Do you watch entire YouTube video ads without skipping to the end?
The answers to these questions revolve around engagement -- something important to advertisers and marketers, but also difficult to measure.
Engagement is at the center of a new movement in digital advertising to buy and sell online ad inventory where time is the measurement of value and price.
This model is called time-spent and is measured in terms of cost-per-second (CPS) or cost-per-hour (CPH).
How does time-spent work?
Online ads have been traditionally bought and sold the same way as in traditional print media and often still use the cost-per-thousand (CPM) metric.
As ad inventory volume increased, advertisers looked to gain more control over their spend and the measurable value of their media buys. This led to the “viewable impression.”
Another approach is to use a “guaranteed time slot” (usually in increments of 20 seconds) to make sure that an ad is actually viewed.
Some might say that one more set of numbers doesn’t make much difference. But there are two very important reasons why the time-spent method is important:
1. Engagement. Buying and selling content according to the number of impressions served would seem to make measuring ROI easier.
But therein lies the problem: It’s too easy.
Treating every impression the same (as CPM does), regardless of the advertiser and ad, doesn’t address how consumers actually engage with content. The viewable impression metric isn’t much better.
The combined result of volume and programmatic advertising technology is that consumers are completely inundated with ads. How do they react? By tuning out or installing ad blockers.
2. Fraud. The second impetus for adopting a time-spent metric is the ongoing battle against ad fraud. Some estimates predict that advertisers will lose over $7 billion to fraud this year. Time spent as a stand-alone metric is a good indicator that a visitor is human and not a machine.
Not Flawless
That isn’t to say that time-spent is a perfect solution.
A true measurement of “engagement” with online ads must go beyond the amount of time an advertisement is visible to an online user.
The cost-per-second or cost-per-hour metric isn’t ideal for every type of ad. Measuring the percentage of a video viewed, or how long an expanding ad is open provides a better calculation of engagement than viewing time for a static display ad.
Also, this model tends to favor publishers, especially those with premium content that consumers view for longer periods of time. While this might be good for brand awareness campaigns, it doesn’t necessarily address conversions.
Could time-spent metrics change ad tech?
Given the general ineffectiveness of the CPM model and the growing pressure to put customer engagement and experience first, time-spent could shore up ad tech in several ways:
1. Reduced inventory: Following the guaranteed time slot model, ads will have to be viewed for longer than the viewable impression standard of one second. This could seem bad for publishers, but reducing supply and emphasizing value could wind up benefiting them.
2. More emphasis on content quality: When ads are sold based on time viewed, the issue of placement and context becomes even more important. Ads placed alongside high-quality, long-form text and premium video will be in higher demand.
3. More complex analytics and programmatic solutions: When it comes to measuring effectiveness, advertisers need better tools such as new audience segmentation features in DMPs, ad servers that match content with contextually relevant ads, and new processes for RTB ad exchanges.
4. New approaches to digital marketing: As advertisers adjust to the shift, they may find themselves rethinking their campaigns in general. If potential customers are exposed to ads for a guaranteed length of time — which would result in a significant increase in brand awareness — advertisers might need to rethink their whole conversion process.
While there’s not likely to be a overnight revolution in ad tech because of CPH and guaranteed time slot media buying and selling, it’s quite possible that more publishers will choose to use this model, forcing advertisers to adjust — but also providing them with many advantages.
So if the time spent "metric" for one ad "exposure" averages 15.867 seconds while the corresponding figure for the same ad on another "platform" 1s 5.344 seconds does that mean that the first one was almost three times more "effective" or valuable and should the former seller be able to charge at least three times the ad rate of the latter? Or, putting it another way, how do you interpret time spent in a meaningfull manner? Don't you need other information--- even data derived from humans like ad recall or motivational response---as well as other electronic indicators? If the first ad in my example, above, generated only 10% higher recall than the second ad, isn't this---providing the research is sound-----a more important barometer of "engagement"?
Reduced inventory, provable viewability, time-spent, content emphasis....golly, it's sounding more like TeeVee everyday
Great piece, Maciej. A new currency is needed, you bring up some valid points that would indeed put digital more in-line with TV (assuming the commentary above is focused on where the puck is headed, which is video). Advanced multi-touch marketing attribution vendors which can address bots/fraud will be in high demand.
This boulder has been pushed up this hill for longer than anyone wants to admit, nnd will keep getting pushed up the same hill until screens are smart/interactive/creepy enough to gauge eyeball tracking on a universal level...
I've wrestled with this one for years.
Take a high circulation/high readership fashion magzine like Vanity Fair with over a million copies. It may be read for (say) 5 hours in that month. So that's 300 minutes across 30 days. The average is 10 minutes per day. With pass-on, the million copies might get (say) 8 million readers.
In TV-speak the average minute-audience would be little more than 50,000. Mind you, for a prestige brand that would be the best 50k you could get!
So while I am a strong believer in time as the most sensible common metric, it is how comparisons are made across and between media. There would need to be masses of re-education done.