Comment: Web 2.0 and the death of the page impression

by Paul Cook, founder of PositiveFeedback, Revolution UK 25-Jan-07

I find BBC2's Dragon's Den entertaining and a good reminder that businesses need to be built on sound principles. So, I was shocked when the panel got excited about a poker affiliate attracting 50,000 hits. No-one asked over what period this meaningless metric was counted, let alone how many people it equated to and whether any had signed up.

It seems funny that, 10 years on, people go on prime-time TV and talk
about measuring a site in terms of hits. Despite the fact that it can't
be done for Flash sites, people have been happy to measure site

popularity in terms of pages. To double impressions you can split

content from one page into a menu and sub-pages. People may stay on the
pages for half the time, but you can serve twice as many ads.

The movement towards table-based web design made impressions easier to
quantify, so they've been the main currency. Yet, if you redesign your
site your page impressions will be irrelevant, so it follows they don't
give the best answer when trying to evaluate which of two sites is most
popular. However, web 2.0 will change everything.

Web 2.0 lets designers present sites in a more synchronous way. While
the user accesses the page, relevant content is constantly requested
from the server. The result is a page impression that lasts for minutes
and generates a constant stream of hits. The solution is to accept that
the web is an active medium that has more in common with TV and radio
than newspapers. Unfortunately, current measurement revolves around
pages, partly because the media owners who dominated the web in the
early days were newspapers.

Web media needs to be measured in terms of users and hours. If a sports
site has 100,000 unique users and one million user hours per week, we
can evaluate that instantly. If we know one million user hours equates
to a five per cent share of the total hours spent by UK users accessing
sport, better still. Panel-based measurement could calculate the size
of the market while site-centric measurement could be used for
individual sites. Of course, measuring unique users is easier said than
done, but these problems are not insurmountable and the result would be
a measurement system that allows sites to be compared fairly. And,
you'd be talking in terms that offline media buyers understand.