Viewability

By Scott Knoll is the CEO of Integral Ad Science

In the past few years, the advertising industry has made tremendous strides in measuring viewability. But buying viewable ads doesn’t always translate to positive results. As a matter of fact, it often translates into worse results.

As true consumer-based marketing becomes a reality, understanding how consumers are exposed to ads over the course of a campaign — and then optimizing that exposure time — has never been more critical. But where are the tools to fix the problem?

A Short History of Viewability

In the early days of digital advertising, the industry relied on a familiar frame of reference to measure results — the same framework that TV, print, and direct response advertising had been using for more than a century. This CPM-based model looks at circulation, pass- and read-through rates, and newsstand sales to calculate impressions — all audited by industry-supported, nonprofit organizations.

Digital advertising borrowed the CPM concept and began trading based on the number of ads served. Jump to today, where viewability has become the industry’s currency, with the Media Rating Council (MRC) establishing minimum standards for measuring viewable impressions for desktop, mobile, web, and in-app advertising.

Here’s what those transaction standards look like today: In video 50 percent of the ad must be in view for at least two continuous seconds, in display advertising 50 percent must be viewable for at least one second, and in large format at least 30 percent of the ad must be in view for at least one second. It’s critical to have one transaction standard for the entire digital advertising ecosystem. Unfortunately, in pursuit of this noble goal, the industry may have caused unintentional harm; the industry transaction standard has somehow overruled good common sense.

The MRC standards were never intended to be a performance threshold for marketers, but, for lack of a better option, they’ve become just that. The marketing industry needs to divorce its media trading standards from its marketing objectives. Just because the transaction standard is centered on the minimum “opportunity to see” does not mean brands should optimize their campaigns toward it. Quite the contrary.

A System in Need

“If I had asked people what they wanted, they would have said faster horses,” Henry Ford once quipped. That sentiment accurately describes the marketing industry, which often finds itself on the bleeding edge of innovation. It is hard for marketers to articulate exactly what they want when what they need has never existed.

Recent research from Integral Ad Science (IAS) shows that about half of the ads being purchased today have absolutely no value. This staggering statistic is not the result of half the ads performing and the other half missing the mark. Instead, it’s because more than half the ads purchased never even had a chance to work.

Publishers and advertisers alike are well aware of the current system’s shortcomings. According to The New York Times CEO Mark Thompson, “We’ve produced an environment which is very good at rapidly matching ads with people. It’s not very good at any kind of discrimination of how that happens, and what the actual effect is on the end user. It’s based on a very impoverished view about how advertising works: Literally, the assumption is if you can hit a consumer with a message the job’s done.”

But hitting the consumer with a message is only part of the equation. Consider the following scenario: A brand has struck a deal with its favorite publisher. Its audience is the most on-target audience the brand could ever imagine. In the deal, the brand has agreed to transact on the MRC standard for viewable impressions (as it should).Any marketers worth their salt understand that one second rarely leads to audience attention, let alone sales.

A few days into the campaign, the brand decides to take a look at some of the data for two consumers. Interestingly, both consumers had the same number of viewable exposures (for this scenario that number is six). Consumer A made a purchase. Consumer B did not.

Both consumers are the same age, have the same income level, exhibit the same behavioral characteristics, and live in the same general geographic area.

There are many things that may have caused Consumer B to not buy the brand’s amazing products, but is the brand seeing the fullest picture? Spoiler: It is not.

What the brand doesn’t know is that while Consumer A has an average in-view impression length of 4.7 seconds, the average in-view length for Consumer B is only 1.3 seconds. While both consumers had the same amount of viewable impressions, they had dramatically different experiences with the brand messages. The total exposure time for Consumer A was 28.2 seconds (the average in-view impression length multiplied by the number of viewable exposures); for Consumer B it was 7.8 seconds. Consumer A had three and a half times as many opportunities to see the brand’s creative as Consumer B, yet the brand paid the same amount to reach both.

Is that why there was a difference in performance? Probably.

Now, reaching a consumer and landing a message effectively is the challenge and the reason why viewability is so critical. Keeping the MRC standards in mind, when marketers say they have a 70 percent viewability goal (that is, they want 70 percent of their impressions to be viewable), what they’re actually saying is that it’s acceptable for 70 percent of their impressions to be in view for one second. But remember, it takes a second just to load the ad; so the reality is that the full ad may never be viewed. Even if seen, any marketers worth their salt understand that one second rarely leads to audience attention, let alone sales.
 
When campaigns are optimized for viewability, the next natural step is to quantify the return on a specific campaign. In the early days of digital, marketers combatted cookie-bombing. Today’s advertisers must deal with “one-second-impression-bombing” in the quest to claim credit when it comes time to determine what drove the conversion.

The one-second standard may also have created a marketplace where advertisers end up buying more one-second impressions. If marketers were to optimize to a one-second standard, they would get a large quantity of one-second ads. Yes, they just meet the standard — however, what ends up happening is advertisers get poorer performance as a result.

Data in a joint study from IAS and IPG Media Lab indicates that 70 to 85 percent of people are exposed to only one viewable ad impression and, of those people, 50 to 65 percent have the potential to see the ad for five seconds or less on a typical campaign. This data establishes a need to understand and optimize against consumer ad exposure.

For the time being, trading on the MRC standard is adequate because the industry needs a common language. However, according to an IAS survey of more than 1,000 digital advertising professionals, just 51 percent feel the viewability standard is sufficient. And if marketers hope to get ahead of the curve, they can — and should — optimize toward impressions that are in view long enough to be consumed and acted on.

Is Seeing Believing?

So, what makes a successful consumer impression? Ideally, it’s how long that consumer sees an ad cumulatively and across all formats — video, display, mobile, etc. IAS research concludes that the more exposure a consumer gets (up to a certain point), the more likely they are to perform the action that the advertiser was hoping they would perform. Whether they went online to make a purchase, recalled the brand, or increased purchase intent, exposure time is the No. 1 driver in getting that consumer to take an action.

In nearly every case IAS studied, a “barbell effect” — with dramatic under-exposure and dramatic over-exposure in typical campaign media buys — was the norm.
 

All brands want to buy viewable inventory because they want to ensure that the consumers they’re targeting adequately see their messages. Empirical evidence suggests that the longer a message is in front of a consumer, the better chance it has of working.

The advertising industry is currently operating on a lackluster concept, simply due to the tools available.

The next frontier for digital advertising will be to increase and improve this opportunity, shedding the technology limitations.
 
There’s More to Be Done

A new way of thinking is emerging, where viewability is no longer a binary metric, but also includes how long and how often a consumer has seen an ad in total — think of reach and frequency as the guideposts. In this context, reach can’t be based on a number that isn’t meaningful, like an ad served. Nor can it be based on whether it was in view. Reach must be based on time, accounting for both exposure time and frequency.

The technology to understand how long a digital ad has been in front of someone clearly exists, so why does the industry treat digital like traditional print advertising, where it never had the technology? And what will it take to move past this misperception?

There are two major components: one is the measurement capability, to quantify the cumulative amount of time that each person has been exposed to the brand message. The second is the ability to optimize for the desired exposure time or exposure level. Optimization is initially happening on the programmatic side, but will then impact the entire ecosystem.

With this new way of thinking and measuring, marketers will know how much exposure time an individual has had to date, before bidding on an impression. Based on that data, they can then decide whether to bid on that consumer, and how much to bid.

Equipped with an understanding of how long an ad is in view on a single impression, marketers can tie the data to specific consumers and understand how, where, and how often (that is, how long in total) they’ve been exposed to the message. This critical and predictive data answers the question brands really want answered: Have consumers seen the ad frequently enough and long enough to take the action we wanted?

And once brands have access to data this powerful, how could any reasonable CMO not want to use it?

 

Skip to content