Sunday, April 30, 2017

The Partnership Fallacy

Remember doing group projects in school? Typically your teacher would partner you up with students who had different skills. Maybe a good writer would be matched with someone who knew math, for instance. In theory, you would each teach the other; the group project would get done, and everyone would learn. But what really happened? Perhaps pushed for time, at some point everyone in the group realized that the job needed to get done. And if you wanted to do well, this meant dividing labor. You broke down the project into parts, as in the well-known "divide and conquer" algorithm pictured here, and each of you took on the part that you understood best. In the end, the project was a success – but the math person did even less writing, and the writer again avoided math. So goes the partnership fallacy:  We partner up hoping to improve our weaknesses, only to divide labor and so make our weaknesses even worse.


Business leaders often fall victim to the partnership fallacy. For example, it may be hard to believe today, but in the early 1990s Apple Computer (as it was then called) was widely known to be bad at making small devices. The company’s earliest attempts to make small computers had flopped. Meanwhile, Sony stood out as the worldwide leader in making cool, miniaturized electronics. So Apple’s leaders decided to form a partnership with Sony. The idea was for Apple to learn about miniaturization while producing a laptop computer. History says that the alliance succeeded, because it created the first “Powerbook” computers. But in terms of learning, the alliance was a failure. Case studies at the time reported that deadlines kicked in, and pushed the two companies to each stay focused on what they did best. In particular, Sony took care of miniaturization, with very little day-to-day contact between Sony and Apple engineers. In the end, they got the job done – but Apple emerged without learning Sony’s miniaturization magic.



What’s more, often partnerships fail even to accomplish their stated goal, never mind learning. If you have much experience, you can probably name an example or two of failed partnerships. Business leaders typically allow these arrangements to fizzle out without much fanfare. Combining firms through merger and acquisition suffers this problem, too. We talk a good story about “synergy” and “learning” and “1+1>2”, but the evidence shows that combining firms typically increases only the variance in performance (not the level of performance as promised by “synergies”).  

The lesson: Partnerships are not a substitute for learning by doing. Partnerships sound great during a planning session. After all, who can argue with the idea of bringing in somebody who already knows what you need to learn? But ask yourself, how did they come to know?  Odds are, they learned the old-fashioned way: by doing. Effective leaders understand how organizations learn, so they avoid the partnership fallacy. Partner if you must, but don’t fool yourself that this is a great way for your organization to learn.


For a thorough academic study highlighting the value of organic learning and growth, read my book on competition.

Saturday, April 15, 2017

Differing without Dividing

Variety is great for innovation. For instance, consider the case of Seymour Cray, the “father of the supercomputer.” In the 1970s, Cray left Control Data to start Cray Research, a company devoted to creating the world’s fastest computer. Cray approached the problem with a revolutionary architecture, so called “vector processing.” By 1976 he and his team introduced the Cray 1, and Cray Research was seen as the Mecca of high-speed computing. John Rollwagen became company President in 1977, bringing business leadership alongside Cray’s technological prowess.


In 1979, Rollwagen brought in another technology genius, Steve Chen, to lead the design of a completely different approach to supercomputing. So as Seymour Cray’s team worked on the Cray 2, Chen’s team worked on the Cray X-MP. Chen’s design built on Cray’s initial innovation, but did so using a revolutionary architecture featuring multiple processors operating in parallel. Released in 1982, the X-MP set a new standard for supercomputer performance, and significantly raised the bar for the team working on the Cray 2.


When we do not know what the future holds, variety helps our organization to discover what is possible. This truth is one reason why we often hear people saying that they want to increase the diversity of their employees. Just like the biosphere, organizations evolve better if they sustain variety.

Yet examples like Cray and Chen’s are rare. One reason is that sustaining variety is expensive. How inefficient to run multiple projects that are trying to do the same thing. But another, bigger problem is that sustaining variety threatens to divide a company. People object to having others in their company working at cross purposes. How can we encourage differences without being divisive?

One way is to live by the adage “disagree and commit.” Here in Silicon Valley people attribute the saying to Intel. The idea is that you should encourage disagreement during the decision-making process, in order to improve the quality of your decisions. But once a decision is made, everybody needs to fully commit to its implementation. Unfortunately, in practice this saying often is used to silence those who see things differently. Often managers say “disagree and commit,” but they are really saying “disagree and shut up.”


I prefer “switch and commit.” The goal is still to end up committing at the end of the process, but during the decision I want the participants to switch roles. The person disagreeing with you needs to take your position and argue it well. Similarly, you must argue the other’s view well. You can think of the approach as devil’s advocacy taken seriously by both sides.

I first tried “switch and commit” when teaching a controversial topic here at Stanford. For the first assignment, the students had to state their position on the topic. For the second, big assignment, they had to write an essay taking the opposite view. (They did not hear about the second assignment until after they handed in the first.) The end results were some fantastic essays, because the authors were legitimately skeptical.

Since then, I have tried “switch and commit” when facilitating hard-hitting business meetings among top managers. The results have been mixed. Many people cannot get their head around a different perspective. But now and then you find an exceptional leader who appreciates the value of differing without dividing.


A readable review of related academic work is Scott Page’s book The Difference.