Victim or Messiah — Terminal values and NPV, and what you can do about it

Kodak. Booksellers. Alarm clocks. The landline. All well-documented victims of the first wave of digital disruption.

Fast forward a few years. The fourth industrial revolution is upon us. Unicorns are no longer mythical animals (although their valuations sometimes appear to be!). Business models that were cornerstones of our economy over the past century — the taxi industry (vs. Uber), the TV industry (vs. OTT services like Netflix), the hotel industry even (vs. Airbnb) — are already suddenly starting to look questionable within a very short period of time.

This unfortunately is before disruption driven by machine learning and artificial intelligence really hits us in force, all enabled by Moore’s Law on processing power and deep learning algorithms deployed at scale. Most proponents of machine learning are guiding towards a period of increase in the cognitive and responsive capabilities of machines, at a gradient that is unprecedented. (If you haven't seen this watch Jeremy Howard's TED talk — specific reference to pace of change towards the end.)

A number of businesses stand to have the shape of their markets radically altered in this new world within a relatively short period of time. To name a few (the list is too long):

1. Car manufacturers are suddenly pivoting and re-labelling themselves as 'mobility providers' as deep learning enables driverless transportation (see Chris Urmson's TED talk to see an application of Jeremy Howard's car classification at work in Google's driverless cars)
2. Medical diagnostics and drug development businesses as deep learning enables new diagnostic tools, drug developments and treatment paradigms (IBMs big focus on Watson is in the healthcare space)
3. Insurers, as the fundamental nature of insured risk and the quantum of claims evolves

And this is just the big picture. At the product and project level within companies, this trend is being rapidly replicated as well. The explosion of data and analytical techniques is enabling disruptive product design, customer segmentation and targeting, all at an incredible pace. Data, tools and techniques that were cutting edge less than six months ago rapidly lose their distinctiveness. As an example, just when we thought mobile apps were the future, we’re now being led to believe that chat bots and Virtual Reality will take over.

In this world, how should companies think about evaluating projects and products? How should equity investors think about valuing companies? What is an appropriate leverage level that debt providers should sign up to? Historically, this function has largely been performed with the venerable cash flow based tool most favoured for valuation by MBA professors — the classic net present value (NPV) model. However any developer of an NPV model will know that the bulk of value (particularly in startup J-curve ventures) often relies on the ‘terminal value,’ which in most cases is a perpetuity after 5 years of explicit cash flows.

While terminal values have never been perfect, one could still take some comfort that changes to cash flows if any would be slow and value could still be realised from investments while business models evolved. However, if the pace of disruption has now so fundamentally changed, what confidence can today’s investors and companies place on this perpetuity, if any? Is money still being put into projects under old paradigms where investors and companies expect that products will have eternal or at least longer lives?

NPV as the Victim:                                                                   

Does the NPV tool rapidly become outmoded as it ceases to have relevance? Assuming that investors are irrational (and arguably some are) or have deep pockets/conviction (e.g., Google with its driverless cars) and look for other ways to justify investments — what replaces it then — how accurate really are the alternatives:

• Earnings multiples for established companies – still anchored to some extent of NPVs

• Venture valuation models — eyeballs, clicks, user numbers, etc. — can these really apply to big companies and projects in any reasonable fashion

NPV as the Messiah (to incumbents):

(As a preface – my personal view is that this scenario is unlikely). Assuming investors are rational, at some point, investment could dry up. If investors have no confidence in the outlook for certain sectors and for the longevity of their investments, then they could cease writing big checks. This is good news for incumbents, as large-scale disruption ends up not being funded, and venture funding only follows incremental improvements. The open question remains as to how deep does investment have to be in order to fund innovation and disruption in this phase of the industrial revolution — according to some, not very.

Clearly these are some important open questions that hopefully is the subject of academic research. However, what is clear is that the NPV in tomorrow’s world will cease to be as accurate a predictor of value as it has been over the last century. The public equity markets are also likely also have a higher level of volatility because of uncertainties in valuations.

What investors must do: 

There are however a few things that investors can and must do:

• Think through investments very carefully, conducting ‘disruption due diligence’ to identify potential weaknesses of business models, but also the opportunity to be disruptors given the capabilities of the target
• Focus more on engendering a balance within management teams and their collective ability not only to manage BAU but also adapt and evolve with the times
• Encourage, incentivise and organise around disruptive thinking and strategic agility within investments — in the words of Uber’s Travis Kalanick when launching UberX — “I don’t care about the brand. If we don’t cannibalise ourselves, someone else will cannibalise us.”