the deal is actually worth doing" is exactly what net present value closes. NPV translates every future dollar a project promises into today's dollars, adds them up, and subtracts what you pay today. A positive number means the project creates value after compensating you for waiting and for risk; a negative number means it destroys value, no matter how impressive the undiscounted total looks.
This guide works through the formula, runs a complete small-business example, covers the decision rule, shows how to pick your discount rate, compares NPV with IRR, gives you the Excel formula, and flags the five mistakes that most often corrupt the answer.
Why a Dollar Tomorrow Is Worth Less Than a Dollar Today
Discounting rests on three facts about money and time, each of which erodes the value of a future dollar:
Opportunity cost. Cash you tie up in a project cannot earn returns anywhere else. If your next-best use of funds earns 10 percent, then receiving $1,000 a year from now costs you the $100 that money would have earned in the meantime.
Risk. A dollar in hand is certain; a dollar promised next year is a forecast. Customers may not materialize, costs may overrun, the equipment may underperform. The further out the cash flow, the more can go wrong, so distant dollars get discounted more heavily.
Inflation. Rising prices shrink what each future dollar buys. At 3 percent inflation, a dollar received in ten years purchases roughly what 74 cents buys today.
Discounting folds all three into a single operation: divide each future cash flow by a compounding factor built from your discount rate. Money expected sooner keeps most of its value; money expected years out keeps much less.
The NPV Formula, Piece by Piece
The formula looks intimidating for about thirty seconds, then it clicks:
NPV = CF1 / (1 + r)^1 + CF2 / (1 + r)^2 + … + CFn / (1 + r)^n − Initial Investment
Each symbol has a plain meaning:
- CFt is the net cash flow you expect in period t — cash in minus cash out, after tax, in that year (or month, or quarter).
- r is the discount rate, also called the hurdle rate: the annual return your money must earn to justify the wait and the risk. More on choosing it below.
- (1 + r)^t is the compounding factor that shrinks each future dollar to its present value. Year 1 divides by (1 + r), year 2 by (1 + r) squared, and so on — the shrinkage accelerates with distance.
- Initial Investment is what you pay today, at time zero. It is not discounted because no waiting is involved.
Each fraction is the present value of one year's cash flow. Add them all, subtract today's cost, and the remainder is the value the project creates — or destroys — measured in today's dollars.
A Worked Example: The Bakery's Delivery Van
Suppose you own a bakery and are considering a $30,000 delivery van that would let you serve office catering accounts. You estimate the van generates $8,000 per year of incremental after-tax cash flow — new catering revenue minus fuel, insurance, maintenance, and the driver's wages — for five years, after which you sell the van for scrap value you treat as zero to stay conservative. Your discount rate is 10 percent.
Discount each year's $8,000 back to today:
| Year | Cash flow | Discount factor at 10% | Present value |
|---|---|---|---|
| 1 | $8,000 | 1 / 1.10 = 0.9091 | $7,273 |
| 2 | $8,000 | 1 / 1.10^2 = 0.8264 | $6,612 |
| 3 | $8,000 | 1 / 1.10^3 = 0.7513 | $6,011 |
| 4 | $8,000 | 1 / 1.10^4 = 0.6830 | $5,464 |
| 5 | $8,000 | 1 / 1.10^5 = 0.6209 | $4,967 |
| Total present value | $30,326 |
Subtract the $30,000 purchase price:
NPV = $30,326 − $30,000 = +$326
The project barely clears the bar: it creates about $326 of value in today's dollars. Note what the undiscounted math would have told you — five years times $8,000 is $40,000 against a $30,000 cost, a "profit" of $10,000. Discounting reveals that nearly all of that apparent profit is just compensation for waiting five years and bearing the risk. The deal is essentially break-even, and if your cash flow estimate is even slightly optimistic, it flips negative.
That fragility is itself useful information. Re-run the math at 12 percent — a rate reflecting a riskier outlook — and the present value drops to about $28,838, giving an NPV of roughly −$1,162. At 8 percent it rises to about $31,942, or +$1,942. One project, three verdicts, depending entirely on the rate. Which is why the most important input in the whole calculation is the one people spend the least time choosing.
The Decision Rule: Positive Means Go
The NPV decision rule is the simplest in finance:
- NPV above zero: accept. The project earns more than your required return, so it increases your wealth by exactly the NPV amount. A +$326 NPV means you are $326 richer in today's dollars for doing the deal.
- NPV below zero: reject. The project fails to clear your hurdle. Your money does better in the alternative the discount rate represents.
- NPV at zero: indifferent, in theory. The project earns exactly your required return. In practice, zero-NPV projects get rejected unless they carry strategic benefits the cash flows omit, because forecasts skew optimistic and there is no margin for error.
When ranking mutually exclusive projects — you can buy the van or renovate the storefront, not both — choose the highest NPV, not the highest return percentage. NPV measures absolute wealth created, so it already accounts for project scale. A small project earning 40 percent can create less wealth than a large project earning 12 percent, and NPV tells you so directly.
One caveat: NPV assumes you can fund every positive-NPV project. If capital is rationed — the bank approved one loan, not three — rank by NPV per dollar invested (the profitability index) instead of raw NPV, so limited funds flow to the highest value per dollar.
Choosing Your Discount Rate
The discount rate is your required return: the annual percentage a project must earn to justify tying up your money at that level of risk. Three practical ways to set it:
Start with opportunity cost. Ask what the money would otherwise do. Cash sitting in a business savings account earning 4 percent sets a floor — no project riskier than that account should be discounted lower. If your realistic alternative is paying down a 9 percent line of credit, 9 percent is your floor. For many small businesses, the alternative is reinvesting in the existing operation, whose historical return becomes the benchmark.
Use your cost of capital for bigger decisions. Textbooks prescribe the weighted average cost of capital (WACC): blend the after-tax interest rate on your debt with the return your equity demands, weighted by how much of each funds the business. WACC is the right hurdle for projects with roughly the same risk as your existing business; full WACC math is overkill for a van purchase, where a rounded opportunity-cost estimate is fine.
Adjust for project risk. The company-wide rate fits average-risk projects. A project riskier than your normal operations — a new product line, an unfamiliar market — deserves a higher rate, often 2 to 5 percentage points above base. A safer-than-average project, like replacing broken equipment with an identical model, can use a lower one. Using one flat rate for everything systematically overvalues risky bets and undervalues safe ones.
When in doubt, run the calculation at two or three rates, as the van example did. If the verdict survives across a reasonable range, you can trust it. If it flips, the project is marginal and the honest answer is "too close to call on the numbers alone."
NPV vs. IRR: Why Two Right Answers Can Disagree
The internal rate of return (IRR) is NPV's sibling metric: the discount rate that would make the project's NPV exactly zero. The van's IRR is about 10.4 percent — the rate at which its $8,000 annual flows exactly repay $30,000. The IRR rule says accept when IRR exceeds your hurdle rate. Here 10.4 beats 10, matching the positive NPV. Usually the two agree. When they conflict, trust NPV, for three reasons:
Conflicting rankings. For mutually exclusive projects of different sizes or timings, IRR can rank first the project that creates less wealth. A $1,000 project doubling your money (100 percent IRR) beats a $100,000 project earning 15 percent on IRR — but the second creates far more dollars. NPV counts dollars, so it ranks correctly.
The reinvestment assumption. The mathematics of IRR implicitly assumes cash the project throws off gets reinvested at the IRR itself — that your 40 percent project lets you reinvest every interim dollar at 40 percent indefinitely. NPV assumes reinvestment at the discount rate, typically your cost of capital, which is far more realistic. High IRRs flatter themselves through this assumption.
Multiple IRRs. Projects whose cash flows change sign more than once — an upfront cost, years of inflows, then a big cleanup or decommissioning cost at the end — can have two or more IRRs, or none at all. The rule "accept if IRR exceeds the hurdle" has no answer when there are three IRRs. NPV always produces exactly one number.
None of this makes IRR useless. A percentage travels well in a meeting, and IRR needs no discount rate input, which helps when the right rate is genuinely uncertain. Use IRR to communicate and sanity-check; use NPV to decide.
The Excel Formula — and Its One Big Gotcha
Excel's NPV function computes present values, but with a trap that has corrupted countless spreadsheets: it assumes the first cash flow you hand it arrives at the end of period 1, not today. The initial investment happens at time zero, so it must stay outside the function:
=NPV(10%, 8000, 8000, 8000, 8000, 8000) - 30000With cell references, if B2:B6 holds the five yearly flows and B1 holds the $30,000 outlay as a positive number:
=NPV(10%, B2:B6) - B1The classic error is including the upfront cost inside the range, which discounts the time-zero outlay by a full period and overstates NPV by roughly one year's discount on the whole investment. On the van example that mistake inflates NPV by over $2,700, turning a marginal deal into an apparent bargain.
For cash flows landing on irregular calendar dates, use XNPV(rate, values, dates) instead, which discounts each flow by its actual date — and unlike NPV, it takes the initial outlay inside the values range, since the dates tell it the outlay happens at time zero. Google Sheets mirrors Excel exactly, with the same NPV timing behavior and an equivalent XNPV.
Five Mistakes That Corrupt an NPV
Most bad NPVs come from bad inputs, not bad arithmetic. The repeat offenders:
1. Counting sunk costs. Money already spent — the $2,000 market study you commissioned last quarter — is gone whether you buy the van or not. Including it punishes the project for a cost the decision cannot recover and can kill deals that are genuinely worthwhile. Only incremental cash flows count: the difference between your cash with the project and without it. By the same token, do count opportunity costs — using a warehouse bay you already own still costs the rent you could have collected.
2. Discounting profit instead of cash. NPV runs on cash flows, never accounting profit. Depreciation is the wedge: it reduces reported profit but moves no cash, so add it back. Conversely, capital outlays and working-capital investments move cash without immediately hitting profit, so include them. A project showing steady accounting profit while consuming ever more receivables can have a deeply negative NPV.
3. Mixing nominal and real inconsistently. A nominal discount rate (one that includes inflation, like a bank quote) must discount nominal cash flows that grow with inflation; a real rate must discount constant-dollar flows. Discounting real cash flows at a nominal rate double-charges for inflation and understates value. Pick one convention for both sides and keep it.
4. Forgetting working capital and terminal value. Growth projects soak up cash in receivables and inventory before they generate it — that upfront working-capital investment is a time-zero outflow, partly recovered at the end. And for long-lived assets, the explicit forecast horizon is not the end of value: a terminal value capturing post-horizon cash flows (or salvage value) often dominates the total. Omitting both flatters short-payback projects and punishes durable ones.
5. Trusting a single-point forecast. An NPV built on one revenue guess is a guess wearing a formula's clothes. At minimum, re-run the pessimistic and optimistic cases alongside the base case. Better, flex the two or three inputs the answer is most sensitive to — usually price, volume, and the discount rate — until you find the breakeven value of each. A project that needs everything to go right is a different proposition from one that survives its downside case, even when their base-case NPVs match.
Tracking the Investment in Your Books
The analysis ends when you decide; the bookkeeping begins when you spend. Recording the outlay as an asset rather than an expense is what lets your future books reconcile with the forecast — the van sits on the balance sheet and depreciates over its useful life while the catering revenue lands in income each month. In Beancount the purchase looks like this:
2026-11-01 * "Buy delivery van for catering route" #capex-van
Assets:Van:Delivery 30,000.00 USD
Assets:Checking -30,000.00 USDPair it with a monthly depreciation entry and a dedicated revenue account for the catering route, and at year-end you can compare actual route cash against the $8,000 you projected — closing the loop between the spreadsheet that justified the van and the ledger that records what it really earned. For background on structuring asset and depreciation accounts, see the Beancount documentation.
Your Next Step
Open a spreadsheet, lay out one real decision you are weighing — equipment, a hire, a marketing commitment — as a column of dated cash flows, and discount it at your best estimate of your required return plus two points in each direction. If the NPV stays positive across all three rates, you have a project worth doing. If it flips, you have learned the most valuable thing an NPV can teach: exactly how much margin your optimism is standing on.
