NPV and IRR: how to tell if something pays for itself

What are NPV and IRR?

They are two ways of asking one question: is this worth what it costs? Both apply whenever you pay something now and get money back over the following years: an upgrade that cuts a bill, a course that raises your pay, a pricier car that lasts longer, a tool for a side business.

• Net present value (NPV) is what everything you get back is worth today, minus what you pay today. A dollar arriving in five years is worth less than a dollar now, because today’s dollar could be earning something in the meantime. NPV shrinks each future dollar to its value today before adding them up. Positive NPV: it pays. Negative: your money would do better elsewhere.
• Internal rate of return (IRR) is the yearly rate the decision earns you. Put another way, it is the rate at which NPV is exactly zero: the break-even rate. If the IRR beats what your money could earn instead, the decision pays.

Both rest on the time value of money. If shrinking future dollars to today’s value is new to you, the present value guide covers it from the start; this page picks up where it leaves off.

Is it worth it? NPV and IRR calculator

Type what it costs today and what it brings back each year after, separated by commas. The calculator shows the NPV at your rate, the IRR, the simple payback time, and each year shrunk to today’s value. The four buttons load made-up examples to play with; the numbers in them are illustrations, not real prices, savings or raises.

The “same flows at other rates” row is worth a look every time. If the answer flips from yes to no between 5% and 10%, the decision depends heavily on the rate you pick, which is the next section.

What discount rate should you use for a personal decision?

Use the rate the money would earn if you did not spend it, at a similar level of risk. That rate is called the discount rate, and it is the most important number you choose. For a person it is usually one of these:

• Money that would sit in savings: what your savings account actually pays.
• Money that would be invested for the long run: a cautious planning figure such as 7% a year before inflation. That is not a promise; the S&P 500 compounded about 10.0% a year from 1928 to 2025, but individual years swing widely and the future can be worse.
• Money that would pay off debt: the debt’s interest rate. Paying off a 22% card is a sure 22% return, so an upgrade has to beat 22% to be worth more than the payoff.

Here are the four example decisions from the calculator, every number invented, at four different rates:

The same upgrade is a good deal for someone whose cash sits in savings and a poor one for someone carrying a 22% card balance. None of the four beats paying off that card. And the cheapest wins never need this math at all: cancelling a subscription you don’t use costs nothing up front, so there is nothing to earn back. The leak check looks for those first.

How do you calculate NPV?

Divide each year’s amount by (1 + r) raised to the number of years away, add the results, and subtract what you pay today.

How do you calculate IRR?

IRR is the rate that makes NPV exactly zero. There is no formula you can solve directly once there are more than a couple of years, so you find it by trying rates until NPV crosses zero. Calculators and spreadsheets do the trying for you.

IRR vs NPV: which should you trust?

Trust NPV for the decision, and use IRR as a feel for the margin. They agree on a simple yes-or-no question (IRR above your rate means NPV above zero), but IRR can mislead in three ways.

1. IRR ignores size

A percentage says nothing about how many dollars you end up with. Say you have $8,000 and two choices (numbers invented):

A has the shinier rate, but it only puts $1,000 to work; the other $7,000 just earns your 7%, which adds nothing beyond it. B leaves you $380 better off in today’s money. When options are either-or, pick the higher NPV.

2. IRR quietly assumes you can keep earning the IRR

Option A’s 28.6% describes $1,000 growing to $3,524 in five years. That only shows up in your account if each $400 you get back is also put to work at 28.6%. Put the $400s in something earning 7% and they grow to $2,300.30 by year 5, which is $1,000 growing at 18.1% a year. That more honest figure is the modified internal rate of return (MIRR).

For the course example, the IRR is 19.7% but the MIRR at 7% is 14.6%. Both say yes; the MIRR is the more realistic size of the win.

3. IRR can have two answers, or none

That happens when money flows out, then in, then out again. It gets its own section next.

When IRR breaks: cash flows that change sign twice

If the flows switch from negative to positive and back to negative, there can be two break-even rates, or none at all. A cost at the end is the usual cause: a cleanup, a repair, a tax bill, a deposit you must return.

The calculator above, like a spreadsheet’s IRR, reports a single break-even rate, so for flows like these, trust the NPV row and the “other rates” table instead of the IRR. Most everyday decisions (pay once, get money back) have one sign change and one IRR, and this trap does not apply.

Does it pay for itself? Payback period vs NPV

The payback period is how long until the money back adds up to the cost. It is easy to work out, and it answers “how soon do I get my money back?”, not “is it worth it?” It misses two things: the time value of money, and everything after the payback date.

A middle step is the discounted payback period: the same count, using each year shrunk to today’s value. At 7%, Y pays back in about 6.4 years instead of 5, and X never does. It still ignores what happens after, so use it as a risk check (how long before I’m safe if things go wrong?) next to NPV, not in place of it.

How do you calculate NPV and IRR in Excel or Google Sheets?

Both have the same five functions. Put the cost as a negative number and the money back as positive numbers. Here the course example is in B1:B6 (−4000 in B1, then 1000 to 1800), the stall in C1:C3, and the tool with real dates in D1:E5.

XNPV and XIRR: when money moves on real dates

NPV and IRR assume one amount per evenly spaced period. XNPV and XIRR take a date for each amount and count time in days (on a 365-day year), with the first date as today. Say the tool costs $2,500 on 15 January 2027 and brings in $600 on 30 June 2027, $700 on 15 December 2027, $900 on 1 August 2028 and $700 on 1 March 2029 (invented). XNPV at 7% is $156.67 and XIRR is 12.2%. Treat the same four amounts as if they came a full year apart and you get an NPV of −$59.15 and an IRR of 6.0%: the money really came back sooner, and that is worth something.

XIRR is also how you measure what your own investment account has earned with deposits on irregular dates. That use has its own guide.

What is a good IRR?

A good IRR is one comfortably above your discount rate. There is no single number: 9% is good if your money would otherwise sit at 4%, and poor if you are paying 22% on a card.

Two cautions make “comfortably” matter:

• The money back is a guess. A raise you hope for, a saving an installer promised, a resale value years away. The less certain those are, the bigger the cushion you want. A 7.4% IRR against a 7% rate, like the upgrade example, leaves almost no room for the saving to come in a bit low.
• Compare like with like. A long-run stock planning figure of 7% comes with years that lose money. A bill saving you are fairly sure of is closer to a safe return, so measuring it against 7% can make it look worse than it is. The rate should match the risk.

Common mistakes

• Adding up the savings without discounting. $450 a year for 12 years is $5,400 on paper but only about $3,574 today at 7%.
• Putting today’s cost inside the spreadsheet NPV range. Add it outside the function.
• Choosing between options by IRR. The higher rate is not always the bigger win. Compare NPVs.
• Counting too many years. Only count savings for as long as the thing will really last, and include repairs and replacement costs as negative years.
• Using pre-tax numbers. A raise is worth what reaches your account. Rebates and credits go in the year you receive them.
• Forgetting the free option. Paying down high-interest debt, or simply not spending, is often the decision every other option has to beat.