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, January 31, 2016

Are Microservices to Ecosystems as Core Competencies were to Conglomerates?

As far back as the 19th century, industrial firms pursued vertical integration strategies. The thinking was that by owning the supply chain from raw materials to retail outlets, a firm had direct control over its entire cost structure, making it better able to squeeze efficiencies out of it and being less susceptible to supply shocks. This was important because, for large industrial firms, competing on price was the primary strategy for winning market share.

During the 1950's and 60's, companies also pursued conglomerate strategies: bringing seemingly unrelated businesses under one roof, sometimes seeking synergies (as Sears did owning a retail merchandiser and retail brokerage - "buy your stocks where you buy your socks"), and sometimes not (as LTV did owning a steel company and an airline). The rationale for the conglomerate was entirely financial: cheap (mostly debt) capital allowed large companies to grow through acquisition, and regulators were less likely to block acquisitions of unrelated firms on monopolistic grounds.

By the 1980s, both strategies had begun to lose favor. The financial benefit had evaporated: high interest rates clobbered the profits of debt-fueled acquisitions and forced divestiture. But the operating benefits weren't there, either. Different types of businesses (manufacturing, distribution, retail) require different types of leadership and have very different cultures. And, within each of those businesses, some functions are differentiating (such as fleet optimization for a logistics company) while some functions are not (nobody beats their competitors by having a superior accounting back office). Management thinking embraced "core competencies": own and hone the things that differentiate, rent and commoditize the things that do not. This also allowed for better matching of capital with company: the risks and returns from a company that owns a fleet of railcars is easier to assess than the risks and returns from a company that owns ore mines, railcars, and foundries. By breaking them up, the individual investor can choose what types of businesses to expose their capital to (a raw materials company, or an asset company, or a refining company), and the pricing of that capital more accurately reflects the risks.

Tech firms today are embracing the "vertical integration" and "conglomerate" strategies of yore by positioning themselves as "platform" and "ecosystem" companies. The thinking is that by combining multiple and diverse capabilities into a single offering, a company creates both efficiencies and synergistic value for counterparties in some activity, such as crowdsource funding or payments. The ecosystem strategy takes this even further, combining unrelated capabilities under one roof (eBay buying Skype in 2005, SAP developing HANA in 2010, Facebook buying Oculus in 2014), often justifiable if only because digital commerce is still in its infancy and nobody is really sure what's going to work and what's not.

But what if you could extract utility-like functionality from within an ecosystem into an independent company? Take payroll as an example: rather than have every Human Resources platform company need its own team of people to write and maintain state and federal witholding rules, hive those off into an independent business that makes them available as a metered service offering, charging a tiny usage tax (say, $0.001) each time it's invoked. The technology to do this is certainly available: code the algorithms as a microservice or a blockchain smart contract, and deploy them in an elastic cloud environment (as usage will tend to spike with pay cycles).

To the HR platform company, there are a lot of good reasons to do this. It monetizes a non-value generative activity (nobody subscribes to a payroll system because their implementation of the witholding rules are better than everybody else's). It throws off utility-like revenue streams that move in lock step with the broader job market. It disaggregates a staid HR utility function that needs to be deployed infrequently (potentially only as often as once per year, when new tax laws come into effect) from more innovative ones that are more exploratory in nature and demand frequent deployments (such as time entry performed through emerging wearable tech). It separates a legacy and risk-averse tech culture from a cutting-edge risk-prone one. It takes a high fixed cost for maintaining a non-differentiating characteristic off the P&L (teams maintaining rule-heavy legacy code are rarely inexpensive). It's stable cash flows would be attractive to debt finance, better aligning capital investment in HR technology. It removes asymmetric risk that can be more accurately insured (smothered in a platform, correct calculations offer no financial upside, while a faulty calculation exposes it to costly litigation and reputational damage).

True, it eliminates a barrier to entry of future competitors. And, while the market would support a handful of utilities to prevent monopoly, thin competition would give those utilities oligopic pricing power. But it creates a one-time financial windfall for the first movers, laggards would be pressured to subscribe by shareholders demanding the same structural benefits to the income statement, and low switching costs would keep utility pricing power in check.

Given that tech is in a period of both cheap capital (interest rates remain low, VC levels remain high, and companies such as Alibaba and Facebook can make acquisitions inexpensively with their own high-priced shares) and rapid growth (growth in consumption of tech products such as social media today mirrors growth in consumption of manufactured goods in the 50s and 60s), it's little surprise that we're seeing a return to industrial strategies past. But technologies like microservices and blockchain could be the modern equivalent of "core competencies" to sweep through businesses. Blockchain proponents already champion the potential of decentralized autonomous organizations (DAOs). With MBAs now eschewing investment banking in favor of tech companies, financial engineering of this nature isn't too far away.