The same project, three metrics and three different conclusions

Simple payback divides the investment by the annual savings and returns a number of years. It is useful as a quick filter, and it has three omissions that matter.

It does not consider the time value of money. A peso saved in year six is not worth the same as one saved in year one. For a company with a significant cost of capital, that difference is large.

It looks at nothing after the payback point. Two projects with the same four-year payback, one with a service life of eight years and the other of twenty, look identical on that metric and are not remotely the same project.

It assumes savings are constant. They are not: the tariff moves, equipment degrades, production changes.

Take a project with investment I, annual savings A, a service life of N years, and a discount rate r that reflects your cost of capital.

Simple payback = I ÷ A. Direct. It answers "in how many years do I recover my investment?" while ignoring r and N.

Discounted payback. It brings each year's savings to present value by dividing it by (1+r) raised to the year, and accumulates until I is covered. It always comes out longer than the simple one, and the gap between the two tells you how much your own cost of capital is costing you.

Net present value. It adds up all the discounted savings over the N years and subtracts I. If it is positive, the project creates value above its cost of capital. It is the only one of the three that takes the asset's entire life into account.

And the internal rate of return is the r that makes the NPV equal to zero: the project's profitability expressed as a rate, which your committee can compare directly against other uses of capital.

It is not in the formula. It is in the annual savings that goes into the formula.

If those savings were calculated by multiplying the estimated production by your average cost per kilowatt-hour, they are inflated, because your average cost includes demand charges the project probably does not touch. The correct comparison is against the cost of the energy the system actually replaces, in the hours when it replaces it.

And if the savings were estimated without measuring, no formula can fix it. The three metrics above are arithmetic on a number; if that number is an assumption, all three return assumptions with decimals.

Tariff. What happens to the NPV if the electricity tariff rises less than assumed, or falls. A project that only works with the tariff rising is a bet, not an investment.

Production or performance. What happens if the system yields ten or fifteen percent less than projected. If the NPV turns negative within that margin, the project has no slack.

Input, where applicable. In cogeneration, the price of gas. In storage, the time-of-use differential. If the case lives on a differential and that differential narrows, the case closes with it.

A project that withstands all three with a positive NPV is a project. One that only works in the base scenario is a presentation.

The assumptions separated from the result. Estimated production, tariff used, discount rate, service life and annual degradation, each on its own line and not inside a final number.

The savings calculation, with the hours in which savings occur. If it is missing, the rest of the proposal is decoration.

And the adverse scenario already run. Do not ask for it as a favor: it is part of a serious proposal, and whoever leaves it out probably already knows what happens when it is run.

Simple payback, discounted payback, net present value and internal rate of return are standard investment-appraisal methods. This article cites no market figures because it needs none: everything above is arithmetic you can reproduce with your own numbers.

The adjusted-baseline and measurement-and-verification criteria that back the savings come from the IPMVP protocol, which is what we use in the diagnostic.