me thinks they are trying to make the business metric fit their technical issues.

what the advertiser wants is: how many people are going to see the ad, and for how long. and what’s the quality of the ad position.

selling ‘premium’ ad inventory is about selling three things:
1) the quality of the volume of ads you have to sell
2) the quality of the content the ad will fit into (and its position in this content)
3) the quality of the brand you have

Selling ads is not as simple as refining the one metric.

Remember up-front that you (as Yahoo!) are measuring web “effectiveness” to maximise advertising revenue (banner-based or Adwords/Overture based). From that perspective, the “views” model isn’t applicable to anything other than diverse portals like Yahoo!, which aspire to be “one-stop shops” (I’m hearing that too often now – indications of a return to portal-ism?). Second, it doesn’t consider the many many different user interfaces to online information today that aren’t browser-based or even web-page-based. What about Yahoo! Widgets? What about the content access through Yahoo! Messenger? Through Yahoo! Desktop Search? Through Yahoo! Toolbar? Third, what about always-on customisable home pages such as the one Google offers? A user could access content all day long and never have to make more than one “visit”. Similarly, a Gmail inbox could be open all day too, with several dozen ads being served. That would count as one “visit” too. And hell, what about content access from Yahoo! Go?

Clearly, measuring the effectiveness of online content is a lot more hairy than you’ve made it out to be, Peter! :)

When the Internet business community began standardizing definitions (which became constructs) circa ‘96, the alignment around “page view” very much imposed a print construct on the concept of web audience.

I always found this problematic.

Advertising is like physics– its all about time and space. In print, it’s space.

Do you think there is any logic to migrating toward the broadcast construct, which is to say moving from space to time? It is not hard at all to envision a day when the concept of a page is anachronistic, yet the alignment of the industry around that concept makes progress away from the construct seem difficult. Maybe its all about time, not space. In broadcast, or let’s just say any time-based media, audience is calculated based on cume persons, and time spent, and expressed in terms of average minute, average quarter hour, average program, or average daypart persons. No one cares about “frames viewed” (if TV is 24 frames per second and I watch an hour, we all just accept that I watched 60 minutes, and no one cares that I was served 86,400 frames.)

It might be easier to shift adspend from TV (still 50 cents on the dollar here in the US) to the Internet if we gave advertisers the time of day, if you see what I mean.

Remember that TV show from the 60s, Its About TIme? Maybe it is about time.

Time may not be the perfect model. But I’ll leave you with these wise words a guy once told me. “All models are wrong. But some of them are useful.”

Whadaya think?

Cheers.

–josh–

Principal
Warp Speed Marketing, Inc.

What drove “visits” to more prominence is the inability of page views to track AJAX and Flash web sites that contain multiple pages within a URL.

I have same concerns with js that if I am using tabbed browser and open multiple pages, my visit is counted as one per session, which makes the loyalty measurement a skewed piece of statistics.

Page visits must be used as an addition to existing metrics (esp for the type of sites I described above), not a replacement to page views.

js