The article In Praise of Dissimilarity from MIT Sloan Management Review has very important implications for R&D management.  The article describes how most managers view similarity based on functionality or product taxonomy (e.g solid state drives and hard drives are similar).  However, another way to look for similarity is based on how different products interact in a scenario or event (e.g. shoes and mp3 players are related through exercise).  This is called thematic similarity.  The article points out that thematic similarity can help focus innovation and provide a competitive advantage.  However, it also raises some important challenges for R&D management.  Lets dig in.
Traditionally similarity (taxonomic similarity) has been seen as a that based on overlap of functions and features:

Whether explicitly or implicitly, the traditional understanding of “similarity” by managers has been a taxonomic one. Simply put, the degree of similarity as traditionally measured depends on the extent to which two objects possess the same features. Personal computers, for instance, all have hard drives, processors and a video monitor.
…
Thus, taxonomic similarity is based on the properties of the objects themselves, and taxonomic categories cohere around shared internal properties. As a consequence, taxonomically related concepts tend to resemble one another.

Thematic similarity is probably as important but often overlooked:

…similarity is not just a matter of degree (how similar are two things), but also of kind (how are two things similar). Two things are thematically similar if they functionally interact in the same scenario or event. For example, an athletic shoe and an MP3 player are related through interacting in a workout theme, coffee and a computer interact in an office theme and a navigation system and a motor via an automobile theme. In each of these cases, the two things perform different roles.

However, managers are trained to focus on taxonomic similarity and hence prone to ignore thematic uncertainty:

When managers ignore the thematic similarity hidden behind taxonomic dissimilarity, they risk overlooking opportunity (as well as misdiagnosing threat).
…
The behavioral theory of the business enterprise has long acknowledged managers’ dangerous tendency to search for opportunity in familiar taxonomic domains.

The benefits of thinking thematically are pretty significant:

Thematic similarity opens up a new area of the dissimilarity space. While Google Maps and Yellow Pages are taxonomically similar services, another Google service, Google Voice Search, and GPS are clearly in taxonomically dissimilar categories. And yet there is a thematic similarity between the two in the context of using cell phones.

Hence themes can actually help focus and direct long-term R&D and innovation:

The new area of thematic similarity holds particular promise for innovation and opportunity search. Focusing on areas of taxonomic dissimilarity can help managers identify novel products or services that result from the combination of strategic assets that are taxonomically dissimilar but thematically related.

As we have discussed many times, innovation occurs at the intersection of technologies.  The more dissimilar the underlying technologies, more disruptive the innovation is likely to be.  Thematic similarity provides a framework to bring normally dissimilar technologies together – and hence drive innovation:

This distant (in taxonomic terms) yet close search for opportunities created by thematic similarity provides a pragmatic guide to how (in which domains) strategists can find new potential for competitive advantage.

The underlying problem is that R&D management processes and cultures are developed around taxonomic similarity:

Taxonomic similarity underlies key frameworks of management such as strategic relatedness, the Standard Industry Classification (SIC) system, the definition of industry boundaries, including the forces within that industry, and the International Patent Classification (IPC). For example, the IPC category F02 (combustion engines) contains internal-combustion piston engines, gas-turbine plants, jet-propulsion plants and so on.

May be we can extend some of the traditional tools such as brainstorming and focus them around themes:

Methods such as brainstorming, which aim at identifying such distant domains, are often referred to in the general management literature. For example, in an attempt to move beyond mere product extension, companies often encourage their developers to think “outside the box”

But true exploitation of thematic similarity will require management innovation. We will need to develop new tools and processes to decide which thematic similarity to explore and how much to invest in it.  One example provided by the article focuses on the integration of GPS technology with cameras. Thinking thematically, this would be pretty straight forward marriage.  However, in reality, this very hard to do.  The skills necessary to design cameras are very different form those required to design GPS receivers.  Even if we can get the two technologists to brainstorm together, actual collaboration though workshops would be rather difficult.  For managers, resource allocation for such development would be even more difficult.  One approach would be to have detailed roadmaps that can be used to engender purpose driven communication between the two groups and portfolio balancing processes that effectively allocate resources for such activity.

…consider an extreme case in which two products are so strongly associated that they are combined in one product but not thematically integrated. Many cell phones sport a camera function and a GPS function. However, the GPS and camera functions have not been integrated in most phones, despite sharing a thematic similarity: Many photos are about places, just as GPS is about places. Thematic integration links these two functions, allowing users to “geotag” the location at which a photo is taken.

Another advantage of exploring thematic uncertainty is the ability to explore all potential competitors. For example, as the article points out Google did not see their business model as amenable to or at risk from social networking:

Google only openly acknowledged the threat posed by Facebook on November 1, 2007, when it launched Open Social, Google’s own social networking platform. In other words, Facebook remained a noncompetitor for Google for more than three years and six months after Facebook’s launch. In fact, Google managers actively dismissed Facebook precisely because it did not fit Google’s taxonomy of activities. Google CEO Eric Schmidt said, “We have address books, and the sum of our address books is the social graph.” And it was not until February 9, 2010, that Google acknowledged the thematic similarity between social networks and e-mail by making a determined foray into exploiting the integration of social networking and e-mail by launching Buzz, a networking service that was closely integrated with its e-mail offering, Gmail.

We will also need new strategic planning processes that can identify competitive threats from thematically similar firms.  More importantly, we will need a better approach to evaluate those threats and find effective ways to respond to those threats.  Finally, thematic similarity can be used to find acquisition targets.  The article points out that Intel believes it acquired McAfee based on thematic similarities.  The problem is that McAfee’sbusiness model is so different from Intel’s that integration of the two will take a very long time.

Actually, Intel and McAfee are remarkably similar thematically. According to Intel, the acquisition of McAfee would boost its strategy in mobile wireless, where it is beginning to produce chips for smart phones. Beyond smart phones, security is becoming a key requirement as new devices, from tablet computers and handsets to televisions and refrigerators, connect to the Internet. The purchase is therefore set to turn Intel, the world’s largest chip-maker, into a leader in security, extending its reach into Internet-connected devices.
…
While experts hope that chips can be improved to make them able to withstand malicious attacks, that prospect is seen as being years away.

Even with time, I am not sure how easy or valuable this integration will be.  May be there is a limit to how much taxonomically dissimilar firms can be before they can no longer be merged effectively.  Furthermore, if integration is going to take many years, can we actually forecast how the market place will function at that time?
A few more questions than answers, but still a very useful concept.