The phases of digital transformation

Enterprise leaders are learning that digital transformation isn’t really about transforming from one thing into another—there is no end state.

Rather, it’s about gaining the agility to continuously evolve—not to just bolt technology onto the existing business but rather to leverage technology to change how the business operates, top to bottom, today and in the future.

This distinction is both important and, surveys suggest, challenging. How exactly does a business gain the agility to continuously evolve? Many enterprises may struggle to answer this question, given that research firm IDC has found that 59 percent of businesses are at a point of “digital impasse.”

As with almost every ongoing process, the key is to start small, with the right emphases and healthy habits, and then to grow. If you run a 10-minute mile today, you’ll need to be able to run an 8-minute mile before you can hope to eclipse six minutes.

Likewise, taking control of a company’s digital evolution often involves several phases, each with its own ingredients. These ingredients may change over time (they involve cloud technologies, application programming interfaces (APIs), and machine learning today) but the principles behind the phases are constant: speed from small teams, scale from ecosystems, and adaptation from constantly-refined data and intelligence.  

Phase 1: small teams, faster innovation and cloud infrastructure

In many companies, large teams move monolithically. These teams often collect project requirements months before development starts. Governance and funding processes are usually heavy, restricting how easily the project’s scope can be changed to adapt to shifting consumer needs, market conditions or competitive dynamics. More often than not, software built this way ends up rife with interdependencies, making it difficult to easily expand the project or use it in new ways in the future.

To begin making strides, a company should consider replacing its large, heavily governed IT teams with many small, lightly-governed teams that can move and innovate faster. Rather than placing an enormous amount of time, money, and labor behind a single monolithic fat application that may rely on outdated or untested assumptions, these teams should release scores of minimum viable products, then let customer feedback guide development, scaling efforts up or down as appropriate.

To create this kind of fast-moving operation, a company will almost certainly need to utilize a modern cloud architecture and decentralize its operational model to place more autonomy within individual teams. Today’s most successful enterprises are increasingly embracing lightly-coupled architectures; modular, granular application design; and funding and governance models designed to let teams take advantage of these tools to work fast, with neither bureaucratic obstacles nor backend complexity getting in their way.  

To be clear, the technology that fuels IT will always be complicated, but that complexity doesn’t need to set the pace for everyone in the business. Organizations can use the cloud to abstract applications from infrastructure, change service behavior and traffic flows without changing code, increase overall agility, and empower developers to work more autonomously and quickly to create business value.

Phase 2: ecosystems, scale and APIs

APIs sit at the core of modern, cloud-first application architecture. At their most basic, they connect systems, but when fully utilized, they abstract backend complexity into an interface that developers can actually use. Because APIs make data and functions available in an easy-to-consume, infinitely scalable manner, an API initiative can accelerate internal development efforts—but that’s just one use case.

Increasingly, businesses are also making their APIs accessible to external partners and developers. This allows those businesses to participate in digital ecosystems that may include billions of other participants. These ecosystems can help a company maximize proprietary strengths while leveraging other participants to fill gaps, provide scale, and open access to new markets.

For example, a business with valuable weather data may make its APIs available to ecosystems of external developers, who leverage that data to make a variety of apps. Some of the apps combine the weather data with APIs from other providers to produce entirely new things, and virtually all of them occupy various platform ecosystems. The API provider benefits because its services have scaled in a way never before possible, the developers benefit because they have richer resources and better tools to work with, the platforms benefit because they remain a hub for transactions between suppliers and consumers, and users benefit because they have more developers competing to satisfy them.

There are two basic ecosystem plays: to be the gravitational center everyone else resolves around, à la iOS, Android or one of the major cloud platforms; or to be one of the participants leveraging those gravitational centers. To make the first play, a company usually needs to be among the first in its industry to capitalize on a major disruption—and to be nimble and well-monied enough to sustain early advantages once competitors start to catch up. The latter approach is somewhat more accessible, though to succeed, a business will still need to invest deeply in not only its APIs but also the developers who turn APIs into revenue-generating apps and experiences.   

Phase 3 and beyond: learning, adapting and machine intelligence

The point of moving fast is to continually improve the products and services offered to customers—and to do that well, a company needs intelligence with which to iterate its current offerings. To an extent, this is about establishing feedback loops between customers and the business, such as proactive monitoring and analytics that provide insight into how a company’s software is being used. But more and more, this process is about making software better through machine intelligence.

Apps that respond to voice commands are becoming more common and refined, for example, and image recognition has reached the point that it’s not just distinguishing pictures of dogs from pictures of muffins (which can be pretty hard for a person, let alone a machine) but also helping to diagnose types of cancer. It’s still early in the mass application of machine intelligence—which means it’s also the period when those who shape the technology will be separated from those who just imitate the leaders.  

Machine intelligence can take many forms. For example, there’s artificial intelligence in which the machine makes a decision, such as a security algorithm neutralizing a bot, versus intelligence amplification or augmentation in which the machine helps a human make a better decision, such as a digital assistant suggesting the user leave for a meeting. Either way, machine learning is likely underneath the user interface—and to excel at machine learning a company needs four things: data, algorithms, compute power, and talent.  

Data is something many businesses already have—but developing the others can be time-consuming and expensive. Luckily, these resources are becoming democratized; many of the top machine learning companies are making their algorithms available via relatively user-friendly APIs, and the cloud makes affordable computer power more accessible than ever before.  

Everyone needs to be fast

Many companies have grown accustomed to operating like gazelles among lions. When a threat emerges, gazelles don’t need to be the fastest or most agile to survive—they just can’t be the slowest.

But this approach just can’t keep up with the pace of business in an age of ubiquitous connectivity and proliferating devices. Today, the disruptive threats—the lions—feed on far more than just the one or two industry laggards. The particularly voracious threats can consume an entire herd! Just think of the way stock markets react when Amazon shows interest in a new industry.

To position themselves for success as the pace of digital change continues to accelerate, companies need to move fast with smaller teams, leverage the cloud and APIs to empower those teams, embrace ecosystems for access to new markets, and invest in machine intelligence to add richer features that will delight customers.