Increasing use of digital applications and economy-wide digital processes challenges our business management. Changes in society, unseen innovation speed, and increased reconfiguration of value networks confuse many industries that enter the stage late. Missing experience with digital technologies, new business models, and the complex influences of a highly networked, global business world are significant barriers to success. This is as valid for management as for the employee level, but the employee level suffers additionally higher fears of job security.
The THERON method (Predict, Involve, Embed) provides a basis to navigate this highly complex digital world by optimal inclusion of all stakeholders in a business.
Please take a moment to read the following abstract from a Wall Street Journal article of August 8, 2017!
Uber Technologies plans to wind down its U.S. subprime car-leasing division to stem unsustainably high losses, according to people familiar with the matter, a major retreat just two years after starting the business. The ride-hailing company is aiming to close out or sell most of the business by year-end, the people said. As many as 500 jobs could be affected by the end of the Xchange Leasing program, or roughly 3% of Uber staff. Executives were prompted to pull the plug partly because they came to a stunning realization: The average loss per vehicle was about 18 times what they had thought.
(Wall Street Journal, August 2017)
What happened? Uber – the most agile of all companies, the master of the digital universe. How could this sharing economy juggernaut fail so miserably? How can we avoid such a disaster? We need to understand the nature of digitalization in order to answer this question, and to learn from it.
The digital economy is a vicious circle composed of technology, business and capital (Exhibit 1). The circle is vicious because more and more of its components will become digitalized themselves, thereby increasing the speed of innovation, nearer to the speed of pure digital players.
Compared to the past, digitalization raises the level of uncertainty in business drastically. The outcome of a given input is very hard to predict. This is caused mainly by the following factors:
This last one is a special case. Digitalization makes things more complicated than they used to be. Finally, everything and everybody will be digitally connected to everybody and everything at marginal cost, but the path towards it is not that clear (see Exhibit 2). Any change will propagate, cascade and potentially amplify many other changes throughout this tightly knit, multi-dimensional grid.
We all know respective stories from the financial services world: computer speed trading hiccups, hacks, cyberattacks, crypto-currency roller-coasters, and uncontrolled credit bubbles. Similar happenings sooner or later will become regular events in most other realms of our second life.
Uncertainty in a system of complexity acts a bit like a mirror maze. What is the original image, what is a first level reflection, what a reflection’s reflection? At every turn and corner, one must quickly learn the reality. This takes a couple of trials – a move to the right, a look backwards, what does the image do? Once you feel safe, you can move forward with a good chance of not bumping into a mirror.
If an organization is not able to learn in such a setting, failure will rule, pile up, and suck in all energy. Over time this leads to cultural and financial distress. Fear will rule.
What about our leaders? Will they lose control? Probably yes! No single person will be able to fully oversee the web of digital competition. The “know-it-alls” of better days are clueless. As the Uber case proves, even the young hot shots easily miss the target by a multiple of almost twenty.
Really frightening, isn’t it? True cluelessness is what frightens us so much about digital change. People feel helpless, overwhelmed, exposed to unknown unknowns, passengers on a journey into the unknown.
Social scientists rightfully compare digital literacy to the three previous disruptive forces of human society: learning how to speak, learning how to write, and learning how to print. All three consecutively have led to radically different models of society.
What will the digital society look like? Will we still have sustained organizations offering jobs, a social home, and hope? What will be the glue of such systems? What will make them strong in intensely competitive struggles and in quite hostile environments?
High-hanging fruits, slow fixes, and hard wins
Let’s face it: There is no easy way out! Digitalization will not go away. Complexity must not be ignored. And competition is fierce by its very nature.
What about our leaders? Will they lose control? Probably yes! .
Trial and error, the prayer of all prophets of agility, might indeed be an effective learning approach, but it is not a very efficient one. Every trial consumes resources and chances in the form of time to market. An almost arbitrary approach pays off only for so-called “Hyperscale” companies, i.e. digital juggernauts with a huge global web of relationships of different kinds and, at the same time, minimal marginal cost of change due to their virtually virtual business system. Any innovation they launch of whatever shape or form will find enough customers at the long tail of their customer base to pay for their investment (see Chris Anderson „The Long Tail“, New York, Hyperion, 2006; Yochai Benkler, “The wealth of networks. How social production transforms market and freedom”. Yale University Press, 2006). But most existing businesses and companies do not start out as “Hyperscalers”; contrary to Facebook or Google, digitalization is not in their DNA (Exhibit 3a).
Most companies will not even make it there at all, and that on purpose! It is simply not in line with their core business, their identity and their values. They are located near the “Old School” corner of our Digital Genotype matrix. The number of customer relationships is relatively low compared to e.g. the sum of all internet users available to Google on a global scale. And their business systems generate a significant share of fixed cost, which inherently generate relatively high transactional costs for change. Even extensive use of robots and automation will not change this in the short or mid-term. To these “Old Schoolers”, errors will cost not just money but precious time as well.
Companies near the “Hard Web” corner deal with a significant product cost share. Thus, their relatively high cost of change will be multiplied by their high number of customers. Failure in the hardware part of their business – i.e. the next iPhone, the autonomous car, or the fully automated supermarket – can make or break their business. Hence, they have to develop effective means to identify potential failures as soon as possible and prevent stranded assets as well as other negative impacts on a larger scale. And even “Soft Spot” businesses must try to avoid too many failures. Due to the fact that they are less connected, failures cost them less money, but they pay with two other valuable currencies: time and organizational energy. Especially in markets determined by critical mass or controlled by network effects, time to market is the hardest currency. And in uncertain times, organizational energy is the glue which holds it all together. Both of these types of opportunity costs are usually not accounted for in P&L statements, budgets, or investment plans. The financial bias of our traditional controlling instruments often blinds us from the real cost of erring – i.e. losing opportunities en masse and wasting energy on corrections and justifications.
So, to all businesses other than “Hyperscale”, trial and success is a vastly more economic approach in a digital world (Exhibit 3b). They must learn how to select the better alternatives for their exploratory activities, and recognize patterns which they’ve seen before in order to draw conclusions prior to the next trial. In essence, they must learn how to learn fast. This is what this paper is all about – double loop learning.
Thousands of books, articles, and papers have been published on the issue of learning organizations over the past 50 years. Starting out with a rather technical focus in the early days of organizational learning research, the focus has quickly shifted to cultural issues, behavioral theory, and finally post-heroic leadership philosophy. This work has been an important contribution to social science and organizational theory. And we have come a long way in putting these valuable insights into practice. But: does it give us concrete advice what to do now? Will these insights gained from organizational theory be capable of guiding us through the myriads of daily decisions when digitalizing our businesses – the formal as well as the informal ones, the explicit as well as the tacit ones, and the conscious as well as the unintentional ones - without paralyzing our organization by fear? Most probably, it will not! This is due to the fact that it would require constant awareness in times of rapid change, high complexity and utmost uncertainty. No organization will be able to meet this requirement. This challenge is very much like nailing jelly to the wall. A nail alone won’t do the trick. We rather recommend transferring the jelly into a case and attach a loop to it.
Our approach is called “PIE – Predict Involve Embed!” PIE replaces utmost and constant awareness by practiced and formalized behavior patterns.
The overall idea is based on findings of how the human mind works. We took the most critical insights produced by extraordinary thinkers like Daniel Kahnemann, Amos Tversky, or Philip Tetlock, and merged these findings with empirical lessons learnt from our own consulting practice and from applied concepts of innovative companies like Amazon or Twitch. And here is what we have come up with:
The core concept of PIE is a systematic and educated prediction of the outcome of business decisions. Any prediction is expressed in numerical terms and qualified with a confidence level. The method is totally standardized and obligatory, and thus provides for tangible benefits:
It is almost common sense today: In times of growing complexity there is no single function in charge of the outcome of an investment anymore. Hence, collaboration in teams must replace cooperation in hierarchical command-and-control structures. Synchronous dialog must replace asynchronous messaging. All parties involved work together on an issue right from the start. There is no leader and no follower. Effective teams almost always must span function borders, organizational units, and legal entities. Hierarchical reporting must be subordinated to lateral communication. And most importantly, the teams must pursue a common objective and report on identical performance indicators.
PIE is able to trigger this change and reinforce consistent lateral team behavior on three levels:
PIE replaces notoriously faulty “benefits collection” at the end of a project with a continuous and holistic alignment of tasks, expected outcome and financial planning. PIE redistributes risk and contingency planning – performed today by top management – into the working teams, and embeds the management of uncertainty, opportunities, and risks into the daily practice of all relevant stakeholders.
Most companies do business cases all the time. So, what is the difference to PIE? PIE differs from traditional business cases in four aspects (Exhibit 6).
PIE does not focus solely on the bottom line. PIE incorporates all elements holistically, from time to market, critical marketing milestones, technical achievements, etc. All success factors are systematically predicted with equal attention.
PIE eliminates hidden biases of “best/worst case” agglomerations by untangling important variables. For example, when looking at the bottom line, costs and revenue will be dealt with one by one, and relevant interdependencies like cost-revenue ratios will be looked at separately. This helps to better understand the underlying cost and revenue drivers as well as any relations.
PIE relies on explicit and tacit business knowledge only. The PIE method cancels out political agendas, power imbalances, skewed data sets and other factors which typically make business cases useless and often even damaging.
PIE establishes effective responsibility and accountability beyond the hierarchical level across the entire team. Opposite to the common practice of collecting and documenting second and third opinions, there is just one common goal for all!
The PIE concept makes organizational learning the focus of leadership in “Old School”, “Soft Spot” and “Hard Web” business situations. Such situations can be found in virtually any company.
Because PIE is rather simple, it can be implemented much easier and quicker than lengthy and lingering transformation programs, provided that the management really supports it. Like any other transformation program, PIE is not a quick fix. But it will deliver valuable feedback and create a positive spirit from day one.
In order to estimate the value of PIE we must fully understand what the term learning means from an ordinary employee’s perspective: learning is difficult, painful, anxiety provoking and time consuming. It costs a lot of the individual’s energy, attention, and emotional stability.
Digitalization makes things even worse in that the anxiety factor is multiplied. This is due to the following three reasons:
The fatal product of these anxiety drivers can hardly be overestimated: Learning will grind to a halt if the leadership isn’t prepared to create and maintain an appropriate cultural environment.
PIE is neither a self-started nor a sure hit, because its focus is on the learning process. Above all, we want to learn how to prepare predictions that are as accurate as possible. However, employees are usually not used to making numerical predictions. At the beginning, team members will feel very uncomfortable when asked to do so. They will cast doubts on the method and on their own skills, and they will wonder about personal consequences of this exercise. However, PIE can become the centerpiece of a learning culture, because it addresses all three anxiety drivers simultaneously (Figure 6).
Learning is most effective when it is based on frequent and clear feedback. Therefore, we must keep track of what we had predicted, we must identify deviations, and we must explain the reasons. Forecasters need feedback from two different sources (Exhibit 7).
These two feedback mechanisms must be formally implemented in order to work efficiently. Typically, a central function in charge of planning should serve as dedicated PIE Support Unit. Beyond methodological guidance this unit should also monitor from a neutral point of view team setup and stakeholder involvement. Team scope, composition, and approach are critical to success. Last but not least, the PIE Support Unit must provide for moderation of adverse reactions, potential conflicts, and dysfunctional behavior – a delicate and very important task which cannot be achieved without the adequate attitude of managers and leaders.
Last but not least, there is a supercritical informal success factor. We strongly promote a leadership model called “Post-heroic Leadership”. This model entails a couple of often tricky and painful demands on top managers.
Invite speaking up, listen to and learn from subordinates!
Utilize it instead of ignoring it! Take the burden of dealing with complexity in a smart way, i.e. discerning important from irrelevant facts based on an analysis of today’s situation rather than falling back upon undoubted truth from years of experience. Solutions which might have been right yesterday may be totally wrong today.
Accept concerns and opposed opinions as valuable input rather than considering them as a sign of disloyalty and ignorance.
Understand failure as a fruitful ground for effective learning. Don’t punish the ones who err, rather focus on the ones unwilling to go back to school!