HOW WILL IT get in shape to meet the challenges of digitization? What help will be available to analyze the enormous data volumes? And what type of organization does a company need to succeed in the digital world? This chapter explores the foundations for this success.

Speed is the hallmark of the digital world. Fast development times, short cycles, and rapid change are factors that overwhelm most traditional organizations. The first section of this chapter describes how companies can build a two-speed IT architecture. This involves a robust IT system deployed for stable functions with minimal need for intervention, and an agile, fast IT system for everything aimed at the rapid pace of digitization. We then move on to data, the new gold. Big data and advanced analytics—vast accumulations of unstructured data and its intelligent analysis—are critical success factors today. And since data theft can jeopardize success, we also describe ways to protect it. With the advent of the Internet of Things, software is making its way into everyday objects, and is also covered in this section on IT and technology.

However, digitization is also changing the culture and organization of companies, especially with the emergence of the chief digital officer (CDO), and we explain what this role does. Companies in the digital age need agile, cross-functional, flat hierarchies, and we explain how to go about it. Everyone wants digital talent, and we describe how to find it and retain it. Also in the new ecosystems, most companies work with a network of partners, whose interaction needs to be managed, and the final section of this chapter describes how.

6.1 TWO-SPEED IT: ACCELERATING THE PACE FOR THE DIGITAL AGE

Three Routes to a Two-Speed IT

Essentially, companies have three options for establishing their agile IT system: They can build on something that already exists in their organization, they can start from scratch, or they can acquire a company with the necessary skills.

Travel group Thomas Cook took the first route. It isolated the customer-facing areas of its IT organization, separated them from routine tasks, and, in so doing, gradually established a two-speed IT architecture. For this solution to work, there must already be processes within the IT organization that are worth transforming. Additionally, it requires significant investment in middleware systems to decouple customer-facing processes from the back end.

Starbucks decided to start from scratch. The firm was unable to identify a suitable starting point in its existing IT system, and therefore established a completely new IT organization separate from the old one, to take on the agile role in a two-speed IT architecture.

U.S. insurer Allstate was in a rush, so it acquired a company with an agile, fast-paced, and innovative IT team. The advantage of this approach is that Allstate lost no time with a long-drawn-out transformation process, or with establishing an agile IT structure over an even longer time. Instead, it could immediately start work on developing attractive digital offerings for customers. Companies that choose this route are convinced that the costs of integrating the new organization into the existing one are lower than the additional profit leveraged from fast market entry.

Companies that build an agile IT can reckon with a lead time of around 18 months before the new system delivers valuable contributions. This type of restructuring revolves around a strong crossover of business and IT. All processes in which fast changes count or that enable the company to set itself apart from the competition will in future be the responsibility of the agile IT team. As such, these IT systems need to be modernized and adjusted accordingly, and strict governance is required to manage this change. As a first step, a close interface must be established between business and IT. Cross-functional teams of marketers and programmers must be formed, products and services must be organized into teams, and a product owner must be appointed for each product and each service to coordinate developments and communicate with the organization. The product owner acts as a coordinator and moderator, but not a manager.

The more independently the teams are able to deploy their talents, and the more independently the agile IT is able to act in the organization as a whole, the faster the company will be able to deliver attractive customer offerings. Discussions with suppliers and consultants who plug holes in the skills of the organization and offer even greater flexibility also run in parallel to this process. The team has shared goals that are achieved in iterative sprints. These teams develop apps for customers, while other teams ensure the stability of the underlying IT architecture on which the apps are built.

The Agile IT Often Accelerates All Elements of the Organization

The fast IT organization typically concentrates on processes in which there is direct interaction with customers and where rapid adjustments are often necessary to remain competitive and meet customer requirements. In many cases, the agile IT element even spurs on the IT engineers who work on the more stable IT, resulting in a self-intensifying system that in turn can accelerate the entire IT transformation.

A major bank, for example, successfully implemented a two-speed IT structure. It soon became apparent, however, that the apps used by customers to move money or view account information also necessitated changes to the back-end systems that provided and modified the customer data. As a result, not only were the customer-facing systems made faster, but the systems in the processing department also became part of the fast architecture. This also required changes to the delivery model to support the fast cycle times while managing the interdependencies.

This could prove the beginning of a move toward a completely agile IT organization. The advantage is that employees who don’t feel up to the challenges of agile IT can remain in the steadier but no less reliable departments with the gentle transition process over a prolonged period. The aim is to achieve an IT architecture that mirrors the architectures of digital giants like Amazon, Facebook, and Google, where technology is fully integrated with business targets and they work in unison to identify, establish, and meet business goals.

This is where digital product management comes in. Managers who lead a development team are not responsible just for an app and its use, but possibly for the entire purchase and payment process through to order confirmation and notification of delivery dates. This kind of end-to-end responsibility leads not only to shorter lead times, but also to increased quality of the end product or service. Managers’ job profiles now include numerous tests and experiments in which market strategies are trialed with different objectives. Customer data and feedback are collected and analyzed, and the results are immediately incorporated into further developments. Other tasks include developing a set of key performance indicators (KPIs) used to measure the success of the team.

The combination of an agile, fast-response IT structure, a digital product manager, small teams of entrepreneurial talent, and cutting-edge scalable architecture has formed the backbone of the astonishing success of the likes of Amazon and Google. Today, companies from the analog world can also benefit from the strategy.

Holger Hürtgen on the path to better, data-driven decisions.

How can Amazon recommend a novel that actually interests me after just a few book purchases? How does Spotify know which hit song will appeal to me next after just a few downloads? And how is it that online fashion retailers are able to recommend clothes that I like even though I’ve only viewed a few of their pages? We are no longer surprised by this. We just expect it. But very few of us know how it works.

Online stores and apps like Spotify collect accessible data about their customers: what we buy, what we look at, which website we came from, which device we are using, what time of day it is, how often we visit, and what else we look at. Advanced analytics software then sifts through this vast sea of data searching for patterns, which the company can then use to make predictions about our behavior. Online retailers, for example, love lookalike modeling, which is the search for statistical twins, on the assumption that someone who leaves behind the same data trail as we do will have similar needs to ours. Once one twin likes a song, video, or pair of pants, the other immediately gets a recommendation for the same item.

Big data and advanced analytics aren’t just useful for retailers. They also help manufacturers and service providers make better, fact-driven decisions. Despite this, many managers and employees still base their decisions solely on personal experience or statistical planning. In the past three years alone, more data has been generated in the world than in the rest of human history. Companies that are able to leverage these data volumes can gain an immediate competitive advantage. Big data and advanced analytics deliver the tools for better, faster decisions, and although technology plays an important role, it is by no means everything. Added value is generated where the technological possibilities and corporate goals intertwine. Because the very core of the company is affected, from the corporate strategy and the business model right through to its growth prospects, once again the executive management must be on board. A big data strategy needs to be driven from the top.

The Three Success Factors of Big Data and Advanced Analytics

For companies that want to drive their businesses forward with big data and advanced analytics, three levers are needed: the company must formulate a vision outlining its objectives; data analyses should be tested in defined applications; and the technical, organizational, and skill frameworks must be in place.

Every successful transformation into a data-driven company starts with a vision. A target scenario is outlined that not only formulates an overarching target as a business case, but specifies measurable targets for specific applications, from customer retention to production optimization. The vision also includes a transformation path that splits the journey through the data landscape into stages, and sets deadlines for achieving these milestones.

At this level, the company defines the points in the business process where it wants to implement data analyses. The most fascinating areas are where companies target the emerging ecosystems and make the decision to enter new territory, from smart homes to digital health. However, dramatic improvements are also possible in standard areas. Big data analysis can be used to optimize customer contact (e.g., to avoid contract terminations), in direct customer communications, and in cross-selling. Internal processes also benefit in ways such as optimizing machine utilization, predictive maintenance measures, and employee retention.

Ideally, the implementation of the big data strategy will start with a manageable but complex application, and one that promises high returns. If results are achieved quickly here, the entire transformation project is given fresh impetus. And once several such subprojects start delivering results quickly, the entire project funds itself.

The Foundations

These partial successes can be achieved, however, only if the company has already laid a stable foundation of technology, processes, and talent. At first, only the basics are needed. Expertise and skills will grow with implementation. Five factors are important to ensure a successful transformation.

Data. In the digital economy, data is the new gold. Google is the perfect example of what is possible. The more data a company collects, the more reliable its data-driven decisions will be. For many years now, companies have been collecting data on a large scale, but they lack the methods to bring it all together and analyze it as a whole. In some cases, the reasons are technical where highly ramified corporations store data in different IT systems. In other cases, the reasons are organizational, where marketing, production, logistics, and HR all operate their own silos.

Analytics. After data collection comes the analysis. Analytical approaches are required with differing perspectives: these are known as descriptive, predictive, and prescriptive analytics. The first explains what happened in the past, the second predicts what will happen in the future, and the third recommends decisions that can shape the future development of the company. New algorithms in machine learning systems with artificial intelligence open up new opportunities.

Pioneers like Amazon and Google make their algorithms publicly available. Tensorflow, which Google originally used to improve its search results, can now be used for image recognition and automated responses for call centers. No one, however, is allowed access to their treasure troves of data, an indicator of how these companies like to differentiate themselves and what they see as a general commodity.

Tools. Data analysts need professional programs that help them manage and structure the data volumes. This can be software such as SAS or IBM Modeler that enable them to quickly integrate machine learning components with the help of graphical interfaces. More companies, however, are using public, open-source software like R or Python for these tasks because they develop so quickly with the active members of their large communities. Also, they contain all the advanced algorithms required. In turn, the users, as the target group of the analyses and recommendations, need simple-to-use software that visualizes the results. Companies buy programs like Tableau to do this, or develop software themselves as decision-making aids.

Translators. In addition to management as the recipient of the analyses and the technical talent that processes them, companies on the transformation journey also need a third group of employees, translators. They need to be able to translate the business requirements to the data scientists and data engineers, and also explain to the decision makers which analyses are possible with which statements. The job requirements for this are high. People who take on the role of translator need to understand the digital space just as well as the business world. In general, companies often underestimate just how much expertise their employees need to successfully use big data and advanced analytics. Instead they turn their attention to strengthening their IT systems, which is also important, but much easier to solve.

Processes. Data, analytics, tools, and employees are all united by processes. It’s worth remembering that the value chain is only as strong as its weakest link. If incorrect or poor data is collected, even the best analytics are worth nothing. The results will be nonsensical. As the early computer scientists put it, garbage in, garbage out. Even if the data and analytics are top quality, things can still go wrong if the process isn’t foolproof. For example, if an agent in the call center of a wireless provider does not follow a recommendation provided by the virtual assistant based on customer data during the course of the call (the next best action) because he or she finds the recommendation difficult to understand, the data, analytics, and tools remain useless.

When used correctly, however, big data and advanced analytics can solve the trickiest of problems. Take an example from the world of consulting. A supermarket was considering expanding its range to include organic products and sugar-free, gluten-free, and lactose-free items. The problem was limited shelf space. Which products should give way? The obvious, simple solution was fundamentally wrong: if stores kept only their most profitable items, important anchor products would be lost, which are vital for retaining customers. Using entropy models and hierarchical clustering, the company developed a model that enabled the retailer to accurately predict the migration effect of an item being removed from sale, as well as the increase in margins following the inclusion of new products. The supermarket chain now uses this model to calculate its optimum range, and is growing twice as fast as its rivals.

Artificial Intelligence: The Next Disruptor

If Google and Amazon are proved right, companies will soon no longer be able to differentiate through smart algorithms; instead, these clever formulas will be shared as general commodities. All the same, the data scientists who create these algorithms needn’t fear for their jobs. They will just see a change in the way they work. Instead, their work will involve selecting the best algorithms for the company and combining them in the best way to mine the data.

The next great leap in development is expected to come from artificial intelligence. Machines and their algorithms will increasingly be assigned tasks that are too complex for humans to manage or that the machines can perform more efficiently. Even today, machine learning systems are being used in industrial pilot projects. In 2016, a Google algorithm, AlphaGo, beat the world champion of the board game Go—and that’s just the beginning.

James Kaplan on the seven steps toward effective safeguards against hacking.

In 2016, resourceful hackers removed $81 million from Bangladesh’s central bank after it joined the global banking communication network Swift; the hackers deployed malware to redirect the money transfers of Bangladeshis. A year earlier, hackers had gained access to the servers of the Sony film studios. They published the salaries of the executive directors, revealed the personal details of employees, and attempted unsuccessfully to blackmail the film studio into canceling the release of a film that made fun of North Korean dictator Kim Jong-un. And in 2014, a talented group of digital natives stole the code that the transponders in the electronic keys of Volkswagen cars use to open doors and start engines, the perfect tool for car thieves.

These are just three examples from a flood of incidents, but they illustrate how vulnerable the digital economy is. The risk is considerable: a McKinsey study estimates that if companies and governments do not take effective measures to combat cyber risks, global losses could run to around $3 billion worldwide by 2020. As the economy becomes increasingly digital, attackers are finding ever more entry points.

They attack from different sides and have completely different interests. Sometimes nation-states are behind the attacks, either to steal information to boost their own economies or to weaken political opponents. Sometimes competitors launch the attacks to steal a rival’s technology and use it themselves, or simply to demonstrate how poorly their rival protects sensitive data and to inflict serious damage on its reputation. Other groups have an ideological agenda. These “hacktivists” may want to uncover injustices carried out by the target company against others, or to promote an ideology, be it anticapitalism, nationalism, or ecological policies. Some simply want to show they can do it. Attacks can also come from inside by frustrated or bribed employees or workers driven by their conscience.

Seven Steps toward Cyber Resilience

In a worldwide survey of managers, two-thirds agreed that cyber attacks posed a serious problem, and could result in significant strategic implications. Only 5 percent believed that their company was truly competent in six of the methods listed for preventing attacks. And 80 percent were worried that the hackers were learning faster than their company. The countermeasures even have a negative knock-on effect on business. Security measures, for example, can delay the launch of new mobile features by an average of six months. Three-quarters of the managers surveyed stated that the productivity of their customer-facing employees had fallen because security requirements had slowed the sharing of data.

There is widespread agreement that the previous security models are overwhelmed. After a time of carelessness until around 2007, IT departments established a type of security network with strict processes and secure technical infrastructures. But these are now increasingly full of holes. To meet today’s risks, it’s time for a solution that focuses on security even at the process design stage. The best way for companies at risk to respond is with a program aimed at cyber resilience. Resilience is a term that business science has borrowed from evolutionary biology. It describes the ability of a system to absorb disruptions and shocks, and to carry on working successfully despite them. Companies can achieve cyber resilience in seven steps.

Priority Lists: Which Data Represent the Greatest Business Risk?

Very few companies have a clear idea of which business data is the most important. Security teams therefore need to work with management in the first step to examine the entire value chain and assess where the greatest risks lie. Is it the data for the design of a new product, a self-learning manufacturing process, or sensitive customer data, the loss of which would lead to a maximum credible accident?

Banks and insurers have assessed their risks in this way for years. They call the approach the “crown jewels” program, and it could be used as a model for many other industries.

Customer-Facing Managers Must Be Part of the Team, and Recognize That Data Is an Asset

Only those people who work with data truly understand its value and will take the topic of security seriously, which is why employees need to be trained on the subject.

At Microsoft, founder Bill Gates personally took this upon himself early on, in 2002. The security of products is the absolute top priority, he wrote in an urgent reminder to all employees. If it comes to the choice of whether a new product should be given additional features or be made more secure, he wrote, security must always be the choice. In 2003, Gates introduced Patch Tuesday at Microsoft, an initiative to continuously patch security gaps in the software.

Cyber Resilience as Part of Risk Management

Cyber security forms part of the company’s risk, and must be managed as such. Risk assessments of online attacks must be integrated into the company’s other risk assessments, presented to the relevant members of management, and discussed at executive levels.

To ensure resilience, prevention of online attacks must form part of the planning of all processes. Even in the early days of Industry 4.0 and the industrial Internet in 2012, GE stated its intention to make security components an integral element of the design of its machines, software, and networks. And in 2014, the group acquired the security specialists Wurdtech to bring additional skills into the firm.

Cyber War Games: Continuously Testing Prevention Systems

Why sit around waiting for ill-intentioned hackers when you can have someone else perform an attack to expose weaknesses before malicious hackers do? United Airlines took exactly this approach, and offered free air miles to clever hackers who could discover weak points in its programs as part of its “bug bounty” initiative. Barclays Bank went so far as to hire an entire department of in-house hackers who attack the bank’s IT systems and then immediately repair any weaknesses. The CEO of a European electronics group even stepped down for two days to take part in a cyber war game to simulate online attacks on his organization.

Companies that understand the reality of the situation also practice their response to a successful cyber attack. If data breaches become public knowledge, the wrong statements can have serious consequences. The correct responses must be prepared, not just for IT managers, but also for marketing, customer service, and of course for PR.

Security Technology: An Integral Element of the IT Architecture

Operating systems, communication logs, and applications are established elements of the IT architecture. Each of them can become a security risk if it offers access points for possible attacks because of poor configuration, testing, and maintenance. Security elements must be built into all components, whether hardware, middleware, or application software, and their resilience must be continuously tested and refined during the development process. Together, all of these components mean a huge volume of potential security problems that can be identified only through persistent testing and ongoing maintenance.

With the trend toward rapid digitization, many companies in recent times have hastily introduced new technologies for which they lack the necessary administrative skills, and are unable to understand the technologies’ interaction with the existing architecture. Budgets are reallocated from the maintenance of the old IT systems to the establishment of new digital skills, which makes sense from a business perspective but often has serious consequences for the medium-term security of the IT architecture.

One important step is to introduce different security zones. A European sports apparel manufacturer, for example, introduced a “play zone” in which online campaigns can quickly be created and implemented. The play zone is isolated from the existing systems with its own security zone. This means if security problems occur, the campaign can be quickly stopped or even deleted without the other systems being affected.

Another key element is ownership of the security elements. Often, the enemy is internal, with responsibility disputes among IT, the security organization, and product development tying up resources and budgets. Security technology, however, should be given top priority in the governance of each company. For example, GE firmly anchored an IT security function on its board of directors. Business units and the firm’s head office are regularly and systematically subjected to security audits. According to Bill Ruh, head of the GE software division: “At GE, we focus on software platform security, protect critical infrastructure elements, and help our customers to perform reliable and secure transactions online.”

Protection Levels: Not All Data Needs the Same Level of Protection

“He who defends everything defends nothing,” said Frederick the Great, who knew a thing or two about attack and defense. Likewise, different processes can be given different levels of protection according to their position in the list of priorities, by employing greater or lesser encryption or requiring passwords of different strengths.

Banks, for example, only employ standard verification procedures for their online banking customers when it comes to routine queries. However, if customers want to transfer larger amounts or make unusual transactions, the banks often require additional codes sent to the customer by text message.

Active Defense, Preferably Before an Attack

In most cases, plenty of information is available about potential attacks, from both external and internal sources. Companies must in the future develop the ability to combine all the available information into a single risk profile, and then construct targeted firewalls that protect their systems from intruders. Once these protection measures are in place, the company will be well fortified and able to fend off attacks. In 2011, for example, defense group Lockheed Martin initiated the Nexgen alliance for the early detection of cyber threats. Network partner EMC, a cloud-computing firm, strengthened the alliance with its acquisition of NetWitness, a specialist in real-time web tracking, automatic threat analysis, and measures to combat illegal hacking into computer networks.

To establish greater security against cyber attacks, positions of responsibility throughout the entire company must be involved: operational managers when it comes to assessing which data is the most valuable, compliance officers when it’s about assessing the possible risks associated with losing customer data, HR when it’s about deciding which employees have access to which data, and the purchasing department when it’s about negotiating security requirements with suppliers that need to connect to the company’s IT system. To coordinate such a complex effort, top management must take sole responsibility.

Just like the World War II Enigma encryption machines, it’s a perpetual battle between coding and decoding, ciphering and deciphering. The race continues, faster, more skillfully, more shrewdly, and more uncompromisingly than ever before. As the value chain makes its inexorable shift to the digital world, the potential gains for attackers grow.

Mark Patel on the five guiding principles that can help traditional hardware manufacturers develop a software strategy.

In October 2016, Ford hired 400 employees from BlackBerry, doubling the connectivity talent dedicated to delivering more software and services to Ford customers.¹

With the transition to Industry 4.0 with its connected machines and sensors that continuously collect data and transmit it to central servers over Wi-Fi, even machinery and equipment manufacturers need to upgrade their products to digital. Other industries that were strictly analog have already made the switch. Today, the average car, for example, has more lines of software code than an Apple MacBook. By 2012, smartphone makers already employed twice as many software developers as hardware developers. And a good two-thirds of all machinery and equipment manufacturers now also offer their customers software solutions.

So how do companies successfully tread the path from analog products to digitally ready products? How do they equip their machines with the right software? They need to bring in new skills and talent, as well as new organizational structures and processes. Ten steps lead the way toward embedded software, five of which center on strategy and implementation. We address these next.

Implementation: Recruiting Talent and Expertise

At first, traditional industrial companies wouldn’t appear to be at the top of the list of potential employers for digital talent, who prefer to head for the Googles of this world or successful start-ups. Industry is perceived as lacking the freedom needed for successful development work or the same levels of technical tools and challenges. How can companies overcome these prejudices?

• • Attract key personnel. If an industrial company is able to attract a star of the developer world, not only does it demonstrate that it takes its software business seriously, but also gains access to digital talent through that person’s network.
  • Participate in the software ecosystem and evaluate potential acquisitions. The advantage of leveraging an ecosystem or acquiring another firm is that the new team can start working productively from day one, which is important when time is of the essence. If the team has been working together a long time, there’s less risk of migration, because most developers don’t want to leave a team that collaborates so successfully.
  • Create the conditions that digital talent expects. This doesn’t mean just a high salary. We investigated what digital natives really value. They expect an excellent job. For most, this means working with the very best technology on challenging tasks, in internal networks, and with the necessary freedom. They expect strong top management that has open contact with the developers. And they want to work only for a company that has an excellent reputation in the industry, has dynamic and socially responsible working practices, and offers workplaces where people feel at ease. And, of course, the salary must be right. On top of the basic salary, there should be short-term performance bonuses and a long-term asset growth plan.
  • Allow groups of developers to work independently. Software developers are used to working in close collaboration with colleagues on projects that usually need lots of test runs and modifications to programs. Since their style of working tends to differ from the work flows for machine development, it’s generally a bad idea to integrate the programmers into the existing organization. Instead, the system works best if the software team forms its own business unit with its own management and its own processes.
  • Employ an integrated communication system. Many companies still follow a sequential development process in which the programs are only written once the hardware has been built. This wastes time. With new digital tools, software developers today can begin long before machine construction is finished, testing their programs on virtual prototypes. To ensure that the entire development process runs smoothly, hardware and software engineers should provide each other with regular progress updates, develop the requirements profile of the new product together, and coordinate their targets and schedules.

For most companies in the manufacturing sector, there is no choice. Their machines must come with software; otherwise, they won’t be competitive. Our recommendations will help on the road toward the digitized product, but are not a magic formula—expect the unexpected along the way.

Steve Van Kuiken discusses the art of the digital transformation with CDOs.

Offense is the best form of defense. We hear it time and again. That’s how Bill Ruh, chief digital officer (CDO) of GE, sees it. “If you don’t figure out how to get more productivity and efficiency in your products, someone else will. In the end, every major company has to be a software company, because if you don’t own this asset, you stand to be disrupted,” he says. When Ruh joined GE, he found a heavy focus on analytics on the company’s own machines. He described this as a defensive strategy. His focus was to make GE go on the offensive. “GE would help customers be proficient on all machines, not just GE’s,” he says.

Bill Ruh is one of what is a small but growing group of chief digital officers in businesses today. At Starbucks, Adam Brotman drives digitization forward; at tech giants IBM and SAP, it’s Bob Lord and Jonathan Becher; and at elevator maker Schindler, Michael Nilles is in the hot seat.

CDOs must be multitalented. They need to be strategists and digital natives, and have a strong understanding of customers and employees. Developing a digital strategy is just one part of the job description; above all, the CDO must also be a skilled networker, both internally and externally. At an external level, he or she will identify the most exciting technological developments and most promising start-ups, and at an internal level, the CDO will convince the entire organization of the merits of the digital transformation. Just as important as knowing what can be expected of employees in all things digital, the CDO must also understand what customers expect from the company, and what technology they will feel comfortable using.

Developing Internal Talent

It’s an important philosophy for Max Viessmann, who heads up digital at his HVAC and refrigeration company. “You need to invest time to get everyone on board,” he says. “It mustn’t be about the digital winners and analog losers.” Many members of the digital team came from the existing workforce. “The external proportion isn’t actually that high,” Viessmann says. Anyone who demonstrated an affinity for the topic was able to join one of the project teams, and those who proved themselves had the opportunity to stay permanently. Viessmann wasn’t concerned with their qualifications. “We have machine engineers, business administration graduates, social scientists … ultimately, it doesn’t matter.” What counted was their attitude. “I spend 30 percent of my time explaining to our employees we’re not all digital natives—that’s what we’re working toward,” Viessmann says.

Digitization has also long since arrived at elevator manufacturer Schindler. “Three or four years ago, we were discussing things defensively: What can happen to us? Is Uber coming to disrupt our industry?” CDO Michael Nilles said in an interview with the Wall Street Journal. “We now ask: Where are the opportunities?” Schindler is connecting its global fleet of elevators to the Internet of Things as part of its digital transformation, with the aim of offering new services to direct customers and their elevator passengers. Schindler has allied with GE and Huawei to develop components for the Internet of Things. According to Nilles, the data collected forms an indispensable platform for developing new services. He believes that in the future, companies will be differentiated by the software offerings they provide for their hardware: “You are not going to distinguish yourselves anymore with a product.”

Chief digital officers with the necessary powers are becoming increasingly prevalent, and renowned companies of all sizes and across all industries are adding the role to their management boards. From tech giants like Cisco and IBM continuing through Volkswagen, CVS Health, L’Oréal, Starbucks, Williams Sonoma, and all the way to the city of New York, organizations worldwide are adopting this function.

A CDO with reach and authority—a Steve Jobs for every company—can help to break down the functional silos, and shape the journey toward digitization as a permanent disruptor.

Julie Goran on autonomous teams, agile sprints, and product owners.

Is it possible to have an organization model that has no fixed structure, and is constantly evolving? Yes, and it works very well in fact. ING, the Dutch banking group, embarked on a journey to transition from a traditional fixed organization to an agile model. In the headquarters organization, the staff was split into some 300 nine-person squads that are multifunctional teams that work toward a specific goal. The squads are grouped into 13 so-called tribes. Each tribe has a dedicated purpose, which is broken down into subpurposes and deliverables for each squad. Squad members follow a set of agile ceremonies to align and work together. There are no managers; they all learn together, work through problems together, and make decisions together. The successes and failures of all the tribes are shared in quarterly business reviews (QBRs), which also serve as platforms for cascading broader business goals and objectives to each of the tribes, and aligning the priorities, resources, and focus areas of the different tribes. Many core processes such as budgeting, performance management, and procurement processes are now fundamentally different in order to support an agile operating model. The success of this new model is very impressive—software releases are now on a weekly basis rather than five or six times a year, and employee-engagement scores are up multiple points.³

Agile companies are almost tailor-made for the digital age with their simple, stable, and effective structures that enable extreme flexibility and fast response times. That is precisely what’s needed, because along with its opportunities, digitization also brings with it plenty of challenges. It makes the business world more volatile with its sudden leaps in progress; it makes it more uncertain, more complex, and more ambiguous. The old way of organizing companies is coming under increasing strain. Hierarchies are too slow for the digital pace, rigid budget plans can’t keep up with demand, and before a project even starts, the constantly growing complexity overwhelms even the most capable executives. In this new age, innovations can destroy established business models, competitors can form new ecosystems outside the industry, and businesses can grow at incredible speeds. The new age needs new organization structures.

The agile organization combines a stable framework with dynamic skills. Again, top management sets the direction. There’s a simple, common reporting structure, and lean processes and values are specified by the head office. Then there are the dynamic components: a culture of fast learning based on trial and error that grasps opportunities and tries them out while willing to accept failure, teams free of hierarchies that work autonomously, and employees who take responsibility.

How Do Agile Organizations Work?

As the first industry to practice the agile method, the software industry laid out three principles in a manifesto. As stated in the Manifesto for Agile Software Development at www.agilemanifesto.org, for development work to succeed, individuals and interactions are more important than processes and tools, just as working software is more important than comprehensive documentation. Collaboration with customers is more effective than any amount of contract negotiations, while responding to change is better than following a plan. It’s a manifesto against bureaucracy and in support of creative freedom, born of the experiences of long-drawn-out projects that seemingly never end, and once they do finally end, the market and customers have already moved on.

Agile companies refrain from the quests for perfection that are so time and cost intensive, and base their models on companies from the digital world, which bring products to market that at first often have only limited features. Agile companies use heuristic methods to complement and improve their products with continuous market testing cycles. Products can be launched with only a few core functions, an approach known as the minimum viable product (MVP). A prime example of a company that takes this approach is U.S. automaker Tesla, which constantly improves the performance and features of its cars even after purchase by offering software updates. This generates revenue because customers want to use these new functions.

It’s a real paradigm shift for many companies, because from now on, products are never truly complete; they’re continuously updated.

To work in agile mode, companies need a new organizational structure. In the future, the core of the organization will be formed by stable, product-based teams working toward their set targets largely autonomously. Depending on the task at hand, these teams will comprise employees from across all functions—including IT, design, marketing, controlling, right through to production, as required. The product owner plays a key role in the team. He or she represents the customer and the customer’s interests, sets priorities, allocates tasks across the team, and coordinates the remaining players, but has no managerial function.

The teams work in sprints that generally last one to four weeks, and always follow the same process. At the outset, the product owner explains the goal. The team then estimates the time and effort needed and what can be delivered by when, and then gets to work. Each day, the team members meet to organize the day’s work, resolve problems, and discuss progress. At the end of the sprint, the team presents the results to the product owner and all other parties with an interest in the product. Sprints always end with a review session: What went well? What went wrong? What can we learn? As a result, each individual team learns continuously, and so too does the organization, from the efforts of all teams.

Organizing product development into cross-functional teams is so successful because the teams are responsible for their products and services right from the beginning. Since the team is already made up of all the important functions, it doesn’t rely on other departments that could jeopardize delivery. This responsibility and autonomy are excellent motivators for the team members.

Many Employees Have to Take on New Roles

When companies adopt this type of organization, many employees have to learn new skills and take on new roles. Also, fewer managers are needed because the teams take on more responsibility. As a result, many of those managers who remain in the organization have to give up their familiar managerial roles and develop into coaches who pass on their knowledge. Managers who retain their managerial positions then lead through visions and values, acting as catalysts that bring together the mind-sets and actions of all stakeholders. And of course, they continue to manage the organization through the most important financial and performance KPIs, coordinate cross-team initiatives, and communicate with all key players, both those inside the company and those with an interest in it: employees, owners, and society in general. Many IT employees also find they have new roles. Whereas earlier their work was far removed from the customer in the back office, many now have to find solutions for customer requirements as the product technology managers of their teams.

The agile organization also redefines relationships with suppliers. Data analysis and software development will be considered core strategic areas for many companies. Instead of outsourcing this work, companies will attempt to do much of it themselves, often in partnerships or with flexible freelancers forming a network that needs to be built up and maintained. Budget planning will also become more dynamic, and will no longer be performed once a year, but instead will follow the methods of venture capitalists. First there will be a manageable budget for the development of a simple product that can be tested in the market. If this product is successful, more funds are gradually assigned. Conversely, if the product fails, the budget is very quickly reduced.

The biggest change, however, relates to the corporate culture. Instead of employing a top-down hierarchy and imposing regulations, the agile organization relies on trust. Self-organizing teams agree to defined targets, work without directives from above, and justify the trust placed in them by delivering results. At Netflix, for example, employee freedoms go so far that workers are allowed to decide for themselves how much vacation time they take. Management doesn’t get involved; instead, the individual just needs the approval of the team.

The trend toward more responsibility rather than control will increase in the coming years. As larger sections of the workforce form into self-governing teams, the traditional management structures will be rendered obsolete. For the employees in the teams, however, it means greater freedom than ever. Together, they will decide on working hours, training requirements, new hires, dismissals, and possibly even salaries. This leads to a better experience for customers and employees. At the same time, the company grows and is successful in the market.

Hugo Sarrazin and Satty Bhens on the upheavals in the labor market and the strategies deployed in the battle for scarce talent.

Digitization of tasks in the form of production robots, chatbots, and digital assistants has sparked fear among many employees, who are wondering if they will still have a job tomorrow. In 2013, for example, economists Carl Benedikt Frey and Michael Osborne caused a sensation when they analyzed 702 occupations in the United States and assessed the likelihood of these jobs being performed by robots or computer programs in the near future. They concluded that almost half of U.S. employees could lose their jobs in the next 20 years.

However, a different picture emerges if we focus on individual activities rather than occupations as a whole. A study by the McKinsey Global Institute shows that with current foreseeable technology, full automation will be possible in less than 5 percent of jobs. It does confirm, however, that automation will be possible for no less than 45 percent of individual activities currently performed by employees across various professions. This percentage could further rise to 58 percent if machine processing and analysis of natural language reaches an average human level. The Center for European Economic Research also estimates that 20 to 30 percent of academic activities could be automated. Although this ultimately means a similar volume of work becoming automated as predicted by Frey and Osborne, the prospects are quite different. Ideally, machines will assume the monotonous side of people’s jobs, leaving them more time to get on with more creative and fulfilling work.

However, not everyone can do that work. In view of recent findings, economists and social researchers predict a polarization of the labor market. They anticipate that creative, highly qualified labor will become more sought-after than ever before, and that people who provide personal services will retain their positions, but employees with average qualifications who perform mostly routine activities will gradually be replaced by powerful computers and robots.

What Can Established Companies Do to Become Attractive for Digital Talent?

Traditional companies don’t normally possess the sex appeal that start-ups do. How can they make up for this deficit? As experience shows, higher salaries aren’t the answer. So if remuneration is at the same level as at other competitors, then other factors count more, such as an inspirational mission and challenging and interesting work. Some large firms turn their disadvantage into an advantage, and tell potential candidates that they want to reinvent their entire organization to be ready for the digital age—just the job for new talent. And this is precisely the approach taken by GE when it explains its mission to digital natives: GE wants to become one of the world’s top 10 software companies by 2020. It’s all about storytelling. Companies that are able to outline their vision in an authentic and stimulating story can win the hearts and minds of talent.

Because birds of a feather flock together, some companies sign up well-known players in the digital scene who are sure to attract other talent.

Other companies take it even further, and, rather than recruiting big names, they simply buy the entire start-up—it’s known as “acqui-hiring.” In 2011, for example, Amazon acquired software developer Quorus, a company that brings together social media and online retail, and develops apps that allow users to get advice from friends before making a purchase. Amazon incorporated the entire team into its development structure.

New Ways to Search for Talent

HR departments often feel helpless when it comes to assessing digital talent, because their normal measures don’t work. Software developer Catalyst DevWorks analyzed the resumes of hundreds of thousands of IT systems administrators, and found no correlation between the quality of university degree and career success. So rather than concentrating on qualifications, the HR managers of the software firm developed their own tests, using clever algorithms to identify the IT expertise and skills of an applicant.

Since then, many HR departments now use online tests, games, and analyses to improve their recruiting methods, sometimes even using psychometric tests to determine how well a potential recruit will fit into the company’s culture. Often, really simple ideas work: one corporation uses a 30-minute test to compare the profiles of applicants with profiles of people who have been especially successful in the relevant field. Its number of bad hires has fallen significantly.

But how can companies track down digital talent? A good starting point is online platforms like GitHub, the biggest online repository of open-source software, where proud programmers upload their software along with their names. If a company identifies a particularly clever solution, it can contact the author—it’s already clear the programmer can do his or her job. Companies searching for talent regularly hold programming competitions on platforms like TopCoder, Kaggle, or HireIQ.

And if the talent can’t come to the company, the company can go to the talent. Many companies are establishing their own digital laboratories near leading universities or in cities that attract digital natives. Walmart, the world’s largest retailer, for example, based its WalmartLabs close to Stanford University.

How Can Companies Retain Digital Talent?

Since digital talent is in short supply, people with the right skills are constantly being targeted by HR consultants across all channels from LinkedIn to Facebook in an attempt to attract them to new pastures. This makes it all the more important for companies to bind their freshly acquired talent with a clear incentive program from day one. Tailored programs put together by specialists like the U.S. company LearnUp are now available to help ensure a smooth onboarding process. Google, for example, was able to increase productivity by 15 percent using this type of systematic training.

Digital technology even makes it possible to identify dissatisfaction among employees and thoughts of leaving. Predictive analytics examine social interaction in teams and raise the alarm if any relevant indicators are identified. HR can then step in and offer the employee a mentor, a new position, and perhaps even a promotion. Once again, Google leads the way when it comes to using digital technology in HR. The search engine giant established a People Analytics Unit, which examines how team dynamics and harmony affect output, and how to best stimulate the creativity of software engineers. The company boasts that its software has enabled it to significantly increase not only the productivity of individual employees, but also the productivity of entire development teams.

Anand Swaminathan on the art of managing a collaborative network.

Apple knows how to do it: Its HealthKit platform brings together players from across the world of medicine. Physicians, researchers, hospitals, and patients, as well as a wide array of clever app developers, all connect to its open platform. And just like a spider in the middle of its web, Apple occupies the central position, connecting the various players such as doctors who want to research illnesses like asthma or Parkinson’s disease, or offering Apple customers the chance to manage their conditions better using apps developed by the company or its partners. The commission that Apple receives from the sale of these apps is just a side revenue stream. What the company really wants is hardware sales and data, which is the new gold of the digital age.

HealthKit is a prototype of a functioning partner network. All players benefit—doctors and researchers because they no longer have to spend time organizing and managing a research network, iPhone owners because their smartphones offer even more useful lifestyle functions, and Apple itself, with increased sales and data. Apple is the pioneer of partner management; after all, it’s thanks to the ideas and apps of external programmers that the company has risen to being the most valuable company in the world.

And Apple’s digital example has now spread to the analog world: The complexity and pace of digitization has rendered the old tried-and-tested principle of do-it-yourself obsolete. In the rapidly emerging new ecosystems that are appearing in every industry, it’s impossible for a single company to take on all roles. A company that wants to break into the digital world needs to be open-minded enough to work with partners.

In the auto industry, there has long been collaboration between multiple partners. What is new, however, is digital thinking. The car of the future will exist in an ecosystem that includes not only traditional suppliers, but also completely new digital companies. The real innovation is the new interaction between numerous partners who present the car as a system consisting of hardware, software, and services. Vehicle developers work with start-ups, large digital companies, high-tech firms, and other digital service providers.

Siemens, for example, has allied with IBM and a small start-up named Local Motors, and is experimenting with an open network. Based on a crowdsourcing concept, numerous independent programmers and engineers designed a vehicle that was almost entirely produced using 3D printers at the customer location, making good on the start-up’s name. It took just two months from the decision phase and product design to final production of the first prototype—a fraction of the normal time it takes in the auto industry.

Coopetition: When Competitors Collaborate

There will always be things to negotiate where the interests of the network partners clash. In the digital world, this is especially true when it comes to sharing data. As such, each partner must decide for itself which data or elements of its data model it wishes to share with the other partners, and which data it believes is so sensitive that it can’t be shared. The partners also need to discuss how to distribute the profits. How will a company profit when its customers purchase products from a partner through the shared platform? How much commission will be paid to the network coordinator from revenue generated by a partner’s app in the ecosystem?

The collaboration not competition philosophy hasn’t quite reached all sections of management just yet, but companies appear much more open to the idea. With the pressures of digitization, attitudes are changing. Direct competitors are bundling their expertise. Take the premium auto brands Audi, BMW, and Mercedes, for example, which jointly acquired the mapping service provider Here from Nokia. Extremely precise maps are a prerequisite for autonomous driving, which is being developed in the industry. Here is the world leader for the very data that all three partners need, and which they even sell on to other interested parties. These types of alliances are known as “coopetition,” a portmanteau of cooperation and competition, And we are likely to hear about many more of them in the coming years.

Over the next few years, the emerging digital mega-ecosystems centered on mobility, smart homes, digital finance, and digital health care, for example, will turn many companies that operate in isolation into network partners. The vision paints a picture of a business world that resembles a model of neural pathways, each connected to countless synapses and all connected to each other.