SNA: Survey dimensions or how to master data complexity

My first approach to SNA/ONA was rather simple: Collect data about who knows whom! This is no more than a binary, quadratic matrix and hence of quadratic complexity. And assuming, that there is an upper bound for the number of contacts a person can have, complexity even rises just linearly.

However, little later the demand for data increased dramatically. I wanted to observe the development of the relationships on the timeline. This is very important for measuring the success of the survey. And so an additional factor to complexity appeared. But this by far was not the end. In business settings it is dangerous to equate contacts maintained just by coffee dispensers with production critical ones. So another measurand had to be added - the quality of the relationship. And we even added its counterpart: the quantity or intensity of the relationship, because there is a difference in meaning between the now-and-then-relation and day-to-day business.

So how much complexity is useful? How many questions can be answered seriously by the participants? Where is the limit? Without clarifying the goal of the survey it is difficult to decide which information is really necessary. But from my point of view one should avoid node data wherever possible and focus on the network. Further on, I think in business settings visualisation is the key. It is hard to explain people - and in particular the management - abstract things they cannot see. Hence the next question to answer has to be: How much information can be presented in a picture? And how?
We probably have to find best-practices visualization methods for the different dimensions of network data preventing confusion and information overload.