There's an article in Fast Company about Google's design lead, Doug Bowman, leaving the company to take a job at Twitter. He says that Google's culture of over-reliance on testing and data paralyzes the design function.
This is yet another sad ending to an all too common issue. Businesses that have to place big bets are typically reluctant to do anything that has not been evaluated to mitigate unnecessary risks. Exercising this type of caution is good, and I am an advocate of evaluating ideas before they are implemented.
Where it goes wrong is when the evaluation methods used are inconsistent with the types of decisions that need to be made. While I don't know the inside details, I would suspect that this is largely the issue for Bowman, as it is for most designers who work with large corporations. In most business disciplines, very linear methods of evaluation can be used to make decisions such as the return on investment of new capital equipment, the potential market effects of price changes, or the usability of web products.
However, it is a designer's job to create new experiences, manage consumer's perceptions, and ensure that products and services are meeting underlying consumer needs. Design elements are used holistically to acheive these goals, and testing individual attributes (such as Bowman's mention that he had to test a 3,4, and 5 pixel border) can lead to inacurate design decisions. Does that mean that design cannot be evaluated? Not at all. In fact, that's where many designers fall down.
Design can be evaluated, but new methods have to be used that evaluate whether the holistic design elicits the desired response. The key is to focus on the results, not on the individual elements. I have had great success with this type of evaluation. The companies that struggle with this type of evaluation typically have not taken the time to figure out what the ultimate goal of the design should be. Lacking that information, designers are left to grope in the dark, looking for something that either a consumer, or a company leader, will say they like. This may work in the short term, but ultimately the wrong evaluation will lead to the wrong results.
How does evaluation happen at your company? Ultimately, are products evaluated based on how well they will achieve the ultimate goals of the person who will buy them? By ultimate goals, I mean the basic reasons the customer is buying a product in your category in the first place. If you're scratching your head wondering what the ultimate goals are for the product you're working on, it may be time to start asking some tough questions.