Measuring traffic from an index to unknown destinations

If the heading of this thread is confusing, it adequately represents my current state.

First, a little background.  We have an index of technical documents reside throughout our website.  Actually, we have many indexes, but for simplicity's sake, let's say there's only one.  The technical documents, numbered in the thousands, are not grouped or tagged and have no common naming or title conventions.  These documents sit within several virtual folders, along with other documents of similar subject matter.  To complicate matters, the index is generated automagically by user selections of criteria from drop-down lists.  The index page exists as a static entity, but parameters are added to its URL depending on the user selection as it refreshes with the resulting index.

I am being asked to find out how much traffic the index generates to these technical documents, as some people are interested in seeing the difference between traffic to the index and the traffic to the technical documents.  It's one thing for people to visit the index, but are they using it and actually going to any of the indexed pages?

As it stands, there is no plan to tag or adopt a common naming convention for these documents.  Since there are so many of them, I can't imagine manually entering their IDs to create a content groups.  At least not in the short term.

Any thoughts on how I can measure traffic to these pages from the index without grouping, tagging, renaming or retitling them?

Thanks,

Mike

Views: 52

Reply to This

Replies to This Discussion

Simple. :-)

Not*that* simple maybe, but not very very hard.

1. I can only assume that tagging is impossible for some other reason, but tagging all the content would be a solution.

2. If you don't have tags, do you have logs for accesses to all the 1000's of documents? You could load those and generate reports from there.

3. Given that the logs probably suck, what I'd do is tag the index pages and dynamically (unobtrusive javascript juju) instrument the links to the documents (ntptLinkTag, so there's a slight delay to allow the tag to fire before the index document unloads).

3a. likely tag the links using the URL of the destination as the lc= and a pv=1 to force it to be a new page view (allowing pathing in the UI)

4. Either in the JS, parse the destination URL (or if the index page has any metadata on the documents, add that to the tag) and add category information as pagetag parameters AND/OR create some page-based parameters to extract the virtual folder information (I'm guessing that's part of the URL)

5. If you don't have access to categorisation/meta data on the front-end or in the URL but you do have a back-end database, you might be able to expose the information in there via a DME.

Did I ever mention that we offer professional services to work with customers on this sort of thing?  :-D

b

Hi Mike,

Assuming that page tagging is not an option (based upon your comment "there is no plan to tag or adopt a common naming convention for these documents") then I think your only hope would be go crazy with URL Search and Replace rules, breaking out each of the paths of the path tree into their own parameters. Once they are their own parameters you *might* be able to find some commonality that could map back to a lookup index else where.

You could also, in addition to above, keep the entire URL/page and take advantage of something from way back in the day, and I think they are still available (Bob or Todd should have them somewhere), plugins called ntlookup. There was one specifically for Vignette (nee http://en.wikipedia.org/wiki/StoryServer) which depending on your setup could solve your issue. The ntlookup plugins take your *whatever* input string and can do a lookup else where (like another database) and either rewrite the value or save the original value and just present in the UI the lookup value.

I hope this gives you a couple other ideas.

Lee Isensee
Director, Solutions Engineering and Product Strategy
Localytics
http://www.localytics.com

 

Bob Mitchell said:

Simple. :-)

Not*that* simple maybe, but not very very hard.

1. I can only assume that tagging is impossible for some other reason, but tagging all the content would be a solution.

2. If you don't have tags, do you have logs for accesses to all the 1000's of documents? You could load those and generate reports from there.

3. Given that the logs probably suck, what I'd do is tag the index pages and dynamically (unobtrusive javascript juju) instrument the links to the documents (ntptLinkTag, so there's a slight delay to allow the tag to fire before the index document unloads).

3a. likely tag the links using the URL of the destination as the lc= and a pv=1 to force it to be a new page view (allowing pathing in the UI)

4. Either in the JS, parse the destination URL (or if the index page has any metadata on the documents, add that to the tag) and add category information as pagetag parameters AND/OR create some page-based parameters to extract the virtual folder information (I'm guessing that's part of the URL)

5. If you don't have access to categorisation/meta data on the front-end or in the URL but you do have a back-end database, you might be able to expose the information in there via a DME.

Did I ever mention that we offer professional services to work with customers on this sort of thing?  :-D

b

Reply to Discussion

RSS

© 2017   Created by Wendy Ertter.   Powered by

Badges  |  Report an Issue  |  Terms of Service