The most common way to show Return on Investment is called “Net Present Value”, or NPV. NPV became the most popular way to calculate ROI because of its simplicity in concept and execution. It’s a three-step process: forecast future revenues, calculate the investment required to generate those revenues, sum all those cash flows together discounted at the appropriate cost of capital to produce an ROI that is comparable across the firm.

For incremental innovation, this approach is sound. The managers who use it have a great deal of understanding on the product, the market, and the relative risk of a project. This is the job after all, and it is not a stretch to extend forecasts when you have dozens of years of trends and historical data from which to draw. In some cases it can be as straightforward as dragging the bottom right-hand corner of your excel spreadsheet.

But for discontinuous innovation, there are no such data. You’re starting at time zero. And while you can make reasonable estimates of how much you’re going to spend and the discount rate, the future sales are totally unknown. Facing this, PMs accustomed to the NPV approach necessarily use made-up data – preferably from a third-party source such as an expensive analyst report on future market potential. Accordingly, the revenue forecast becomes a market share percentage of that total available market when the product is a big hit.

And this is where the problems begin. You see, the revenue forecast is just as unknowable to the finance department which reviews the analysis as it is to the PM who produced it. And since the PM knows that he’s competing against other projects for the scarce resources of the firm, he has to produce a sufficiently high NPV to get above that threshold. Moreover, finance knows this as well, and so discounts the forecasts even further, which encourages the PM to bump up the projects as far as credulity will allow.

In the end, what is submitted as a discipline in diligent research and careful planning is actually an exercise in tweaking the excel spreadsheet until it generates an NPV big enough to be compelling but not so big to be farcical. And what’s most ridiculous about this scenario is that it is an open secret that each side is making it up. Still, once approved by management it becomes the official plan and this is where the problems expand.

It increases risk.

As mentioned, unlike startups in a corporate environment multiple initiatives all compete for the same set of finite resources. This means that when deciding between two projects that are otherwise equal, the proper move for finance is to select the one with the higher NPV, which effectively means “the one with the larger revenue forecasts.”

By definition this favors larger projects over smaller ones – and specifically excludes the ability to experiment, since projects focused on small-scale market testing have negative NPVs are not even considered. Consequently, this increases the budget of any potential project and the subsequent risk associated with it.

It discourages reality.

An NPV-based project is static. A great deal of effort goes into the analysis of course, but once complete and approved everything shifts to measuring forecasts v. actual. And since this occurs before the project itself is initiated, it requires companies to make the most important funding decision at the precise moment there is the least amount of information.

Moreover, any information that contradicts the original forecasts calls into question the entire financial justification of the project. This produces a clear incentive to delay, obfuscate, or outright reject reality in favor of the original plan. Moreover, when it’s time for quarterly reviews, if revenues do not match the original projections — as is almost always the case for projects dealing with uncertainty — the only solution to maintain the same NPV is to simply make larger revenue forecasts at points still further into the future. This is the dreaded “NPV Death Spiral” and begins the day that NPV is selected as the basis for funding the project.

It is bad economics

Yet perhaps worst of all is that most savvy innovators know that all of this is true, and as a consequence they don’t innovate at all. Remember, innovations don’t build themselves. At some point a PM needs to take the lead to make the idea real. But if required to show the NPV, then that PM knows he will be held to account for revenue projections that he intuitively knows are fabrications. Should those fabrications prove correct, he’ll be acknowledged at having done the job correctly. However, if those targets are missed the PM will be seen as unreliable at best. In other words, very little upside with lots of downside.

Alternatively, if the idea is good enough, the PM can leave that job and start the company on his own. He still might miss the target and fail which puts him the same position — but the upside is dramatically different. This is why it’s not at all uncommon to see an employee leave a firm in the same industry only to return by selling the result back to his original employer. The economics simply don’t favor the person taking the risk in a corporate setting.

Summary

This is the Innovator’s Dilemma, expressed in finance. It’s a reflection – or consequence – of the accounting systems we used to manage innovation. The fundamental flaw in the NPV approach is that it demands high degrees of certainty from the outset and does not adjust when conditions change. This works for incremental innovation, but utterly fails when considering breakthrough opportunities. And it’s why we need a different accounting approach like innovation options which are specifically designed to address the ambiguity inherent in discontinuous markets.