Industrial Entropy: Technology, Capitalism and the Rise of the Agile Firm

Note: This post is the text of a talk I gave today at Defrag 2013. Thanks again to Eric Norlin for the opportunity to speak and hope this fulltext fills in the gaps left by my onstage free association.


The technology industry has changed sides. After decades as an ally of the multinational corporation — enabling scale and conferring competitive advantage — the accelerating pace of technology change has reached a point where most large organizations can no longer keep up.

As the traditional enterprise comes under threat a new form of tech-enabled multinational is emerging to replace it. Unlike the firms that came before them, these new global leaders have harnessed the accelerating pace of technological change as their most potent strategic weapon.

Coase and “The Nature of the Firm”

In 1937 Ronald Coase — who died this year at the age of 102 — offered economists the first theoretically satisfying explanation for why companies exist. Up to that point, efficient market theory suggested that the ideal market environment would be a flat landscape of specialized professionals trading amongst each other for needed goods and services.

By enumerating all the “hidden costs” associated with free-market exchange — things like search and information costs, bargaining costs, keeping trade secrets, and policing and enforcement costs — Coase helped explain why firms arise, bringing in-house those functions that carry too heavy a burden of hidden costs when purchased on the open market.

WWII and the rise of the multinational firm

Coase advanced this theory just before the Second World War kicked off an intense 30-year cycle of industrialization and globalization, in which massive, vertically integrated firms like General Motors, General Electric, IBM, and AT&T came to dominate huge segments of the global economy.

Technology as an enabler of industrial scale + dominance

Technology was a key enabler of this concentration of power among a small number of firms. Computers and the people who ran them were scarce and hideously expensive — only the largest and most profitable firms could afford to invest in automation and computer-intensive optimization, and the advantages they reaped through their investments in technology only increased their dominance.

The microprocessor era shifts the balance of power

But the peak years of the multinational conglomerate — the Nifty Fifty era of the 1960’s — also witnessed the birth of the microprocessor. Intel founder Gordon Moore first posited what is now known as Moore’s Law in a paper published in 1965, and the geometric logic of that phenomenon has been steadily chipping away at the foundation of the enterprise ever since.

First software deconstructed the industrial value chain…

In the earliest days of computing, large firms used IT to disassemble major segments of their supply chain.  Enterprise Resource Planning (or ERP) systems allowed corporations for the first time to create reliable digital control systems for inventory management and the factory floor.

Extending these systems to major trading partners reduced the Coasean “hidden costs” to the point that critical components and subsystems could now be handed off to other, more efficient firms, lowering the overall cost of production.

At the same time, the rise of newly-industrializing economies in Asia and Latin America — aided by IT-enabled “intermodal” or container shipping — supported the development of super-efficient cross-border supply chains that allowed sourcing and pre-assembly to be shifted to vendors overseas.

Then software deconstructed knowledge work…

The PC era brought the cost of computing down to the point that most high-value knowledge workers inside the firm could be provided with digital tools. And the same process of disaggregation began to appear in the management layers of the enterprise.

WordPerfect and VisiCalc transformed highly proprietary, document-based knowledge work into standardized digital files that could be copied, shared and modified by workers inside or outside the company.  This reduced friction in the transmission and reuse of information and further reduced the “hidden costs” of distributing knowledge work among trading partners. As a result, more and more previously “core” departmental functions — finance, accounting, marketing, sales — could now be farmed out to outside partners without a loss of fidelity.

The tail is now wagging the dog

You wouldn’t be in this room if you didn’t already know where this is headed. I’m guessing that everyone here knows what BYOA, BYOD and SaaS mean; and that I’m not talking about the weather when I say Cloud, or Virtualization, or Software-Defined Networking.

IT began as an enabler of the disaggregation of enterprise work, but has become a mirror of that work: we have passed from an era of personal, individual and portable tech to one of personal, individual + portable work.

When every knowledge worker has the power to choose his or her own technology tools…

…and through those tools can collaborate — through standardized networks and APIs — with both humans and machines based anywhere on the planet…

…with almost any kind of supporting data and information instantly searchable and freely available…

…many of the ”hidden costs” that Coase identified as the reason for firms to exist are no longer present.

Some of these traditional hidden costs have now been replaced by new bottlenecks — the acute scarcity of modern software engineering talent and skilled data scientists are just two current examples — but the speed with which new bottlenecks can now be addressed by adaptive technology makes controlling for these kinds of hidden costs a weak foundation on which to build an enterprise.

What does Industrial Entropy look like?

The economic and social implications of this disaggregation are profound. As just one example, consider the historic disconnect between corporate profits and employment in the U.S. economy. Private employment began to diverge from GDP and labor productivity about 15 years ago, causing employment (and household income) to effectively flatline even as GDP and productivity continued to grow. When work is disaggregated from the firm, firms are still able to grow revenue and profits, but employment plays a smaller and smaller role in that growth.

Reductio ad absurdum?

It’s tempting to just project these lines forward, and those who do so quickly arrive at a modernized version of the pre-industrial craft economy, with a global network of happy sole proprietors freely maximizing their comparative advantage with the help of frictionless digital marketplaces.

The only problem with this bucolic vision is that is doesn’t account for the recent, rapid rise of a new kind of multinational corporation – the kind that quietly enables all this frictionless sharing and trading. The companies that make headlines in our industry are the ones that are in contention to run this frictionless economy, and – marketing spin aside – there’s nothing pastoral about them.

Facebook, Apple, Google, Amazon, eBay – and more recently, Uber, Airbnb, Dropbox, GitHub and oDesk – are all on a mission to break down traditional sources of friction in the economy, with the simple aim of moving those freedom-enabling transactions onto their platforms.

Revisiting the Nature of the Firm

So what we find ourselves with is not a linear progression to a craft economy, but a new and even more radical polarization of economic actors: a massive global pool of contingent labor on one end, and on the other, a tight concentration of massive-scale — but relatively small headcount — global trading platforms for labor, goods, information and entertainment.

Does that mean it’s time to reconsider Coase + redefine the nature of the firm? Or does this polarization just help us see the Coasean firm more clearly for what it is?

Venture Capital and the Firm

At this point it’s useful to take a brief side trip to the venture capital business.

About five years ago I switched sides, from a series of founder/operator roles in software startups to become a “professional” investor in tech companies. It’s been a fascinating journey and I still feel like I’m just getting started, but a few lessons have been driven home very forcefully by my experience.

Rule #1: Capital seeks dominance

The first and most germane to today’s discussion is that the capital markets are dominance-seeking; capital wants to create + capture surplus; monopoly rents are best if you can get them (and keep them), but market leadership to the point of dominance is a nice second-best. Mike Maples calls these kinds of companies “thunder lizards”, and they’re the ones that drive the majority of venture returns.

Rule #2: Achieving dominance is slow + “planful”

The next big lesson, which is really a corollary to the first, is that venture is a slow business, not because VC’s all take August off to sit on their yachts, but because prosecuting dominance is not a short-duration activity.

True dominance in today’s global, tech-enabled economy is achieved only through incredible levels of sustained excellence, and requires successful firms and their leadership teams to pass through a series very difficult organizational and capital markets state-changes. Starting a software company is checkers; building one to dominance is three-dimensional chess, with a good measure of luck and timing thrown in.

Rule #3: Dominance has become fragile

The third lesson, which may seem like a contradiction of the previous one but really confirms the degree of difficulty associated with achieving dominance, is that Moore’s Law has made dominance more fleeting than ever. The pace of technology innovation – of the kind that has the power to threaten the existence of entire industries – has accelerated to the point that most organizations operating at any meaningful scale simply can’t change fast enough to keep up.

Conclusion: Dominance has become a capability, not a state of being

The implication of this change is that the nature of competitive advantage – which for decades has been built around relatively stable things like brand, capital strength and established distribution networks – has begun an accelerating shift toward a much more fragile and hard-to-scale capacity with a one-word description: “agility”.

Agility isn’t a thing, it’s just a label for a tightly interwoven set of human capital, business process and leadership assets that, in combination, allow the best large organizations to use the pace of technological and cultural change as a weapon against their competitors.

So, why does the firm exist today?

Seen through this lens, it’s not crazy to suggest that “winning” modern firms exist not to avoid hidden frictional costs that sit outside the organization, but rather because their founders seek advantage – a chance at monopoly rents – by enabling and accelerating the disaggregation of a chunk of the economy to the point that no other economic actor has the power to challenge their dominance in that domain.

Building the operating systems for a “frictionless” economy, while keeping pace with accelerating technology change, is a game that only the best-led and best-managed firms can play. Sustained agility at scale – the kind that Jeff Bezos has prosecuted for over a decade at Amazon – is out of reach for most of the large global firms that exist today.

In effect, the technology industry has changed sides, from being the handmaiden of concentrated industry to becoming its nemesis. Collectively, we are enabling the rise of a new kind of global firm whose primary aim is to drive Ronad Coase’s “hidden costs” of trading as close to zero as possible, while reserving that last tiny fraction for itself alone.

Postscript: Technology, Capitalism + Human Beings

My topic today was the changing nature of the firm, but it’s hard to talk about the increasing polarization of capital and labor without at least acknowledging some of the more troubling human and societal implications of this shift.

The best feature of the vertically integrated multinational corporation was the fact that it provided good jobs to people at almost every point on the employment scale — from unskilled laborers to PhD scientists — with a meaningful chance at progression along the curve.

The disaggregation of the firm – and the relentless march of technology and automation – has radically shrunk the number of good full-time jobs available at every point in the labor curve except the very top. The near-term implication of this change is the concentration of wealth at the top of society. The longer-term scenarios include things we haven’t seen in the first world since early in the last century: widespread poverty, social unrest and despotism.

Our collective responsibility – even as we lay the rails of the frictionless economy – is to make sure we’re creating productive, meaningful roles in society for the largest possible number of the world’s people. We’ll all be better off if we do so, and we really won’t like the results if we don’t.