I’ve just been building a volumetric model for a large volume consumer-market website. We wanted to make sure that we sized the initial launch infrstructure correctly, and had an understanding of the cost of growth (of both transactions and data volume).
It’s been a doubly useful exercise – as well as giving us the figures we needed, it has also highlighted an unexpected consequence of (an otherwise very sensible) archiving policy decision. This has allowed us to tweak a policy and create a new approach to one aspect of marketing, saving thousands of pounds per year.
Certainly good anecdotal evidence to support the old programmers adage that “it’s cheaper to fix it on paper…”!