It can be positively dangerous to be too much of a subject matter expert. The following comes from a body of work originally developed by the CIA as guidance for analysts.
Consider the following simple experiment. Read the text in the three triangles below.
Did you notice that in each case the article is repeated? If so, you have exceptional powers of observation, were lucky, have seen a similar exercise before, or are a professional proof-reader. Mostly, we see what we expect to see. In fact, the more fluent you are in english, the less likely you are to see the repetitions – the more likely you will unconsciously “correct” the “faulty” text.
As a more relevant example, consider the case of the two Germanies at the end of the 1980s. The Soviet Union was collapsing and had announced that it would not interfere in the affairs of its satellites, and unrest was growing in East Germany. The Berlin Wall was about to fall. German reunification was becoming more and more probable. Yet is was often the German specialists in the foreign ministries and intelligence services who had the most difficulty seeing this. They needed considerable prodding from their more generalist supervisors to accept the significance of current events.
This by the way is an alternative angle on the problem of disruptive innovation described by Christensen in his book The Innovator’s Dilemma. Many established businesses suffer because they miss or resist new products. Christensen ascribes this to either financial or organisational factors – the new products are less profitable than the old, or the most powerful people in the organisation are those in charge of the established businesses; they have no desire to cede power or influence to those bringing forward the new. The German example suggests another reason – they simply cannot see what is happening.
In a world which is changing fundamentally, deep domain knowledge on its own can be dangerous. It predisposes you to see what you have always seen, and blinds you to the new. This is not to deprecate it altogether, but to say that it is probably necessary but certainly not sufficient. It needs to be complemented by someone who has skills in complex and ambiguous situations in general, not knowledge of your situation in particular.