Over the past few months I’ve been reading through chapters of The Innovator’s Solution by Clayton M. Christensen.  It’s written in pretty thick business jargon so it takes a little while to get through, but the content is pure gold.  At this point I think it has officially become my favorite book.  I also believe that it is the best book about Ad Tech that never actually mentions Ad Tech.

One chapter I read today that particularly resonated with me was about the dichotomy between deeply integrated vs. modular products.

Apple Computer’s approach to building computers in the 1980’s is the classic example of building products in a deeply integrated fashion.  The reason Apple was able to push the performance of their products to exceed that of their competitors was because their products were vertically integrated across hardware and software.  By producing all parts of the computer in one company, Apple was able to innovate very quickly and achieve cutting edge, market-winning performance from their computers.

A contrary example is IBM’s modular approach to building computers.  Rather than maintain full control over the hardware and software of their products, IBM outsourced the development of their software to Microsoft and the production of their microprocessors to Intel.  This modular approach allowed each supplier to specialize in a specific niche, optimize their processes and drive costs down.

I’ve read many books that talk about the merits of Apple’s approach versus that of IBM, however in The Innovator’s Solution Christensen favors neither method outright.  Rather he states that it is the market circumstance that dictates the optimal product structure.

Christensen’s point of view breaks down quite cleanly between the following two market circumstances.

1) Product is “not good enough”

If the performance of a product category is “not good enough” to meet the needs of a particular market, then a deeply integrated approach will win.  An integrated company can push performance barriers much faster than a modular company that has to worry about maintaining standardized interfaces with external suppliers.

2) Product is “good enough”

If product performance is “good enough” to satisfy the customers in a market, then a modular approach will win.   After the performance of a particular product meets or exceeds the needs of the market, customers in that market will no longer be willing to pay a premium for increased performance.  At this point, a modular infrastructure using standardized interfaces will allow you to work with external suppliers and minimize your costs while still providing a product that meets customer needs.

Important to note that markets are not static and product categories can move back and forth between “good enough” and “not good enough” states.

Entering the 1990’s, the bandwidth of optical cable fiber used by telecommunications companies was more than good enough to support voice telephony.  Since the product was good enough to satisfy the market, a modular approach won the market.  Corning produced the optical fiber, Siemens made the fiber into cable and a variety of other companies made the multiplexers, amplifiers, etc. needed in the value chain.  However in the mid-1990’s with the growth of the internet, the existing optical fiber telephone network went from being good enough to support most voice applications to being not good enough to support high speed internet access.  Because the market was now in a state where the quality of the product did not satisfy customer needs, a more integrated approach was needed in order to push the limits of product performance.  Corning, who previously had only made the optical fiber for the cable, also started making amplifiers.  By tightly integrating the performance of the fiber with the performance of the amplifier, Corning was able to produce a higher quality product and push the limits of performance.

Aligning your approach to the nature of the market is very important and making the wrong choice can have disastrous consequences.

In 1996 the U.S. government passed legislation intended to allow independent Competitive Local Exchange Carriers (CLECs) to compete with larger dominant phone companies for providing telecommunications services. The law gave the independent CLECs the right to sell phone and internet service directly to residential and business customers, install phone services at the customer’s location and then “plug in” to the switching infrastructure of the incumbent telephone company.

This modular approach to providing phone service would have driven costs down had the telecommunications market stayed in a good enough state.  However, since the internet was driving higher performance needs, this modular structure was a major failure.  Many CLECs attempted to provide new high speed DSL service to customers but found that it was very difficult to provide these new services while having to rely on the infrastructure of the telephone company.  The vertical integration of the telephone company, along with their integrated technology, service network and billing infrastructure provided them a significant advantage over the CLECs.  As a result none of the CLECs could compete with the phone companies and almost all of them went out of business.

Putting down the heavy, hard cover text for the day, I can’t help but appreciate how applicable these business lessons are across different industries.  I used to think that Ad Tech was unique in its depth of issues and complexity.  What I realize now is that I just haven’t spent enough time looking around.

When to Choose an Integrated vs Modular Product Architecture
Tagged on:         
  • Eh.

    Did you ever read R. Coase on transaction costs economics? I think that explains it better.

    The difference, under that paradigm, comes down to the cost of “innovation” inside the company, or outside the company.

    If the cost to communicate, collaborate, and develop products is lower inside a single company, you will have vertical integration. If the cost is lower in the broader market, a vertically integrated company will fail.

    It’s complicated a small amount by the idea of total available investment. In an ideal world, a single company and a broader market would have equal access to resources – thus, the model selected (integrated or not) would depend solely on transaction costs. That isn’t the case; in fact, it’s rarely the case in the real world.

    Access to resources – personal, knowledge, etc – is measured in both time and money. Frequently, the broader market has far greater access to those resources than any single company can amass, except in certain limited situations (e.g. IBM, back in the day).

    Consequently, the cost of collaboration can be greater in the broader market, but the increased access to resources means the net pace of development is significantly faster.

    Example: open source. (Most of the companies funding open source projects/developers could just as easily have that team in-house; but the net rate of development would be lower and/or the costs higher).

    Open source is an even better example when looking at time-to-market. Most open source projects, even if they’re quite good, never really make it “to market.” Those that do typically have companies behind them – companies that can handle issues internally, because those issues must be internal (the cost to working within the broader market is too high).

    A slight re-phrasing of this is to note that significant work can be done in one area in the broader market, but there exists work to take a nearly-ready product and turn it into a market-ready product that may require investment by a single company.

    Consider the example of Android, particularly in its earlier days.

  • Pingback: 6 Innovative, Scary Marketing Trends That Will Put You In the Lead in 2015 | Cardwell Beach()

  • Pingback: Modularity vs. Integration in System Design – Hightech Forum | iExclusives()

  • Pingback: Modularity vs. Integration in System Design – Hightech Forum | APG Editorial()