Comments for Page View Metric Dying - But What Will Replace It?

Page View Metric Dying - But What Will Replace It?

2008-02-28T21:13:59Z

We've all seen the signs. Ding dong the page view is dead... well, dying. First Compete announced that they would be using attention-based web metrics, or Attention Metrics for short. Then Facebook announced that they will move to a similar metric. Perhaps most importantly, Nielsen NetRatings announced last July that they would stop using page views for comparing popularity on the web, and move towards more attention based metrics. Also, Microsoft announced this week the release of a new ROI measurement tool called "engagement mapping".

]]> This is a guest post by Muhammad Saleem, a social media consultant and a top-ranked community member on multiple social news sites.

The reasoning is simple enough: While unique visits and page views are useful in measuring how much incoming traffic a site has, it isn't exactly a good or accurate way of measuring impact or even engagement. You could have high incoming traffic (for example, any site that is hugely successful on social sites) but if there is an incredibly high exit rate and only 30 seconds to a minute spent on the site, the traffic numbers don't mean much (i.e. not all traffic is created equal). Furthermore, the rise of new web technologies such as AJAX which don't require page reloads to refresh elements or modules in a page, or video embeds (such as from YouTube) that allow you to watch a video and then browse related videos without ever refreshing the page, are making page views a mostly inaccurate measure and rendering it largely irrelevant.

While most people agree that page views are becoming irrelevant, the same people are uncertain about the future. For example, many agree that attention-based metrics are the future. Attention metrics calculate the total time spent on a site or interacting with a page (or element on a page in the case of Facebook applications) as a percentage of total time that people spend online, to measure a site's relative importance on the web. However, there are many others, like the Tel Aviv-based Nuconomy Studio and even Yahoo's Buzz, that believe using factors like comments on posts, ratings from users, number of times something is shared, and clicks on ads as a measure of how popular something is is a better/more accurate metric.

The problem it seems, arises because there is a disconnect between the advertising industry and the publishing industry. The reason why there is an eternal quest for traffic, not only in terms of unique visitors, but also maximizing page views per visitor, is because advertising networks let you in on the basis of how much traffic you're generating, and your eventual income is based on the number of impressions (and clicks). While it is true that the page view as a metric is on it's way out, this isn't going to happen unless a new metric comes from within the advertising industry, which, with over $20 billion at stake, has the most to gain from a more accurate way of determining where to spend their money.

But it's not that simple either. As Scott Ross explains, different web technologies and applications have unique effects on different sites. What technologies you use and how they effect engagement and interaction on your site may depend on the size of your site, the niche you operate in, and a host of other factors. In fact, the metric that is most applicable could even change from page to page depending on the content on those pages. That being the case, perhaps one metric that is applied to everyone is just not enough and just not practical/efficient. As web technologies evolve, the page view is bound to die as a metric, but unless the advertising industry can get it's act together and work alongside the publishing industry, a good set of new metrics that would be widely adopted is not imminent.

Comment from evohollywood on 2008-02-28

2008-02-28T23:05:28Z

I think the future lies in producing a dynamic metric that represents a sort of "unified theory" addressing and accounting for all the different types of engagement a website can receive. I'm imagining something resembling a quadratic formula. Engagement = (Unique Users x Events (i.e. clicks, video views, ad interactions, comments posted) x Time On Site)/(Number of Different Content Pieces Per Page + Number of Unique Advertisers Per Page).

The result is something like an engagement density, directly increasing every time a unique user interacts with a specific element of a site whether its an ad or a piece of content and decreasing the shorter the time spent on the site or the more advertisers displayed on the page. With Microsoft moving towards "Engagement Mapping" and the other big players following I'm sure some standard metric like this will emerge.

Evan Moore

Comment from Peat Bakke on 2008-02-28

2008-02-29T00:19:14Z

Yes! The value of a web page has absolutely nothing to do with it's technical qualities. Hits, pages, bandwidth, load -- these are all terms that measure server activity, but have no direct connection to the human experience of engagement.

The quality of the experience can only be quantified within the context of that experience. Measuring the time on a page is important if the value of your content is oriented around a time based experience. Click funnels are important if your content is process oriented. Visitors are important if your site is community oriented. Comments are important if interaction is important.

At the end of the day, if you don't know what your goal is, your statistics will be useless. I'm skeptical of any single figure that sums up the experience across all web sites.

Comment from Shahar Nechmad on 2008-02-28

2008-02-29T00:37:02Z

I think you are absolutely right.
The biggest problem with the engagement metric (which of course I'm a big believer in) is that it is hard to compare between sites.
An engagement metric for an e-commerce is very different than the one for a social network.
That said, I do think that over time we will see the evolvement of an engagement metric for different verticals. One for blogs, one for video site, etc.

As for the advertisers... Actually they suppose to be the biggest advocate of this new metric. We at NuConomy simply see with our customers how measurement of engagement gives you a much better and productive way to understand your ROI from campaigns, users, content and many other things.
It's in the best interest of the advertisers to start and move there.
The problem of course is that the advertising industry moves very slowly, an it will take a long time until we will see major shift in using the engagement metric not to just measure yourself but also as an advertising currency.

Comment from Jason Lankow on 2008-02-28

2008-02-29T06:32:01Z

This is also particularly valid in the case of sites like myspace.com that force users to navigate multiple unnecessary pages to get back to a main screen rather than implementing a more fluid transition that required fewer page loads. great post!

Comment from Medical Furniture on 2008-02-29

2008-02-29T09:06:27Z

Perhaps the best thing is to create specific context for given industries and then comparing trends them with others within the peer group. You could then have a series of clusters of websites which have a measurable average of interaction. Ofcourse it goes without saying that I believe it is not how long someone spends on a page that matters. Is what they do there that can make the difference between a profit and nothing.

Comment from schoondorp.nl on 2008-02-29

2008-02-29T09:22:03Z

If you compare websites with tv programs, the number of visitors is as important as the number of viewers - in either case you can't be certain what the people are doing, if they are watching at all etc. The whole discussion about attention metrics sounds good, but is in fact an attempt to find fake certainty in an unpredictable environment. Having different metrics for different sites will obscure the whole industry. I would say: copy what is common practice and don't try to suggest that you can measure people's attention. Of course you can on some sites, but on the whole, measuring visitors is as good as measuring viewers - it is a rationale to come to a price..

Comment from Just Like Fred on 2008-02-29

2008-02-29T13:08:32Z

The whole idea of measuring quality with quantitative parameters is, if not absurd, then off target.
It's nice to see that someone is trying to come up with better alternatives.

It all boils down to the assumption that popularity equals quality.
If that was the case then Shakira would be a greater artist than Dylan. In my opinion, she is not. Some might disagree, and that is just fine.

Comment from PJ on 2008-02-29

2008-02-29T14:40:51Z

I just want to point out that NetRatings didn't state that they were dropping the page view metrics. The statement was that more emphasis would be placed on the metric in order to account for engagement compared to page views. The perfect example for this is Google Maps which has low page views but high time allocated. There is certainly no talk of doing away with the unique visitor metric; it is still the best way to measure reach and I cannot see that changing.

Whilst I like the idea of a wholistic metric in theory, it just cannot work in practice because, as Scott Ross suggests, there are many factors at play and each site is totally different. Such a metric would be interesting but I could see it being exactly like the Fifa ranking, totally irrelevant and unreflective of the actual ranking of each team.

We have to get used to the idea that users, time and views (in that order of preference) will be used and publishers/agencies alike will pick and choose the metrics they deem the most important

Comment from Henry on 2008-02-29

2008-02-29T17:46:09Z

This is an intriguing concept. It makes sense pragmatically because 1 visitor who stays and interacts with the site for 20 minutes is definitely more valuable than 3 visitors that stop by and look at the site and then leave after 30 seconds. We are noticing the difference in these phenomenon's at www.writiki.com on a daily basis. It's always different.