Stickiness

How important is this statistic?

[cyril kearney]

Let's say a user enters my site at noon and he looks at ten pages. The last page is logged at 12:20.

That would mean that he was exposed to my site for 20 minutes. That measure is called stickiness.

Does anyone use a number like this? How do you use it, I mean what value do you get from it?

How do you handle the built in error? You don't really know how long the person looked at the final page.

[engine]

I'd add to the stickiness definition with visitor returns to the site on a regular basis.

Stickiness on an information site may be totally different from an e-commerce site. An e-commerce site would classify stickiness by regular re-ordering.

I'd also suggest that some people will rapidly go through a site by printing out the pages. In addition, they may leave the page open while they continue to surf in another window (or go down the pub without switching off their machine).

[tedster]

On another thread, someone mentioned noticing that their Google traffic gets fewer page views per visitor, but that may be because Google search results brings them closer to exactly what they were looking for. One more variable in the stickiness equation.

I can't see a scientific use for time spent per visitor, unless perhaps the sample is very large. Even then, the analysis depends completely on the purpose of your site. Going back to Google as an example, I'm sure they're delighted to see shorter times per visit but more repeat visits. But where ad impressions are an important part of the business model, it's a different picture.

>> You don't really know how long the person looked at the final page.

One of my clients is building a compex cookie system. One proposed feature is using onUnload to register the time a visitor leaves each page. That could handle the last page issue that cyril brought up, except for the visitor who goes off using other windows for their browsing session and never unLoads yours.
But is the overhead justified by useful information? I'm not convinced it is.

[skibum]

Like engine said it would very much depend on the type of site. Probably the longer the better on an ad supported site. Any site selling products from widgets to luxury homes, long visits could indicate very high interest or inability to find information ==> site needs rework. Combined with some other stats like conversions, path reports, single access sessions, etc. it might be posible to put together some kind of picture as to what the true story is. Any conclusions would probably require a lot of inferences and may or may not be on the mark.

[ggrot]

>>But is the overhead justified by useful information? I'm not convinced it is.

You might be referring to development overhead, but as far as processing and bandwidth costs this is really simple for the server. Just use the javascript onUnload to create a new image object and load it with a url of a 1x1 image on your server. It gets logged in with everything else and shouldn't affect the server at all.

[Brett_Tabke]

Time On Page (TOP) is an interesting proposition cyril. I've looked at the stat and the issues surrounding it and they haven't lead me to any conclusions. There doesn't appear to be anything there that I can base site building on.

Time on site is more a reflection of the style of site than a user metric. I think it breaks down to the style of site, the size of page, and the type of page they are reading. Some sites have large rich pages that are going to have higher TOP values.

The one place the time on page stat is important, is on larger sites where branding is so important. TOP is a good indication of the impact of your branding. If you are building for branding, the higher the TOP, the more successful the branding.

I think the best way to to use TOP info is to compare it to where you where 6 months, or 12 months ago. I still don't see being able to use those conclusions to a large degree.

[Laisha]

>How do you use it, I mean what value do you get from it?

Yup. It depends, too, on the type of site you're talking about.

We run a site that has, among other things, a news roundup for news related to the war on terror. We update this page several times an hour, and it has a 300-second refresh.

We have at least two people who "park" on the page all day, apparently while they are at work, judging from the times. Now and then, both make forays into other parts of the site, but most of the day, they're both just there on our news page.

I don't know quite HOW to interpret this, but I suspect it's a good thing: I would like to think that it means they use the page rather like I use the TV news, which is on (in the background) all day in my office.

I think that's a good thing.

Another site I have is about search engines, positioning, etc. If those two people were parking on THAT site all day, I would not think it was a good thing.