Is “Stickiness” Good Usability, or Page Views?
By WellcomeMat on Mar 6, 2007 in WellcomeMat
We recently pushed out a ton of new features on the site, and have a million more to go before the end of March. However, over the course of the last month or so, we have noticed a very strange development that I think begs a very important question:
Does improving usability decrease total page views per visit?
Our last major release was about 80% usability improvements. Yes, we have a long way to go, but what is important is the effect that it had on the average number of page views per visit. In every major category that we watch for growth, WellcomeMat saw a very healthy improvement (total number of new members, total number of video uploads etc.). However, the average number of page views per visit went down almost 13%. It seems strange to me that page views are still being used to declare the “stickiness” of a site.
Popularity: 9% [?]
On Mar 6, 2007, Jason F. said:
If you’re equating stickiness with page views, then a key usability feature would seem to be the ability for users to SEARCH…locations, video content type, videographers and other categories. I have a feeling you’re going in that direction…right?
From a videographer’s standpoint, each set of new features is like a visit to the candy store.
—j
[Reply]
On Mar 6, 2007, John Schroeder said:
It is an interesting question. And yes I do think that there is something about people having to stumble around a site that leads to higher page views. Increase the usability of the site and they go where they want get the info they want and move on.
I think that some people still use page views as a measuring stick because its easy to track and it’s universal. From one site to another page view stats are easily comparable. But what one site wants from its visitors compared to the next site could be quite different. These differences are harder to compare.
I have certain goals with my site that are ever evolving. Measuring how I am doing with these goals is much more important to me than page views. Your new members and new video uploads as you know is what will build your business.
[Reply]
On Mar 7, 2007, WellcomeMat said:
Thanks for the thoughts John. Our goal will always be to have the best usability possible, and we won’t bend where that’s concerned. Efficiency is what we have our largest debates over.
[Reply]
On Mar 7, 2007, WellcomeMat said:
J–
Yes…we will re-incorporate search (it was still live until two months ago or so) when it is the best it can be. But, page views will not see a large increase because we try to eliminate as many clicks between our users and desired content. In other words, if we can make our search one page, using AJAX/Flash/PHP, you you can bet we will.
[Reply]
On Mar 7, 2007, Jason F. said:
As a point of clarification, I may have misinterpreted good “page views” as the number of sucessful searches for content - like the ability to access/view all the vacation rental properties on Cape Cod - versus bad “page views” as it relates to the number of clicks a user would need in order to find the information they’re looking for. Does that make sense?
The Ultimate One Page Search: I know you guys are good for it!
—j
[Reply]
On Mar 7, 2007, WellcomeMat said:
I see where you are going with this. To me, the worst thing site developers could do is load a new page completely to serve additional functionality that doesn’t require a full page. Ajax and Flash alleviate this by allowing you to reload only a portion of the page. This is what most people refer to as “2.0″ (the most annoying, but applicable, catch phrase). What is cool for ad revenue sites is that people are now figuring out ways to change the URL address when someone loads new content. This is not good for advertisers, still using the out-dated page view number to justify spending. The URL changes and you have a new page view, whether you are serving up one page, or 1/10th of a page.
[Reply]
On Mar 9, 2007, Tony - vidlisting.com said:
Agree that page views don’t make sense as a performance indicator when using AJAX.
Would “total video views” or something along the lines of “video search results actually watched” work better instead as possible metrics?
At the end of the day, feature improvements on a video site should make watching videos easier. I’d think that improvements in usability would, in some way, result in more videos watched even if page views go down because of AJAX-related decreased page loads or efficiency improvements related to searching/filtering.
Tony
[Reply]
On Mar 9, 2007, WellcomeMat said:
Tony,
You’re right, and we watch total video views closely. They are always meant to be the center of attention on any WellcomeMat page. Great ideas man.
[Reply]
On Mar 11, 2007, Florencia Coelho (Argentina) said:
Hope you profit from my investigation in web metrics.
NYT
Internet ratings service comScore is altering its method of counting Web usage to take into account many Web 2.0 innovations, including AJAX-driven pages, which cannot necessarily be counted in page views.
http://arstechnica.com/news.ars/post/20061225-8493.html
The use of pop-ups as an advertising technique to serve content is creating headaches for metrics companies like Nielsen//NetRatings and regulators.
The web traffic conundrum and CNET’s shrinking page views
Is the web and its advertising the most trackable medium ever? Not
exactly. Despite the hype over the Internet being so trackable, the
“dirty little secret of Silicon Valley is that no one knows exactly
who is going where on the Web,” according to BusinessWeek’s Sarah
Lacy. All the third-party services such as comScore, Nielsen, Hitwise
and Alexa have their weaknesses, and they rarely line up site traffic
with the site’s own internal servers. Lacy believes the best hope for
fixing this is for the big and small online ad players to work with
measurement services to improve their offerings. She also thinks new
metrics such as time spent or a measurement that combines a set of
factors might emerge. “Until that happens, though, the Internet will
have to deal with the discrepancies,” Lacy concluded. “And Web
metrics, like company valuations, will remain a crapshoot.”
One example of the crapshoot nature of gauging traffic is the recent
brouhaha over CNET’s declining page views on comScore. TechCrunch
blogger Michael Arrington made a stink over it, noting that September
2006 traffic for CNET was 616 million page views vs. 1.37 billion in
September 2005. But the problem is that Arrington didn’t mention that
CNET had redesigned its pages “to streamline navigation to create
better user views and reduce the number of page views per user,”
according to new CNET CEO Neil Ashe in a conference call. While
Arrington and others believe that independent tech blogs are eating
into CNET’s traffic, CNET VP Dan Farber countered that CNET and ZDNet
have their own blogs that compete well with the independents. “There
is no lock-in, big or small,” he wrote.
Web Numbers: What’s Real? (BusinessWeek)
http://www.businessweek.com/magazine/content/06_43/b4006095.htm?chan=tc&chan=technology_technology+index+page_more+of+today’s+top+stories
CNET Is Bleeding Traffic (TechCrunch)
http://www.techcrunch.com/2006/10/13/cnet-is-bleeding-traffic/
CNET Networks Q3 2006 Earnings Call Transcript (Seeking Alpha)
http://internet.seekingalpha.com/article/19042
CNET, Traffic and the Blogosphere (Seeking Alpha)
http://internet.seekingalpha.com/article/18829
[Reply]
On Mar 12, 2007, jf.sellsius said:
IMO, Page views MAY be a useless state.
Needle in a haystack, inefficient search leads to more page views–but ultimately frustration and a bail out. Efficient sites MAY lead to less page views because they find what they’re looking fo faster or more page views because people are enjoying the site and want to see more.
A better indicator is measuring the user navigation clicks to see where they go & where & when they bail out.
Measuring repeat visitors may be a better indicator of your site’s efficiency and value.
We did a post on this relative to blogging. We talk about “magnetism” vs. stickiness.
http://tinyurl.com/2wtlgk
[Reply]