Tuesday, May 19, 2015

Is It Time To Pay Attention To Attention-Based Models?

Yesterday, The Financial Times announced a new digital advertising metric – ‘cost per hour’ (CPH).  They devised the model with Chartbeat.  The result?  Increasing marketing effectiveness by measuring not just whether an ad is seen or not, but for how long.

Currently, this new metric is available only for digital display ads, not video ads.  That will soon change.  After all, video ads supply even more data – when someone actually starts watching the ad itself.  And, when they stop watching.

While everyone has always assumed that the more time spent with a piece of advertising, the better the brand recall, The Financial Times now gives us actual numbers. 

According to a white paper written by Nikul Sanghvi, an analytics consultant on the project, when comparing ads seen for five seconds or more with ads seen for less than five seconds, all metrics showed significant improvement:   Ad Recall (+79%), Brand Familiarity (+55%), Brand Association (+51%), and Brand Consideration (+58%).

What’s more, these results were shown to increase the longer that the ad was seen.

Would this translate equally well to video advertising?

The assumption would probably be yes.  But there is something even more relevant about time spent with video advertising than display advertising. 

Unlike display advertising, video advertising runs on the dimension of time – :15, :30, :60.

 Which means if we can interest people in watching, a thirty-second ad would be more effective than a fifteen-second ad and a sixty-second ad would be more effective than a thirty-second ad.

We can now measure peoples’ interest in video ads through time spent with the ad.

So why aren’t they sold that way?

The Financial Times sells attention for display ads on a cost per hour basis.  Could time spent with video ads also be sold on a cost per hour basis?


How to do that?

We'll leave that for another post.

No comments:

Post a Comment