|Visitor path anlysis - any value in this ?|
What use is this when the most popular paths are single clicks ?
I've looked at lots of stats (using WebTrends and other packages) for all kinds of websites, and I've yet to find anything useful in the Visitor Path section.
Where is the value in this when all you get is a long list of mostly single access pages ? Most visits are single page visits, and then (much further down the list) 2 page visits. This is no good to clients, so what's the use in having this section ?
This can't be the way to analyse routes through the website, so what is ?
If the overwhelming majority of paths are single access sessions that in itself is an indication that there are some problems somewhere. Ad placements may be bad, the site architecture may be confusing, the entry points may not provide visitors with whatever it is they expect. Maybe there is no need to go past the first page?
For analysis of paths with more than a single access session, simply throw out the single access sessions and look for common paths. When people do dig into the site, where do they go, what do they do, where do they bail, do they accomplish the desired goal?
It's never black and white but some inferences can be made and a few hypothesis testesd.
Fom2001UK, I tend to agree with skibum. While every site gets one-page visits, if they represent the majority of the paths, it sounds like there is a real stickiness problem - people arrive, and don't like what they see. Further analysis could include where the one-page vs. multi-page visitors are coming from. Perhaps the one page visitors are finding the site while looking for something else entirely.
I'd recommend looking at the most popular entry pages (presumably the biggest one-page paths, too) and trying to figure out how to induce visitors to navigate around the site. Is the entry page attractive and connote something useful and professional? Are there navigation links that are inviting and obvious? Are the links specific enough to let people know what they are getting? For example, IMO people would be more likely to select a link titled "The Top 5 Widget Design Mistakes" rather than the more cryptic "Design Criteria".
Assuming you have some preferred paths - e.g., those leading to an order, and information request, etc., be sure that you steer your visitors in the desired direction.
I'm not buying that at all. The sites I'm looking at have averages of 6 and 7 page views per visitor (WebTrends tells me that in the main summary).
So why is this not reflected in the visitor paths ?
Typically they'll list a single access to the homepage as the first path (10 - 20% of hits). This is followed by a long list of single page accesses (with very low percentages).
There's something wrong with the way I've configured it, or the way WebTrends is reporting it.
I'm blaming WebTrends :-)
In my experience it's always useful to see where people have left the site. Unfortunately this isn't as clear cut as it first appears.
Although in theory if people all leave when they hit a certain page it would tend to suggest that this page is poor or lacking in information, it can in many cases be that this page contains useful information that answers the users original query and they leave happy.
I would say that particularly on large sites the pathways users are taking can be very valuable information. this does rely on the fact that this information is correct ;). It may give an indication of how good/bad your site navigation is..
Fom, 6 to 7 page views per visit sounds pretty good based on my experience, and why all your common paths would be one-pagers is hard to understand. (Certainly, having a number of one-page paths is very common. Many people arrive, take a look, and leave - I've seen sites where the home page alone was the most common path, followed by a couple of other entry pages that also became exit pages.) I suppose if there was a very high degree of randomness to the long visits, none of them might appear in your top paths... This seems unlikely, though - I usually find that there are some common paths even on large sites, e.g. "home page (usually a major entry page) - obvious/popular link1 - obvious/popular link2)". If you are looking at the visit numbers, and they don't seem to match the paths, there may be a problem with tracking the paths. Webtrends does have some difficulties with AOL and other ISPs that request from multiple IP addresses on the same visit. If this was a major problem, though, I doubt if you would be seeing 6 or 7 pages per visit. Multiple IP problems tend to depress that number, sometimes to the point of having more visits than page views.
Looks like it might be time for some number-crunching to see if the path data is way out of whack with the rest of the data, or whether you are just seeing some really strange visit patterns. :(
We have the same type of thing going on with our WT stats. 10-14 pages per visitor, but WT shows a solid 20% who only hit the homepage. Does anyone know if it cleans out spiders when it makes this stat? I'm thinking it does not, hence the skewing.
I get the same problems with NetTracker, .css and .swf files etc seem to be logged in the paths, which in turn screws up vistor duratiopn and paths summaries/averages. I dont even look at them anymore!
I had the same problem with the path report on NetTracker. As soon as I turned on the host resolving the paths came together.
Its also intresting to see the page views by search engine. Google returns the smallest page views (per user)for me and its because they bring the user closer to the info they wanted to see.
>>Its also intresting to see the page views by search engine. Google returns the smallest page views (per user)for me and its because they bring the user closer to the info they wanted to see.
This statistic could also be partly explained because the user uses the google cache to see the 1st page from your site during their visit. An alternative explanation could also be that Google is returning irrelevant results :)
>>This statistic could also be partly explained because the user uses the google cache to see the 1st page from your site during their visit. An alternative explanation could also be that Google is returning irrelevant results <<
Hmm cache, maybe. Irrelevant, Nope.. Product pages, Google gives em just what they asked for :)
The data WebTrends spits ot with respect to page views per visitor would seem to include spiders. I don't know that for sure but it seems only logical. WebTrends often reports that with a large percentage of single access sessions, the average time spent on the site is up to 10 minutes. This I believe is another faulty measurement in WT which probably includes spiders. The median is a more realistic measurement.