Welcome to the new seo game of slow clicking link bots.

Seriously, how are they going to tell how long a visit is on your site if there's no real way to tell when a visitor has exited and/or returned?

Well unless maybe they are going to use some pixel beacons across websites. But that's nasty.

[edited by: amznVibe at 11:45 am (utc) on July 10, 2007]

That's right. I spend five minutes ordering from an ecommerce store, but twenty minutes watching a streaming video. Goes to show that the streaming video site is so much more attractive. No. Goes to show that the ecommerce site did it's job and did it well. if it took more time then I'd probably have given up part way and shopped elsehwere.

Open a page, get called away & return 20 minutes later and continue.

Click a link that opens in a new browser and continue using the new window leaving the previous one open.

A lot of hogwash!

I think this is a good metric for longer articles and blocks of information, like newspapers or wikipedia. Though vincevincevince has a point about ecommerce, it needs a different way to determine popularity. But I also imagine that store owners know this and won't pay it much attention. This, to me, shows that there is no one good metric to use. Neilsen needs to offer different reports on traffic.

Give it up Nielsen. You have a flawed product that misleads naive media buyers. How will scrapping page views for time spent change that?

most ecommece sites aren't in the content business and, therefore, dont' have to attract advertisers. They do not care about nielsen rankings. So one couldn't easily compare a video site with a commerce site, in my opinion.

this move will make nielsen rankings account for sites that are using AJAX. Pageviews, as an advertising metric, is slowly dying... and the fact that a nielsen is willing to adopt more accurate ways of measuring traffic is a good sign of their ability to adapt.

Another interesting point is that a site which attracts you to stay for a long period of time probably has a good reason that it does so. Or: you can't compare apples to guava.

Examples:

Factual content website with targetted advertisments - user will be reading and notice adverts for much of the time he is on the site - higher page-time is better for advertisers

Streaming video site - will engross the viewer in the video 99% of the time - actual 'advertisment susceptible time' is much less than the time spend on site

Online game (e.g. board games) - high user concentration on the game, only a short period per page view that advertisments will be noticed

Downloads site - the long page views are downloads, during which time the window is minimised and the visitor is elsewhere - same goes for streaming radio

To me, Nielsen is terrified of the fact Google now knows far more about all aspects of site and advertising performance via search click tracking and the Analytics product and could wipe out all competition in the sector with ease if they started releasing that kind of ranking data.

Nielsen - Never listen to this guy.

Well this sounds along the line of Google's Search Patent explained here [seomoz.org...]

Google could start using Google analytics data to track which web pages for which queries receive click through and the length of time spent on the site to determine relevance, and this could affect organic SE rankings.

I agree this is stupid for ecommerce as usability problems can increase time on the site.

As a current and past Nielsen family (different product groups) I think their past stats were skewed for various reasons, and their attempt at tracking the net is way off as well. I'm truly surprised anyone listens to them at all. I can think of at least two or three sites I have bookmarked and visit at least a couple of times every single day (heh, G Adsense, ought to rank high on SOME people's-I-know ratings :) to look at a few numbers. Sometimes I even leave those windows open all day! How about when you have multiple computers which you use a KVM box to switch between like I do? Or multiple active computers in a house (I have 5)? You can often only see one screen at a time. If they're smart they can track what window is actually on top and only count that, but mult. computers will still be an issue.

Surely the fundamental problem is that you are trying to measure audience of video (in effect, TV), usage of software (online apps), readership of traditional sites, and a host of other things with the same measures. There is no sane way of doing it. No one tries to used the same measures for TV and magazines, do they? (except insofar as circulation is roughly the same thing as reach).