DisruptiveInnovation.Org - Latest Discussions

Modularity in Business Architectures

Horace Dediu, 10/19/2016


(This is Part 3 of a multi-part series by Horace entitled "Modular Revolution." The first entry is here, the second is here.)

Modularity was introduced in competition theory builds on the notions of value chains and expresses the divisibility of a problem into components that may be separated into distinct businesses which can collaborate to outperform an integrated or non-modular alternative.

Fundamentally, modular business architectures are constructs for collaboration through well-defined interfaces. In contrast to technological modularity and a framing as a technology performance comparison question, modular businesses are compared to the performance relative to competitors and to the performance that customers want or need.

Modularization entails specifying how components interface with each other in standardized ways. Through standardization, modularity takes degrees of freedom away from designers and as a consequence, it forces the company and its product to back away from the technological frontier of what can be done.

Hence, as a general rule, companies who chose to compete with an interdependent system have a competitive advantage when products are chronically underperforming. As long as the customer wants to have more performance than the best products on the market can provide then integrated firms have an advantage.

The advantage of an interdependent architecture comes less from the components, but more in the artistry by which the engineers knit the components together into the larger system. It the the interface, not the components, that yields a unique advantage.

But when the performance of those integrated products exceeds the functionality that customers will pay a premium price for—when the products over-serve or overshoot the ability of buyers to absorb improvements— then the customers may accept better products, but they lose a willingness to pay premium prices for those improved products.

When this happens, alternative solutions will enter the market and compete by being fast or flexible. This new basis of competition demands that they make their products more modular. Interdependence and modularity are not binary notions. They are part of a continuum.

Modularity standardizes the way by which components fit together — physically, mechanically, chemically and electronically. As these transition from being de-facto to de-jure standards, engineers cannot choose not to adhere to them. This causes it to be hard to differentiate products. All competitors have access to the same technologies, as embedded within components, which are sold in the markets. And the way the components fit together also are standardized. 

If we look at the the automotive market, we see how modularity has grown through the past century. In the 1900-1920 period automakers were struggling to overcome core product performance issues. Vehicles which performed well were exceedingly expensive. Ford popularized the automobile because he was able to build a product that was designed for mass manufacturing.

The Ford production system required deep integration of the supply chain. Famously, Ford created a system that required the ownership of steel mills, mines, and forests for raw materials. Only by placing these modules under control of a production system, could the costs be brought low enough to ensure a vehicle was affordable to an average Ford assembly line worker.

However, over the next few decades, GM, Toyota and external forces including regulations, evolved the production system to allow far more modularity. GM competed on faster product introductions and a broader portfolio of models. Toyota eliminated wasteful inventory and re-work and just-in-time supplier delivery.

For the sake of having access to suppliers, engineers agreed on standards. Regulators inserted requirements for standard interfaces. There is a standard interface between the fuel and the engine: octane ratings and purity standards allow all vehicles to use a fixed set of fuel grades. There are standard interfaces between electrical systems. Batteries are all 12 volts and all on-board systems operate on 12V. Tempered glass and other safety regulations favored certified suppliers offering compliance without continuous investment.

Suppliers learned to build increasingly complex sub-systems. Suppliers now supply standardized transmissions (ZF), electronic fuel injection (Bosch), electronics, ABS systems and a myriad of other components.

The process of modularization has meant that “good ideas” are adopted quickly. Suppliers developing technologies are able to market the improved solutions to all manufacturers and as a result it “diffuses” into the market as manufacturers adopt it and insert it into their product pipelines. Graph 3 shows the speed of adoption by manufacturers of various automotive technologies indexed to the same starting date.

And in this graph we also see how the hybrid technology is growing far more slowly than other improvements. The industry has modularized but this particular module seems to resist adoption. What’s more worrisome is that the hybrid is half electric. Is the sluggish growth of acceptance of hybrid going to be solved by an all-electric equivalent or is the electric vehicle (EV) going to suffer from an even slower adoption?

Understanding the hybrid anomaly should help us predict the forces that are acting on adoption for other systems that have fundamental impact on society.

What forces other than performance could we possibly consider.

Collaboration and Conformability

Rather than looking at the problem from the top-down, let’s examine it from the bottom-up. In reality, being “good” isn’t part of the adopter’s or producer’s decision process. The process of producing and purchasing is not judgmental, it is however logical.

Let’s break the problem of adoption into two sub-problems. One from the customer's point of view and the other from the producer’s point of view.

From the customer, we need to ask if what is on offer is easy to accept. Goodness is on offer all around but because buyers with the means to buy are usually fully allocated time to use and time to learn is scarce. In other words, we have to ask whether hybrid technology is “conformable” to existing, manifest behavior. Buyers typically accept complex solutions slowly and only if they have dramatic (systemic, or across-the-board) positive impact on their lives. These complex, system-wide solutions come at typically high costs. Conversely, buyers accept simple solutions quickly when they have minimal negative impact on their lives even if they have marginal positive impact.

For example, education is complex, time consuming and inflexible. It also has systemic impact and transformative power. Customers therefore shop for education rarely and only obtain the minimum amount necessary. Alternatively, television programming is simple, conforms to their schedule and is flexibly delivered. It has mild effects but does not change lives. Customers therefore absorb hours of it every day and would be uncomfortable doing away with it for even a few days. TV conforms while education doesn’t. TV is quick to be adopted, education is irritating.

From the producer, we also need to ask if what is on offer is easy to accept. Does this technology present itself in a way that it is easy to adopt into an existing production or distribution network. In other words, is hybrid technology easy to build and easy to sell? Is it delivered through a value network that enables plug-compatible sourcing? Producers typically accept technology solutions slowly if they have to do the development themselves or if they have to build new supply chains, or if they have to insource problem solving, or if they have to create new distribution networks. Conversely they accept technology quickly if it’s “on-the-shelf” and provided by existing suppliers and sold through existing distributors, delivering similar margins through similar cost structures.

As we desire rapid adoption for hybrid technology, we have to ask if it’s conformable or systemic. Is it delivered through a collaborative network or as an integrative leap? In other words, it easy to buy and easy to sell? Is it essentially modular?

In studying over one hundred technology adoptions, our research has shown that conformable and collaborative solutions expand industries and are quickly adopted while systemic integrative solutions create industries but are slowly adopted. So much so that we have not found any slow modular technologies or any fast integrated ones.

Through this lens we can ask what are the attributes of hybrid technologies which cause it to be slow.

From the consumer’s point of view, hybrid cars are eminently conformable. Far more than electric cars and, at least in the US, more even than diesel cars. There is no change in behavior necessary to use one. It takes no more time to learn, no more space to park, no further to travel to fuel it, no more money to own it and no more anxiety to live with it. The car slots perfectly in the existing way people live their lives. This conformability allowed the hybrid car to be “a module” in the way it fits with the potential buyer’s intentions. The car has natural, well understood interfaces to the buyer’s other considerations. High grades then on conformability.

It’s when we look at the offering from the producer’s point of view that the problem appears. Recall that hybrid technology was under development by Toyota and Honda for decades. The reason is that the addition of an electric motor to the gasoline powertrain is non-trivial and has consequences throughout the vehicle’s design. The transmission has to be re-designed. The starter, the accessories and the control systems have all to be re-designed. Toyota even eventually styled the car in a unique way. As a result, hybrid is anything but plug-and-play for a manufacturer.

The problem for the average automaker is that if they don’t have the technology in-house they cannot source it. If they can’t buy it then they can’t build it. There aren’t any suppliers they can turn to who have this technology on the shelf. Contrast this with the high speed adoption of disc brakes, air conditioning, in-car entertainment, radial tires, electronic ignition. and even diesel technology. Suppliers quickly built capabilities to deliver these components as “modules” to manufacturers and allowed them to be inserted into an existing production and distribution system. Hybrid technology has “systemic impact” within the whole value chain. It is integrative and it is not modular.

Toyota has developed this technology at great expense in time and money and is, understandably, unwilling to license it. In the US, Toyota has approximately 80% market share in hybrid technology.

Any other manufacturer who climbs the hybrid learning curve is similarly protective of their results. This concentration of learning also explains why in Japan hybrid technology has taken off. Toyota and Honda have far greater market share there and are able to saturate the sales channels with it. Even so, after 18 years only 30% of sales are hybrids.

Rapid adoption of a technology requires a modular approach. This is nothing more than stating that it requires collaboration. Collaboration not just in terms of supplier components but also in manufacturing, distribution, sales and support. Collaboration is exhibited in the creation of ecosystems and in the post-sale network effects. Collaboration can reach extreme degrees where the more you look the more you see. It recurses and fragments problems into atomic parts.

The absence of modularity does not mean that adoption fails to happen, just that it does not happen quickly. Modularity is a qualification of speed.

We also take the point of view that modularity spans into two domains: the supplier and the consumer. As supplier collaboration couples with customer conformability, modularity exhibits the essential push and pull which drives adoption. 

By observing the causal forces acting on a technology we can be more precise about how quickly it will be adopted and also identify the impediments to that adoption. If the problem to be solved is important enough, energy should be applied to overcome these impediments and not wasted on ineffectual incentives.