Are "Big eBook Deals" worth it?

Just read "The 'big deal' approach to acquiring e-books: a usage-based study' by Terry Bucknell, from Serials 23(2), July 2010.  This article is an evaluation of a "big ebook deal" based on usage, specifically a Springer collection deal by the University of Liverpool.  Here are some of the conclusions in interesting points:

• eBook chapter use was comparable and connected to ejournal use on the same platform.
• A good number of titles were used per subject area (40-60%) except for Math and Stats (not surprisingly).
• There was a reverse correlation between ebook and print book versions of titles, suggesting that there are clear preferences per title.
• Most titles were used at least once within two years.
• Past usage is not always a good predictor of future usage:  high-use titles in one period were not so in another.
• Mostly long-tail usage.  There were no clear "winners" dominating usage.
• Cost per chapter were good compared to available benchmarks and ejournals.
• Overall, this was a good deal, even for STM subjects for which ejournals are often stated as preferred.  (This suggests difficult budget issues.)

This makes me want to do the same kind of analysis for the collections in my library to see if the numbers are similar, especially since this kind of acquisition in only growing and many of the conclusions are contrary to our intuitions.  For example, this suggests that:

1. We should avoid overlap between ebooks and print books.
2. PDA programs are not as useful as they may seem, at least in large collections.
3. We need to push for better budgets to accommodate ebook collection acquisitions.