If you persist data (e.g. store it in a database) chances are that the data will be used by managers, statisticians, stake holders etc. to analyze the workings of their software-supported undertaking to make executive decisions. This analysis can only take place by methods of aggregation. There s no one in the world who can look at a million rows of raw data and glean insight. The data has to be summed, averaged, standard deviated etc. to make any sense to a human being.

A few examples of areas where data aggregation is important:

• Public Health (CDC, WHO)
• Marketing
• Advertising
• Politics
• Organizational Management
• Space Exploration

lol. Take your pick!

Very important, what else is there to say? I work at a large hospital and not only do we have numerous departments using Analysis Services cubes we develioped but they rely heavily on the daily totals and different aggregations they can derive from these cubes by simple browsing. Without the very basic capability of being able to aggregate on some portions of your data you might as well write it on paper (IMO).

Say you have data over every individual sale. Looking at these individual purchases could be interesting some level level(e.g. whne a customer comes and wants a refund) However, I cannot send those 20 million records to my boss at the end of each month and say "Heres how much we sold this month".

This data needs to be aggregated and summarized on various levels. The business would not operate if the marketing guys couldn t get an aggregate for each product, the regional boss couldn t get an total aggregate over a time period and so on.

Our databases have millions of rows, of course we rely on aggregation for managment information, not to use it would be to put too heavy a load on the database in order to run large reports which would impact heavily on the users of the database. I can t think of many cases where the database contains business critical information that managers use to make decisions where aggregation would not be needed for managment reports.

I view data aggregation like data in a grid and being able to group, order, and sort columns. In a large grid of data, this is very important. It s really the difference between looking at a pile of numbers and looking at meaningful data.