I consult, write, and speak on running better technology businesses (tech firms and IT captives) and the things that make it possible: good governance behaviors (activist investing in IT), what matters most (results, not effort), how we organize (restructure from the technologically abstract to the business concrete), how we execute and manage (replacing industrial with professional), how we plan (debunking the myth of control), and how we pay the bills (capital-intensive financing and budgeting in an agile world). I am increasingly interested in robustness over optimization.

I work for ThoughtWorks, the global leader in software delivery and consulting.

Sunday, June 30, 2019

Come On and Take a Free (Regulatory) Ride

Within 24 hours of last month's post on regulation of technology firms, new anti-trust probes into Google and Apple were reported by the Wall Street Journal. Around the same time, Renault stirred up a hornets nest of regulatory furore over its proposed (and subsequently withdrawn) merger with Fiat Chrysler. It's worth examining the prior in the context of the latter to understand why the regulatory free ride Big Tech Media firms enjoy will continue for quite some time to come.

Instead of thinking about companies as financial entities that reward investors, think of them as agencies that fulfill public policy. Let's start by looking at the automotive industry. Economies benefit from the physical mobility of their citizens: the more affordably they can get around, the more flexible they are as a labor force, the more likely they are to get out and spend, etc. To be competitive in the transportation business, a company needs scale; that means transportation companies are large direct employers. The complexity and sophistication of modern cars means that automakers have expansive supply chains, making them large indirect employers as well. Transportation is an industry of great national importance: jobs create middle-class households that spend money, pay taxes, and vote.

But transportation creates a lot of other problems such as energy consumption and pollution, so cars are regulated for their safety equipment, the types and quantities of energy they use, and the amount of pollution they generate. In this way, government can get its citizenry mobile with fewer negative consequences. It isn't a stretch to say that automakers are executors of public policy on the combination of transportation (mobility is good), labor (jobs are good), energy (fuel efficiency is good), and environment (clean air is good).

It takes quite a while for a society to figure out the consequences of its innovations and the price it is paying directly or indirectly for the benefits it enjoys, hence regulation lags commercial innovation for a very long time. For example, corporate financial disclosures were not regulated with the first stock sale or even the first equity market crash: the Philadelphia Stock Exchange was established in 1790, but the Securities and Exchange Commission didn't come into being until the 1930s. Similarly, cars were not subject to emissions regulations with the sale of the first car; delays in emissions regulation stemmed from the fact that the good of affordable mobility outweighed the bad of pollution in the minds (and perhaps the wallets) of policy makers for a very long time.

Let's look at the aborted Renault-FCA merger in this light. Aside from the tantalizing potential for the best buffet cuisine ever in the history of annual general meetings of shareholders, in financial terms this appeared to be a rational merger. A combined Fiat Chrysler Renault with a side order of Nissan and Mitsubishi would have girth in all of the numbers automakers have been traditionally measured on: production volume (third largest, globally); market access (full-scale distribution networks in every major car-buying market including the US, Europe and China); and margin (Dodge and Jeep light trucks.) In short, the proposed merger would have yielded everything except for the specific electronic vehicle technology that governments are adamant that auto manufacturers must deliver. Even there, the combined R&D spend would effectively double, reducing wasteful redundancies: US$8b goes a lot further than US$4b twice over. Given the fact that for the time being EVs are doing more damage to automaker P&Ls than being a source of industry disruption, that combined R&D spend would go a long way to charting a course for a successful future of the two companies.

Unfortunately, the tie-up of the two firms quickly ran afowl of labor policy. Part of the sales pitch of the merger was that Renault would gain access to the US market through Fiat Chrysler dealerships. But it doesn't make sense for a dealer to be selling a Renault Clio side-by-side a Fiat 500, with a Mits Mirage in the showroom just through that door over there and the Nissan Versa in a showroom across the parking lot. It makes even less sense given both dealer and manufacturer are losing money on the wholegoods sale of the small car. There is a stronger argument to be made that these products on nearby showroom floors will simply cannibalize sales from one another, much as the Jaguar F-PACE did to the Land Rover product lineup. If that proves to be the case, than the combined companies will stage their own private fight clubs of their products. Sooner or later, the rationalized product portfolios will result in rationalized production capacity. By way of example, look at the lineup of Fiat Chrysler products: the Chrysler and Dodge brands have no small cars; FCA meets US CAFE requirements with the small Fiats. And, of course, Kraft made similar assurances about maintaining production facilities before acquiring Cadbury only to abandoned them one week after the completion of the merger. The leaders of Renault and FCA said they anticipated no plant closures, but it was difficult to take those assurances very seriously.

In financial terms, consolidation made sense because a consolidated manufacturer can produce just enough small cars (which are by and large unprofitable) to offset their production of light trucks (Chrysler, Nissan pick-up trucks), designer vehicles (Alfa Romeo), performance vehicles (Nissan GT-R, Alpine A110) and exotic vehicles (Ferrari), all categories where automakers make money. But in political terms the merger did not make sense, because it is easier to ramp up production of the facility producing (currently) money-losing small cars to offset any gain in money-making larger cars than it is to perpetuate a lot of factories producing small cars. Gains for a factory producing small Fiats in Italy really would translate into reductions at a factory producing small Renaults in France if the consolidated automotive fight club rewards the Fiat model over the Renault model. While the combination would make financial sense and make it more likely that the two companies would survive the transition to electric cars, the loss of jobs was too great a risk for either the Italian or French governments to bear.

What does this tell us of the ramp-up of anti-trust probes into Apple and Google?

Google is a big media firm which it monetizes through advertising. Google did not invent advertising, mass advertising, or targeted advertising; it found ways to provide a scale of advertising at a level of reach greater than any outlet had ever created before, which it can sell to advertisers at a price lower than what traditional media could offer, with more accurate targeting to boot. Google provides this ad space in media outlets that did not exist previously, ranging from browser-based search, to geographic search (maps), to in-home search (Google Home), to in-car (Android Auto), to many other media outlets. Google's imagination for the possibility of monetizable moments is many decades in front of that of any government.

In the process of becoming Big Media, companies like Google have become Big Surveillance, both directly (the data they capture and analyze) and indirectly (the multitude of 3rd party firms in the analytics chain). Cross-analyzing data creates infinitely targeted opportunities to pitch a consumer transaction: by capturing payments (Google Pay), Google knows you have an affinity for Mediterranean cuisine; by capturing your location, Google knows you travel to San Antonio and stay in the same hotel every week; by capturing your boarding pass info, Google knows your flight is running a bit late. It might be a great convenience - in fact, it might even delight you - if Google Express not only suggested a Mediterranean restaurant, but built an order based on your previous menu choices and scheduled the order to be delivered to your hotel in time for your late arrival.

Consumers are not the only ones who stand to benefit from such analysis. The accelerometer in the phone and the spacial-temporal data captured by maps mean that Google know whether you drive the rental car you get in San Antonio every week in an aggressive manner. The rental car company and your auto insurer would like to know that, as they could price in the risk of accident based on characteristics of your actual driving patterns as well as the responses by the drivers around you to how you drive. (Yes, Progressive Insurance has offered something like this for years, but with Progressive the insured is making an explicit decision to expose driving data; Google could choose to collect that data and provide it to insurers and rental car companies and law enforcement agencies without consumer consent). The contract between insured and insurer has traditionally been based on reported events, specifically moving violations and accidents. The insured had a great deal of leeway in deciding how to behave as long as the what happened showed no indication of trouble to the insurer.

Western societies. and the United States in particular, are based on the rights of the individual, the wellspring of personal freedom. Individual anonymity is a contributing factor to protecting those rights, as the individual is free to make decisions - what to do, how to drive, what to eat, what to invest in, what to read, how to be entertained - without fear of repercussion for the choices they make. Continuous surveillance all but eliminates the anonymity of the individual, exposing not only what people do but providing potential clues as to why they do them because a person's decisions and actions expose their own preferences, beliefs and values. Plus, any one person's decisions erode the anonymity of the people around them: the data collected on one person may indicate an ailing parent or struggling child. Continuous surveillance raises some thorny questions for western societies. What are the consequences when the individual effectively loses decision rights over how they do things to the companies they do business with? What is to stop a company from predatorily monetizing a condition that a person wishes to keep discreet? What prevents an individual from being targeted for the values they are perceived to hold through the decisions they make? What if the algorithms conclude the wrong things about an individual based on the data that it has?

Western governments have historically intervened (ostensibly) on behalf of the individual to level the playing field with large companies. The rise of the American middle class in the first three decades of the 20th century created a large new category of individual investor, enabling legal but ethically questionable (that is, exploitative) financial reporting practices by corporations; with middle-class investors wiped out by the market crash and confidence in public markets all but non-existent, the US government created the SEC. The shift of the American economy from agrarian to industrial in the 19th century allowed for legal but ethically questionable labor practices by industrial firms; as labor unions arose, government eventually supported (and in many cases, enabled) worker organization. The plot is no different this time around, with Big Media engaging in legal but ethically questionable consumer practices. The scale is much different, though: 20th century Big Industry was large in economic might but still human in scale; Big Surveillance is large in economic might and machine in scale.

The anti-trust probes revealed earlier this month target the commercial activities of Google and Apple. These are regulatory probes within the constructs of established policy stemming from the industrial era. The question any probe has yet to ask is one of public policy: if and how to protect individual freedom against constant surveillance and activity derived from that surveillance at machine scale. Before it can answer that question, government must be able to frame the problem in that light, define what rights individuals really have, can conscionably waive and under what circumstances they can be waived, and what rights the individual has no right whatsoever to expect.

Labor policy is well established, so it wasn't difficult for either the French or Italian governments to conclude the risk of combining state champion companies was too great. Data and information policy is not well established and it will take time for the US government to define it in such a way that companies that rely on Big Surveillance do so in a manner that advances desirable, multi-faceted public policy, and not simply exploit its absence. The scale and complexity of forming that policy, in what is new regulatory territory for every government, means that the regulatory free ride will continue for quite a long time.