For years, the page view served as one of our most simple metrics for gauging the size and scale of a web presence. One might call the page view a staple. As you read this, however, the page view is dying a slow death, losing much of its meaning and relevance to discussions about reach, engagement and overall impact.
Why is this happening now, after publishers, media buyers, research and technology companies have learned to rely on the page view as a basic unit of measure? It all has to do with underlying technologies and how they've changed the way we interact with web pages.
In the early days of the web, most content was consumed in page-size chunks. A publisher like ESPN might run a piece about a Florida Marlins game that contained a static box score. The content of the page didn't change materially over time, so it was logical to measure the readership of that particular article by looking at server logs to find out how many times it was accessed. Since a page view is essentially that -- the number of times a page was accessed -- this simple metric became the foundation on which to base our sense of how much traffic a website, a content area or an individual article would receive over time.
These days, however, content is more dynamic. Someone reading up on baseball box scores at ESPN might see that score is updated multiple times as the game is transpiring. Newer technologies allow for people watching the score to see it updated without having to refresh the page they're currently viewing. Thus, multiple exchanges of information can occur within a single page view. This dynamic alters the value equation: one request for information = one page view. What we're seeing is, in effect, page views within page views.
There's another issue with page views, too. Since the industry has adopted page views as a basic metric, various problems have surfaced in deriving intelligence from the data. Since we've been using page views as a surrogate for reach and engagement, sometimes we forget that page views don't necessarily equate to web users requesting content. In a well-publicized case last year, Nielsen//Netratings cut its traffic estimates for financial site Entrepreneur.com when it discovered that pop-up ads served by Entrepreneur were being counted as page views. Cases involving ad campaigns and traffic-driving deals designed to boost traffic numbers are often publicized less. These are the types of problems we encounter when we use surrogates for measurement and lose sight of what really matters to advertisers, publishers and end users.
As this controversy brews, one of the key questions being asked is whether our entrenchment in the current measurement methodologies will stifle the development of easier-to-use web pages and applications. After all, if a single AJAX page view takes the place of 10 static page views, it's a huge concern for web publishers. That one AJAX page view will currently be counted once, while leaving things static will promote higher numbers within syndicated research. Is it any wonder, then, why some web publishers might steer clear of AJAX and other such technologies, since widespread use might damage comparative measures of traffic?
In many cases, traffic numbers have taken years to build. Why damage them now, even if dynamic technologies might provide a better user experience? The stakeholder who stands to lose here is the end user, and since end users are the anchor for the whole value chain, failing to serve their interests does a great disservice to the industry.
To their credit, many online publishers have moved forward with technologies like AJAX and Flash, putting the user experience first (and web measurement be damned). Yahoo!, for instance, has embraced AJAX, relying on the technology within its Yahoo Mail service and others. It also hosts a number of resources within its developer guides that encourage the use of AJAX for things like displaying maps. While it's refreshing to see a publisher put users first, it makes one wonder what the overall effect will be on Yahoo's measurable traffic and overall ad revenue. It also begs the question of whether a few large publishers can force a change in how we measure traffic online.
As publishers graduate to dynamic updating of information, the page view loses more and more of its meaning. Whether this is a good thing or a bad thing largely hinges on how long advertisers and agencies will continue to rank advertising opportunities by the number of page views a site is able to garner.
If page views can't be a viable long-term metric, then what other measures ought we to be using to assess opportunities?
One might suggest that we need that all-encompassing definition of engagement that advertisers, agencies and publishers have been struggling with for the past couple years. An engagement metric would be advantageous, since it would get closer to what advertisers seem to be most interested in: opportunities for exposure to a message in a compelling environment. But the years of bickering should give us pause. How long before stakeholders can agree on a definition? And once that's decided, how long before syndicated research companies like comScore and Nielsen//NetRatings can both measure it and build up the requisite historical data to make comparisons meaningful?
Others suggest that a modified page view might be a short-term Band-Aid. The industry can continue to derive meaningful comparisons, for the time being, by reclassifying certain XML requests as page views. But this approach likely doesn't have long-term viability, as it's likely to open up a debate concerning which XML requests are valid requests for wanted information.
One thing is certain. The page view, as we've come to understand it, is no longer a reliable surrogate for audience engagement.