NPV and IRR — Two Questions Every CFO Must Be Able to Answer

The Economist Who Survived a Death Sentence and Invented the Future

For three years he fought — not against equations, but against his own lungs. He became a vegetarian, wrote a bestseller on health, lived to eighty, and in the meantime described a principle that sounds simple yet changes everything.

His promoters were the theoretical physicist Willard Gibbs and the sociologist William Graham Sumner — a combination that explains much. Fisher thought like a scientist and wrote like a human being. Econlib describes him as one of the clearest economic writers of all time — someone who had the intellect to use mathematics throughout his theories and the good sense to introduce it only after he had explained the principles in words.

In 1907 Fisher published “The Rate of Interest.” In this book he recorded a proposition that sounds simple yet changes everything: the true value of any asset is the present value of the future income that asset will generate. Not the purchase price. Not the book value. Not how much someone paid for it yesterday. Value is the sum of what it will yield tomorrow — translated into today’s money.

This is the heart of NPV. And this is the question you should ask before you spend a single penny on digital transformation.

What Is NPV — Explained Without Formulas

Net Present Value is the answer to the question: will the money I invest today in this project come back to me tomorrow in a larger amount than if I had simply left it in a bank account?

It sounds trivial. It is not.

PLN 100,000 today is not the same as PLN 100,000 in three years. Over those three years the money could have been working — on a term deposit, in bonds, in another project. Inflation erodes its purchasing power. Risk grows with every passing month. NPV takes the future benefits of an investment — savings, revenues, freed-up labour hours — and translates them into today’s value, accounting for the cost of time and the cost of risk. Then it subtracts the amount you must spend now.

If the result is positive — the investment creates value. If negative — it destroys it. If zero — you break even, and your time and effort were spent for nothing.

Fisher knew that value has a temporal dimension — not just a quantitative one. A dollar today and a dollar tomorrow are two different dollars. He called this principle the theory of impatience and opportunity — impatience and opportunity. Impatience, because people prefer money now to money later. Opportunity, because money today can be invested and multiplied.

Every CFO in the world makes decisions in the shadow of this principle. Even if they have never heard of Irving Fisher.

Why ROI Is Not Enough

Most companies considering process automation calculate ROI. Return on investment. The formula is simple: (profit − cost) / cost × 100%. The result: a percentage. “ROI is 240%.” Sounds good. Tells you almost nothing.

Because ROI does not answer the question “when.” A project with ROI of 240% spread over eight years is an entirely different proposition from a project with ROI of 120% that pays back in twelve months. ROI does not discount future cash flows. It does not account for the cost of capital. It does not show which quarter the investment starts to earn its keep.

NPV does.

CFO.com puts it neatly: “We’ve heard managers say a decision was ‘strategic’ because it couldn’t be justified financially. But if it never turns out to be financial — it’s not particularly strategic.” NPV enforces honesty. It does not let you hide behind a vague “strategic value.” It forces you to translate that value into money and verify that the result is positive.

NPV of Digital Transformation — What It Looks Like in Practice

Let us take a concrete scenario. A company processes 3,000 invoices per month. The current cost of manually processing one invoice — let us say PLN 50 (a conservative estimate, covering labour time, error correction, and systems costs). Annual process cost: PLN 1,800,000.

After automation, the cost drops to PLN 12 per invoice. Annual cost: PLN 432,000. Annual saving: PLN 1,368,000.

Cost of implementing the automation: PLN 400,000 in year zero.

A simple ROI calculation says: (1,368,000 − 400,000) / 400,000 = 242%. Impressive. But incomplete.

NPV digs deeper. We assume a discount rate of 10% (the company’s cost of capital). We calculate the present value of savings from each year — year one is 1,368,000 / 1.10 = 1,243,636. Year two is 1,368,000 / 1.21 = 1,130,579. Year three — 1,027,799. Sum of present values of three years of savings: PLN 3,402,014. Minus implementation cost: PLN 400,000. NPV = PLN 3,002,014.

Three million in net present value over three years. On a single process.

Of course this is a model — reality is more complex. Savings may grow (because the company grows and invoices multiply) or decline (because some benefits materialise earlier). The discount rate depends on the industry and capital structure. But even under conservative assumptions NPV shows something ROI cannot: the scale of value translated into today’s money.

The Cost of Inaction Has Its Own NPV

There is another dimension that is rarely discussed. NPV works in both directions.

If automating the invoice process yields PLN 1,368,000 in annual savings, then every year without automation costs your company the present value of that amount. Deferring the decision by twelve months is not “preserving the status quo.” It is the active loss of PLN 1,243,636 in present value — assuming the same discount rate.

PwC’s 2025 study reports that 92% of operational leaders admit their technology investments did not fully deliver the expected results. But “did not deliver” does not mean it was not worth investing. It means they calculated incorrectly. Or did not calculate at all.

NPV does not provide guarantees. It provides discipline. It forces you to record your assumptions, translate cash flows into numbers, and price the risk. And then — to compare the result against alternatives. Because there is always an alternative: do something else with that money. Or do nothing — and bear the cost of inaction.

Two Geniuses, One Question, Two Answers

Enter Keynes. Thirty years after Fisher, in 1936, in “The General Theory of Employment, Interest and Money,” he asked the complementary question: not only how much is an investment worth, but how quickly will it pay back.

Keynes called his answer the “marginal efficiency of capital.” He defined it as that rate of discount which would make the present value of the series of annuities given by the returns expected from a capital asset during its life just equal its supply price. The economist Abba Lerner later clarified the terminology, but the essence remained: Keynes described the mechanism we now know as IRR — Internal Rate of Return.

Fisher asked: how much is this investment worth? Keynes asked: how quickly will it pay back?

This is not the same question. And understanding the difference between them changes the way you make decisions about money.

What Is IRR — Without Equations, With Meaning

NPV gives you an amount. PLN 3 million in net present value. Good information. But a CFO looking at two projects, both with NPV of three million, still does not know which to choose. Because one requires an investment of PLN 500,000 and the other — PLN 5 million. One pays back in a year, the other in seven years. NPV will not show this.

IRR will.

IRR is the rate of return at which a project’s NPV equals exactly zero. In other words: the point at which an investment stops destroying value and begins to create it. But more intuitively — the speed at which your money comes home.

If a project’s IRR is 45%, it means that every pound invested in that project works for you at 45% per year. If your cost of capital — WACC, the weighted average cost of raising money — is 10%, then the difference between 45% and 10% is pure profit above the cost of financing. The larger this gap, the safer the decision.

The rule is simple: invest when IRR exceeds the cost of capital. Do not invest when it falls below. This is the logic that corporations have applied for decades in capital budgeting — from new factories to acquisitions, from IT infrastructure to automation programmes.

The Trap Keynes Didn’t Warn About

Now I will change pace, because IRR has a dark side that most articles ignore.

Imagine two projects. Project A requires an investment of PLN 100,000 and generates an IRR of 80%. Project B requires an investment of PLN 2,000,000 and generates an IRR of 25%. Which do you choose?

Instinct says: A. 80% sounds remarkable. 25% sounds solid but not exciting.

Calculate the NPV. Project A at a discount rate of 10% generates NPV of PLN 120,000. Project B — NPV of PLN 1,800,000. Fifteen times more value. But IRR suggested A was better.

This is the scale trap. IRR does not account for the size of the investment. It does not tell you how much value a project creates — only how intensely it turns capital. A small project with an enormous IRR can be less valuable than a large project with a moderate IRR.

Alchian noted this in 1955. Hirshleifer confirmed it in 1970. Ranking projects by IRR can reverse depending on the discount rate — the value curves of two projects can cross, and the project that looked better at a low cost of capital turns out to be worse at a higher one. This is not a theoretical quirk. It is a situation that arises regularly in automation project portfolios.

There is also a second trap — the reinvestment assumption. IRR assumes that all cash flows generated by the project are reinvested at the same rate of return as the project itself. If your project has an IRR of 45%, IRR assumes that every penny it generates will immediately find another home also yielding 45%. In reality? More like 5–10%, because that is what the market pays for safe capital. NPV does not make this assumption — it discounts cash flows at the cost of capital, not at the project’s rate of return. This is why NPV is the more honest metric.

Two Tools, One Decision Dashboard

Since IRR can mislead — why calculate it?

Because a CFO needs both pieces of information. NPV answers the question “how much?” IRR answers the question “how quickly and how intensely?” Together they create a complete picture.

Imagine you are presenting an invoice process automation project to the board. You say: “The NPV of this project is PLN 3 million over three years.” The Chairman nods. Then asks: “And what is the return on capital?” You say: “IRR is 68%.” The Chairman leans forward. “And our cost of capital?” “Ten percent.” Silence. “So the project returns capital at a rate seven times higher than our cost of financing.” The decision is made in minutes, not weeks.

This is the power of the NPV + IRR pair. NPV legitimises the investment — it shows that it creates value. IRR legitimises the pace — it shows that capital works intensely. Neither tool works as well on its own.

That is why in the QA10 knowledge base both concepts appear side by side.

IRR of Automation — Why It Comes Out High and What That Means

Automation projects have a specific profile that typically generates a high IRR. Three reasons.

First — low entry cost. Implementing RPA or an integration layer in a Zero-Trust philosophy costs a fraction of what replacing an ERP system would. With a low investment denominator, even moderate savings generate a high rate of return.

Second — fast time to value. Invoice process automation starts generating savings from the first day it goes live in production. Not after two years, not after a “stabilisation phase.” From day one. That is a short payback period, which boosts IRR.

Third — repeatability of savings. Every hour saved repeats every week, every month, every year. The cash flows are stable and predictable — ideal raw material for a high IRR.

But — and here I return to the trap — a high IRR does not automatically mean the project is the most important in the portfolio. It may turn out that automating a minor process delivers an IRR of 120%, but saves PLN 50,000 per year. And reorganising the supply chain delivers an IRR of 22%, but creates PLN 4 million in value. Which comes first? NPV says: supply chain. IRR says: invoicing. The truth lies in combining both answers — and in the context that no formula alone can supply.

How QA10 Uses NPV and IRR in Conversations With the Board

In most companies the conversation about automation begins with technology. “We need RPA.” “We need a new system.” “We need AI.” The board hears the cost, sees the risk, and asks: “Can’t we wait a bit longer?”

At QA10 the conversation begins with numbers. The AiP Audit — our process analysis — does not end with a process map. It ends with a spreadsheet in which every identified process has four numbers: the current cost of the activity (from Activity Based Costing), the projected cost after automation, the NPV at the assumed discount rate, and the IRR.

These four numbers form the language in which a CFO makes decisions. Not “this process is worth automating because it is tedious.” But “this process costs PLN 1.8 million per year, automation costs PLN 400,000, NPV is PLN 3 million over three years, and IRR is 68%.”

The first version of that sentence is a request. The second is a business case.

And here the circle closes. Process mining shows what is happening. Activity Based Costing says how much it costs. NPV says how much the change is worth. IRR says how quickly it pays back. RPA does the work. Zero-Trust ensures your ERP remains untouched. TRL 9 guarantees the solution is not an experiment.

Seven concepts from our knowledge base. One coherent decision architecture.

Irving Fisher wrote his theory so clearly that economics students could read half his book in a single sitting. Our AiP Audit report is written in the same spirit. Not because we simplify — because complexity hidden behind simplicity is a sign of maturity, not naivety.