Strategy Frameworks and Decentralization
The strategy implications of decentralized Systems of Control
Today’s strategy frameworks are highly valuable tools. They allow us to ‘step back’ and look at our businesses and our industries with a different perspective. Used properly they force us out of our comfort zones, and make us address opportunities and challenges that we don’t think about in our day to day operations. Understanding where existing companies are in their strategies helps startups position themselves around new products or new value propositions.
In a previous post we introduced the concept of analyzing opportunities by looking at the centralized-decentralized spectrum, something we called Systems of Control. We can also use this as a lens on top of existing strategy frameworks that may, again, allow for a different perspective — a different conversation. Strategy frameworks are never exact; they apply well in some cases, and less well in others. That is to be expected. Their main goal is to get us to think differently, and Systems of Control is intended to do the same.
In its simplest form Systems of Control involves making decisions along a single axis, from fully decentralized, ◯, to fully centralized, △. Systems of Control involve authority; in decentralized systems individuals have full authority — in centralized systems they have given up some authority to other groups of individuals in exchange for gains in safety and efficiency. Governments, industries, corporations, and business models can be mapped across the spectrum. For businesses this relates to how much control and transparency you give to users compared to your competitors. This is often a multi-faceted question across multiple variables.
Today’s world is often highly centralized, the result of specialization and economies of scale.
By mapping an industry on the spectrum, and then choosing a specific value proposition, businesses can stake out a spot that leads to a sustainable differentiated position, and also better consider where threats to their business may come from.
For example, we may plot personal transportation systems like this:
Owning your own car means you can make all of your own decisions. What type of car to own, how it is driven, where it is parked, how it is paid for (lease, buy), where it goes. Having a timeshare in a self driving car (of the future) is also pretty liberating — you ask for a car, and you get it — anytime, anywhere. Current ride sharing applications like Uber and Lyft give the rider great flexibility, and also give drivers more freedom than taxi drivers…but they are still run and controlled by highly centralized companies. Juno is an example of a ride sharing company that gives drivers ownership in the company; it is a more decentralized approach as drivers now have more say in the operations of the company. The fact that ride sharing companies are on the same side of the spectrum as owning your own car, and may be financially superior, is leading the shift in attitudes that millennials have — many may never need to own a car. At the other end of the spectrum, when you travel by air you must transit through highly centralized hubs, stick to a schedule controlled by the airline, go through security, pay extra for legroom, etc. Other than finding the best priced flight you can, you have little control. Likewise buses follow predetermined routes — you morph your path to match the bus, not the other way around.
Intuitively anyone entering personal transport is going to map themselves across this spectrum. They may, in fact, be more precise and map out all the different elements/actors that go into a final score. Perhaps two or three variables are required to do this effectively: for drivers, for riders, and for the company behind them, and in the case of personal transport, the overhead structures that make the system function (lights, intersections, overpasses, flight paths, etc.).
When looking at this spectrum there are attributes that generally apply to the extreme ends of the scale. These are not universal; there are many interesting and thought provoking counter examples. The easiest way to think of these attributes is to consider human society overall, and our progression to the right: from fully decentralized hunter gatherer groups, to highly centralized countries…even authoritarian regimes.
(Aside: Vitalic Butterin, of Ethereum, has noted that language is a good example of a decentralized system that is very efficient. There is no central authority for English; it arises from the choices we all make and can be quite different from locale to locale. This is not counter to our thesis — we don’t claim all systems centralize; we simply claim that many of them do, and those that do largely have the attributes shown below.)
We have presented the thesis that we are at a point in our technology maturity where technologies are starting to compensate for some of the drawbacks of decentralized approaches, and that given the choice users will migrate towards more personal authority; towards more decentralization. The experience of Lyft or Uber is better than the experience with a taxi.
Centralized systems are also more specialized, and more amenable to automation and AI improvements — the core central functions of existing businesses will be driven to low cost and low value. Therefore we would expect more challengers to be entering the market ‘to the left’ of existing participants.
Thesis: We will start to see movement from △ towards ◯ as decentralized systems become more efficient, as more specialized functions are automated, and as users choose to have more control and sovereignty.
With that as our thesis, how can we apply some of today’s existing strategy frameworks, overlaid with this System of Control, to gain insight into the future?
Innovation
It is generally accepted that innovation starts on the decentralized end of the spectrum, and progresses towards the centralized side.
This is well documented by Timothy Wu in The Master Switch and Steven Johnson in Where Good Ideas Come From. On the decentralized side you start with lots of open discussions and ideas, shared amongst many people allowing for serendipity and cross functional mixing. As the ideas mature, more components get locked down and maintained as proprietary, corporate entities take control, and a business model is applied. In the end, most ideas end up within existing corporate structures and models, weighted heavily towards centralized control. The limiting factor on further centralization is that at some point the cost of internal transactions exceeds that of external transactions, simply due to the overheads associated with very large entities. Between that and antitrust enforcement there is a limit to how centralized things become.
Many large companies facilitate this left to right process by establishing research labs that operate outside the main control functions of the larger entity. These labs are intended to generate ideas in a more open and less constrained environment, and then figure out how those ideas can be deployed within the existing corporate framework. This has had mixed success to date.
In the future, we expect to see this same left to right progression of new ideas, but the end point — where centralization stalls out — may be further to the left. Further, there is the opportunity to move not only product, but also process and philosophy from more decentralized approaches into production environments. This will not be due to less efficient internal systems, but rather because a new more decentralized solution will actually be more efficient — many of those cumbersome internal systems will simply be replaced with zero marginal cost software and robotics. You should assume that a new entrant will do this; you don’t want to be caught unaware. Thus it is useful to think through how integrated into your existing processes, channels, and business models new ideas and products should be driven. It is possible that trying to maximize “synergies” with existing businesses may actually miss the mark. You should map out your most centralized processes and models, and challenge them. These are typically systems that have been highly tuned and are very efficient — specifically because they allow less freedom in exchange for that efficiency.
The obvious examples of more decentralized systems invading existing business are open source software, and more recently blockchains and ICO’s.
Gartner
The Gartner hype cycle can be trivially overlaid with Systems of Control, as the time evolution through the cycle typically matches the move from decentralized towards centralized.
Often the hype is generated because the new technology promises a progressive value proposition as well as a new technology. It breaks the old models and gives more people more control or more freedom — this is part of the seduction. Then everyone starts to realize that things are not quite as great as they seem; things are more complex, need to be controlled more, require more integration, are better with a simpler business model. These effectively centralize the offering, making it more understandable and easier to fit into existing models.
However, viewed under our increased decentralization thesis, we can imagine this curve changing, from the older grey curve to the newer purple one.
While the entire hype cycle may actually move faster in time, due to more automation and AI, the more important observation is that technologies reach general acceptability at an early point in the centralization lifetime. They target a point in the System of Control where they provide more control, transparency, and authority to their users. However, this is uncomfortable for existing business processes and industries, and can lead to even greater conflict in acceptance of the new model. Ignoring this is perilous — because the more decentralized solution has a different model, disruption is quite possible, and the previously expected plateau may not last as long as expected; you can’t milk the old system for as long as you thought. Microsoft Windows is a good example — even five years ago most people expected Windows to last a long time, but with most applications moving to the Cloud and with alternatives like smartphones, tablets and Chromebooks, the plateau for Windows now looks quite limited.
SaaS is a good example which has occurred over the last ten to fifteen years. Since SaaS offerings have less “centralized integration” into your business process, their adoption can be faster than traditional software solutions. The services update automatically, reducing maintenance cycles, and often gives you more control — SaaS offerings have lots of dials, built for multiple users, that can be given to all subscribers. This has replaced the older model of configuring a product for a client and then installing and locking it down on premise. The first SaaS services to go through the hype cycle took significant time to get traction as the new model had to be accepted by the market. However, as subsequent services make the transition, they move faster, and generally compete on even more ‘openness’ than previous ones: API’s, data exports, connectors, etc. The newer competitors make it easier to control and move your own data, making the lock-in lighter. Adobe, as an example, made the transition from on-premise software to SaaS. This involved a serious change in business models which short term capitalism forces impede.
A, somewhat more speculative, future looking example would be running shoes. Buying runners today is still a ‘pick something close to what I want that almost fits from a store that has inventory’ value proposition. Decisions about styles, sizes, and arch support are made at central headquarters; stores may or may not have the inventory you need. You get to make the choice between a finite, predetermined set of shoes that the designer dreamed up, even if you are buying online. You have some, but not a lot, of control. The business model for shoes is simple as well — you pay the price that is advertised. New shoe offerings are coming that give you more choice; not only ensuring you get the exact right fit, but also choosing colors and styles that match your lifestyle. You may soon even buy a shoe that is custom produced in the store, as opposed to at a factory. The new business model may be menu driven: You choose the type of sole, the quality of fabrics, and then, ala carte, you add features and functions that you want. Add it all up, and that is the price you pay. In the short term you may pay more; it is a unique and high value service. However, over time you may pay less. Centralized overheads get reduced. Stores do not need to carry unnecessary inventory, pay for stockroom space, or manage inventory and shipping. Instead, raw materials are delivered to the in-store production machines, and there is very little waste.
Ultimately the user has more choice and control; more authority over the shoe they end up with. Old fashioned pre-built shoes will compete on price for a while, but ultimately custom built runners will replace them. Custom made shoes are working their way through the hype cycle today; subsequent footwear offerings will transition much more quickly.
As new products transition the hype cycle they will mature at a more decentralized value point. This is an important consideration for companies watching technologies progress through the curve — don’t expect them to compete with you head on; they will have changed the value proposition.
[ Aside: While far from a full analysis, we can use this view to speculate about the future of work.
As automation and newer products enter the market, the ‘plateau of productivity’ will lose jobs. The more centralized, repeatable and specialized a job is, the more it is subject to automation or redundancy. On the other hand, as new products enter the market in a more decentralized mode, more jobs are created at the leading edge. These are the jobs that we don’t yet fully understand — the Airbnb hosts, the ride sharing drivers, etc. They are created to support the new form of product that exists ‘to the left’ of the existing products. ]
Crossing the Chasm
Related to the hype cycle is an analysis of Crossing the Chasm, where we can again overlay the System of Control view.
The grey line represents the original curve; the purple line is what happens when a newer more decentralized entry hits the market. The phases in the normal cycle get shifted as business models move on the System of Control axis. These strategic models, which we think of as static, can actually change over time.
Here we show that the chasm occurs earlier in the product lifecycle. If the new value proposition is more friendly to users, because it gives them more authority and control, the uptake will be faster than a new product that does not change the value proposition. Products that hit this correctly grow very quickly (Uber, Snapchat), and differentiate themselves through fast user growth. They hit the chasm earlier, and cross it earlier than copycats of existing solutions.
It is possible that more decentralized models also have more longevity. They will have automated more core backend systems and they can run profitable businesses for longer while still maintaining a great value proposition. The laggards will join earlier and stay for longer.
This is the impact of ‘moving to the left’. Historically this was not possible because scaling and managing a significantly more decentralized system was too complex, and the overheads outweighed the advantages. This is starting to change as software is used to manage the complexity, and as centralized functions become cheaper and cheaper to support and run. Further decentralized trust models, often built on blockchains, mean companies are more willing to give more authority to end users.
Strategic Horizons
McKinsey’s Strategic Horizons is a simple way to think about the time and resources that a mature business puts into different aspects of its business.
- Horizon 1 projects are core businesses; these are paying the bills and typically recommended to account for 70% of a firm’s investment.
- Horizon 2 initiatives are emerging opportunities that should extend and leverage Horizon 1 projects, refreshing them and giving them extra time and/or value. This is traditionally 20% of a firm’s investment.
- Horizon 3 explorations look at disruptive new ideas that can fundamentally change the company.
Horizon 1 projects tend to be highly optimized and run for bottom line results. Horizon 2 may be inefficient but hold great promise for the top line. You need more flexibility for Horizon 2 project management to allow them to find their optimal fit before moving them into Horizon 1. People working in Horizon 3 need lots of freedom; then can’t be locked down into existing processes, procedures or channels.
Similar to the chasm model, the grey lines represent where the switch in Horizons are today, and the grey dashed line shows the typical linear progression from left to right of ideas to market.
Traditionally Horizon 2 projects are slotted into existing Horizon 1 infrastructure — the same business model, the same channels, the same operational infrastructure, the same philosophy. But, if competitors are going to disrupt that model, Horizon 2 projects should have the ability to update or change traditional Horizon 1 operations and philosophy. That may mean moving some Horizon 1 infrastructure to a new Horizon 2 approach as opposed to the reverse. Instead of the flow of products and ideas being strictly left to right in the above, there may be times when products, people, and resources, should move from right to left. This is shown by the purple dashed lines above.
Likewise Horizon 3 projects often struggle to get into Horizon 2, let alone Horizon 1. They just don’t fit into the current models, and therefore get dropped. This can be because of the differences in the specialized and optimized central processes as compared to the less optimized and less understood decentralized approach of Horizon 3. By rethinking this, and considering what it would be like to move the Horizon 3 ‘environment’ into production, as opposed to the Horizon 3 ‘products’, you may reconsider the traditional Horizon pipeline.
Changing roles
This change in Horizon perspective can inform the type of people that manage different aspects of the business. Geoffrey Moore outlines “the four zones”, which detail how leadership, measurement, and operations are different as ideas progress in their lifecycle. Once more on our Systems of Control axis, this can be seen as the left to right transition through the zones.
If you attempt to manage all four zones with the same style, the same KPI’s, and the same type of leadership, you are doomed to failure. Highly entrepreneurial people will fail if they are locked in tightly controlled environments where every process must be followed exactly. Operational types, even if put in incubation labs, may never succeed as they do not know how to operate in a fully open autonomous environment. The highly valuable Transformation leaders who can straddle between the open decentralized innovation world and the highly structured and process driven Performance and Productivity worlds are going to be in higher and higher demand.
The shift towards more decentralized systems shifts your needs to the left: organizationally and investment wise. Many large corporations today have “Labs”, that they often put in Silicon Valley. They may not have a transformation zone at all, simply hoping that products will emerge from their labs and end up in their performance business units.
In the future, however, they will have to put even more investment and more great people into the earlier phases. They will have to figure out how to motivate and compensate intrapreneurs, and also find individuals who know how to transform early innovations into scalable businesses. If you were to map most companies today, the majority of people would be in the Performance and Productivity zones. In the future, the majority of your people will be in Incubation and Transformation. This has significant implications on the entire organization, and many products and services will be built to help companies make that change.
Of course, part of this transition is because more and more Performance and Productivity jobs will be automated — they are the specialized rinse-and-repeat jobs that are most amenable to software and AI. But the other component is the change in value proposition, pushing more control and authority towards end users. This requires more sophisticated Incubation and Transformation systems to manage the engagement, feedback cycles, and rate of change that is implied.
Of course there are many other strategy frameworks that are used. Mapping the System of Control thinking over top of them can be enlightening.
Summary
We believe future businesses will have more decentralized attributes than today’s. This is a function of increased software and AI impact, combined with the trust models behind blockchains, and the increase in efficient edge manufacturing and personalization. These help to mitigate the major detractors of decentralized systems: efficiency, trust, and complexity.
Strategy frameworks are a valuable tool to give you new perspectives on your business or to position your startup. However, you should not think of these as static models — there will be a ‘shift towards decentralization’ that will put these models in a new light. By looking through the lens of Systems of Controls you may ask different questions, or shoot for different results, than otherwise.
As startups are looking to disrupt industries, they are often choosing a new location on the System of Control axis. Moving a person from point A to point B is not a new idea; doing it with a significantly different value proposition (real time, simple payment, full visibility, feedback mechanisms) gives users more control and changes the model significantly — thus the new ride sharing economy. Using compute power to solve interesting problems has been around for decades; putting that processing power directly in people’s hands has changed all the models. Many ideas start out by giving users more control and transparency, and then gradually become more centralized. This will not change….however, the point at which they reach equilibrium may well become more decentralized. This will have dramatic impacts on many industries.
