A few weeks ago, I did a post on using the WSS usage reports even though they are hidden when MOSS is installed (The Other Usage Reports). Well, this week I discovered a quirk with them that’s worth noting.
The setup:
I have a client that is storing documents in SharePoint. These documents are getting categorized several ways with additional metadata. We are then targeting, filtering, rolling up, and displaying links to these documents on various pages using the Content Query Web Part (CQWP).
The catch is that some of these files are too large to be pushed into SharePoint, others are not actually files, but instead are external web sites or Lotus Notes applications. To the end user however, these are all just resources to be categorized and displayed. To those ends we are using the Link to Document content type in our document libraries.
The impact:
When you create a link with the link to document content type, an actual .aspx redirect page gets saved into your document library. To the end user, this solution is working great, but this week the site admin got curious about the usage stats and his confusion helped me uncover something strange.
Within the context of the MOSS usage reports (for a publishing site at least), these link to document .aspx pages are excluded. The reasoning for this exclusion may be that these are intended to be documents that live elsewhere in your SharePoint environment and you would want to count their hits in with the actual document and not the link. 
Whatever the reason, it’s unfortunate for us because these are links to external resources and it would be nice to know whether the link here is being used.
Within the context of the hidden WSS usage reports, these redirect pages show up, but the numbers are not accurate. It seems that the intelligence that excludes the search engine hits from documents in a library or publishing pages (.aspx) in the pages library is not applied to the link to document (.apsx) items in a document library. The good news is that it’s pretty easy to spot the indexing hits – if you have 50 redirect links and half of them all have exactly x number of hits, it’s a fair bet that you can subtract x from the hit counts on the other half. The bad news is, well, pretty much everything else. It’s going to be difficult to explain the problem to your users. If your users aren’t thoroughly interested in the stats, they won’t go through the trouble to get the accurate. And if they ARE thoroughly interested in the stats, then they’re going to want them accurate and this solution will likely not be sufficient.
Conclusion:
Meh. I think it’s widely accepted that out of the box SharePoint usage reports are not good. They are difficult to interpret, but at least they’re inaccurate J.
