Forum Moderators: DixonJones
There are some massive differences between the recent stats (for the new site) and the previos stats (old design).
Average visit length has nosedived (from 12 minutes to less than 1 minute).
Total Visits have rocketed (more than doubled), but page views have NOT (nowhere near the same increase).
The client is asking why there is a huge discrepancy between the stats now and previously. My first response is that you can't compare stats from different sites (oranges and apples), but this is not going to satisfy the client.
I need a better explanation as to why webtrends is now recording way more VISITS than before.
Any ideas or a reasonable explanation?
Are you using the same stats program now as before? Any configuration changes there, in how a "visit" is measured? Is he hosted the same place, and on the same server platform?
Where is all this new traffic coming from? Search engines? External links? Bookmarks? SE spiders? Did his standings in the SEs change drastically? Did the page names from the previous site get changed? If so, was an automatic server-side redirect put in place to send visitors to the appropriate page?
Are people staying a shorter time because they can't find what they're looking for, so they give up? Or are they finding what they're looking for faster than before, so they complete their business and go on their way in far less time than before?
Does he do online sales? How does the conversion rate compare before and after?
What about other visitor interaction? Are people using any contact forms, requesting quotes, etc.? Is he getting any feedback from visitors? Has the volume of feedback gone up or down? What are people saying? Are they complaining about the new site?
Have you done any usability testing? Did the new design substitute Flash-based or JavaScript navigation for old-fashioned href link-based navigation? Is the new site entirely in Flash? Does the html and css validate?
It could be he's just getting spidered massively for some reason -- does your stats program count spider visits as "visitors"? It could be he's doing outstandingly in the SEs, hence increased traffic, but the new site is terrible and unusable and people leave quickly, disappointed. It could be some popular site is linking to his site but misleading visitors about what they'll find there -- causing them to follow the link then quickly press their "back" button. It could be the new site is SO much more usable that people can accomplish their objective in a fraction of the time that the old site required.
There are a lot of things to look at here with no quick and easy answer. Is he willing to pay you to spend the time needed to research this anomaly?
Are you using the same stats program now as before?
If not, then that's the place to start the investigation. The biggest differences occur when switching form a logs only based system to.... well to a non-logs based system.
Bottom line is that there's going to be a logical explanation for this and that it's going to be in the logging or sessionizing. It's NOT going to be much different visitor behavior, I'd bet.
Turns out the configuration in the software for a "Visit" had been changed from 30 minutes (default setting) to 1 minute.
Arghhhhhhhhhhhhhh! What a disaster :-(
That explains it all. I must have been testing something, and forgot to switch it back to the default setting.
I mean, can you change the setting for just one profile, without affecting all your other profiles?
