Thursday, September 27, 2012

Attribution? Or, Arbitrary?

There’s a lot of talk about attribution these days.  Most of the talk centers around whether an online ad can still be effective if no one clicks on it?

Those who say “yes” usually have a horse in the race.  Publishers, media agencies, etc.

Those who say “no” are usually pompous pundits who write blogs like this one.

Dax Hamman though, recently had an interesting viewpoint.  Mr Hamman is chief revenue officer at Chango.

He mentions that the major complaint about view-throughs is that you just never know.  Just because the cookies say that the person who went to the website was exposed to the ad sometime in the recent past, doesn’t mean the ad influenced his decision to go the site.

Mr Hamman also mentions that view-through windows – the length of time between the appearance of the ad and the visit to the site – is adjustable.  Typically, it’s 30 days.  But, it can be as short as 24 hours.

When the view-through window is 30 days, view-throughs can account for over 90% of site visitors and over 90% of page views once a view-through user arrives.

That’s a big number, 90%.  It’s almost 100%.

What 90% means is that only 10% of those who go to a brand’s website do so because someone told them about the brand, or they saw an interesting commercial about it the night before.  You’ve probably never done that.  No, neither have I.

Let’s also keep in mind that according to comScore’s recent study of 12 big brands advertising online, 31% of 1.7 billion impressions sampled were delivered, but never viewed.  This is because an impression is counted when it leaves the server.  Not when it is in front of the user.  In fact, the impression doesn’t even have to be above the fold to count.

In other words, impossible to see counts as highly as seen.

So, what do we have?  Only 69% of ads served are even possible to be seen by the viewer.  And yet, 90% of site visitors can be attributed to a view-through?

Really?

Mr Hamman’s argument would have been far more compelling with a smaller number.  When anyone claims that anything is close to 100% responsible for something occurring, I’m usually 100% sure of only one thing.

He's wrong.


No comments:

Post a Comment