Share Of Attention: Advertising’s Newest Time-Based Metric

marcguldimannddtData-Driven Thinking" is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Marc Guldimann, CEO at Sled Mobile.

Viewability, the concept that advertising should be seen to be paid for, has taken the advertising industry by storm. It's interesting to read the conclusions people draw about viewability: It is a currency, it might be a currency, it's not ready to be a currency or will be a currency once standardized.

Most are directionally correct, but trying to classify viewability as a currency is a mistake. First, as a matter of semantics, a currency is what you transact with once a price for an amount of a good is established.

Here’s an easy way to think about it: When you fill up your tank with 10 gallons of 91 octane for $30, neither the gallons nor the octane are the currency – they are units of measurement of the good you are purchasing. The $30 is the only currency in this example.

It’s the same with the dollars on your media plan. Gallons, the metric of sale, is equivalent to impressions in digital advertising. When people lament about new currencies, they are really referring to new metrics of sale.

The metric of sale’s purpose is to represent the “amount of thing” contained within a single unit of a good. The amount of gasoline in a gallon is extremely consistent so it’s a great metric of sale.

Advertising’s “thing” is more nebulous, but at its core the advertising industry is about buying and selling attention, so it makes sense that the metric of sale for advertising should attempt to measure attention. Today display advertising uses the impression, but it does a poor job of yielding a consistent amount of attention. As a result, it’s no wonder that the industry is rushing to anoint new currencies metrics of sale.

The right question to ask is: “Is viewability the right way to measure attention?” Or, more specifically: “Does viewability give me enough information about how much attention is paid to an ad to transact on it?”

Time In View

The answer is that it’s getting close, but viewability doesn't have the granularity required to be the pricing metric for a transaction in its own right. Instead, a derivative of viewability – time in view – has been cast into the spotlight as the savior of display advertising.

For example, there might be six viewable ads on a slideshow page where you spent three seconds, as well as one viewable ad on an article page where you spent four minutes reading. If today's MRC standard for viewability were the measure of attention, all of those ads would be considered the same. Granted, viewability is an improvement over the vanilla impression metric, which has been widely used since the inception of the banner ad.

There are a couple of products trying to increase the resolution of viewability metrics by selling display ads based on seconds of viewability, also known as time in view. Buys are made with typical adjacent banner sizes in bucketed time increments.

Buying banners on a time-in-view metric solves for the value delta between two ads that were on screen for different amounts of time. It’s also an amazingly useful tool when measuring the impact of media but doesn't address share of attention. There could still be eight ads in view at the same time, not to mention the attention paid to the content that the reader is there to consume in the first place. Without quantifying the amount of attention paid, it’s hard to transact on time in view.

Share Of Attention

Technological means can be used to quantify the amount of attention paid to an ad. Chartbeat, for example, measures attention in large part by tracking movement or interaction via a mouse, according to Gigaom.

There’s also eye tracking. Every phone has a camera that could be used to track eye movements. But sending a message to users along the lines of “The application ‘News App’ would like to access the location of your corneas” is sort of a big ask. Many consumers would object to this kind of default tracking, and the value exchange needed to drive opt-in at scale would be cost prohibitive. But with Google Glass v2 around the corner, you never know.

A simpler, albeit potentially less accurate, way to measure attention is to use interruptive ads. If an ad is interruptive, it captures 100% of attention paid to the medium while it’s in view. Interruption has worked for TV, radio, print, video, social feed and even Snapchat ads.

They aren’t perfect. There’s a very reasonable backlash against ads that break experiences in the name of capturing attention or driving performance at the expense of experience. Some interstitial ads have tiny close buttons, while others split the page and move content around. Others, like most pre-roll ads, force audiences to spend time with ads.

But the best interruptive ads, which I call “politely interruptive,” are built with an understanding of their environment and give readers control of how long they want to spend with the content. The amount of time an individual chooses to spend with a message may offer the granularity and specificity required to be a metric of sale.

Thirteen years ago, the digital direct-response advertising was revolutionized by the creation of a more accurate metric of sale: the click. Since more accurate metrics of sale mean less risk borne by the advertiser, direct-response advertisers have been able to deploy a disproportionate amount of capital digitally since then.

Now, measuring share of attention down to the cost per second could do for digital brand advertisers what CPC did for the direct-response industry: revolutionize it by removing buy-side risk.

Follow Marc Guldimann (@guldi), Sled Mobile (@sledmobile) and AdExchanger (@adexchanger) on Twitter.

Popular On AdExchanger Right Now:


  1. It's inevitable that something like a Cost Per Second model will eventually have widespread adoption. Time in View appears to be something of a holy grail to advertisers -- keeping that ad in view for long periods of time makes a "real impression". It has the advantage, even beyond the obvious one of deep branding, of giving the consumer the time and space to respond to a call to action. We see this in the world of corporate sponsorship. Now it's time for digital brand advertisers to have the power and flexibility of that kind of exposure and the model and metric that goes along with it.

  2. We should be careful talking about mouse tracking as validated means to track what's being looked at. Based on a study done by Google, there is only a 6% correlation between mouse tracking and eye tracking. Since viewability is the potential to be SEEN, eye-tracking is truly the only way to know if you're ads are actually being looked at.

    • Thanks for the comment, Steve.

      You've touched on one of the weaknesses of adjacent advertising.. banner blindness. However, this article contemplates the value of time spent with interruptive ad formats which capture 100% of a readers attention while they are in view.

      • If the article is only referring to interruptive ads or full page take overs where you only see the ad, you will have two problems.
        1. Consumers do not like having to wait for the content they expected to see (for those who play freemium games, the war commercial that you have to watch gets annoying)
        2. Not every advertiser will run full takeovers therefor the so called attention currency is not for all ads, hence not a currency.

  3. Dylan Dharmadasa

    Very presumptuous. Would like to see some publishers and agencies collaborate to demonstrate whether there was actual more efficient distribution of ad dollars by using time in view as a transaction method. Also the tech behind time in view seems to be questionable. Very inconsistent results between tech providers. I prefer MRCs suggested method to transact against viewable impressions. A 10-second in-view creative is not necessarily more valuable than a 1-second in view creative, yet the suggested method would value it 10x higher. Huge scope for error in attribution here.

    Also if you were to transact in time in view metrics, unique audience would then have to be factored in. Does this imply you would transact as time in view per user? How would you then determine the value of an individual impression?

    A lot more work to be done.

    • Thanks for the comment, Dylan.

      I'm not suggesting that there's a linear association between time and value, but there's little question that time spent with a message is correlated with brand lift across metrics - several peer reviewed studies from Yahoo and Microsoft research back this up.

      The big win from buying on time spent is de-risking of ad spend. Also, it supersedes the need for viewability or other quality metrics that might be applied to impressions.

      The time that a reader chooses to spend with an interruptive message is an excellent way to measure the value of an impression. And, to you your point, yes the metric should be applied to users.. just like frequency capping.


Add a comment

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>