An excerpt from a recent chat with a friend of mine. (She's the most knowledgeable person I know at web analytics.)

The conundrum... 

As we progress with personalisation by using tools like Test and Target or BT Buckets to serve different content to different segments (such as individualised headlines or navigation based on prior behaviour) or targeting products to different geographic areas (like high end products for wealthier zip codes), the people who manage the business are seeing different content to their customers.

In addition, it can be hard to say at any stage what content has been targetted at what customer, as algorithms work and change so quickly.

My fear is that this can lead to bad decision making.

My answer... 

The first thing that occurs to me is that this is not a new Governance issue - but a new take on an old issue. I think the essence of it is how to package information and data for decision makers about complex problems of which they may have little direct knowledge or first hand experience, etc.

I think the solution is 2-way...

1. Decisions makers need to realise complexity is such that they will never know everything and need to "trust the data" as far as possible.
2. Web managers also need to know that they can never know everything, but need to extract & present the most critical information needed to steer the ship (with warnings about levels of data confidence).

In the past, many Senior Execs were not *educated* in what online was about (& often not interested) - and as such made decisions on first impressions, conjecture, rumour, what they heard in the pub, etc.

The challenge then was to educate senior mgt before they made investment decisions. Generally, this was reasonably simple as decisions were made infrequently and had quite low impact. (See my old article on ALA about this: http://www.alistapart.com/articles/educatingstakeholders/)

A newer side of this problem is that even if Senior Execs *get the web*, they may be dislocated from the experience of customers due to complexity.

A simple example of this is...

A Senior Exec speaks German, but manages an enterprise with a significant Japanese audience. She often looks at the Japanese website but it is gobbledygook to her. There is no way she can put herself in the seat of customers.

How does she make decisions for the Japanese market?

Well, she needs to trust the data. Identify key information and work off that basis. The challenge for her staff is to present such data (about Japanese online trends & behaviour) in a way that a German can understand.