Break it Down: How to Effectively Evaluate Media Quality and Measurement Solutions

In an earlier post, we introduced the concept of DV’s Authentic Impression® – a proprietary, MRC-accredited metric that assures both media quality and media effectiveness.

To be counted as Authentic, an ad impression must satisfy all of the following criteria:

• Fully viewed
• By a human
• In a brand safe environment
• Within the correct geography

DV is the only company accredited to measure and de-duplicate impression quality in this manner – enabling brands to use the Authentic Impression as a definitive measure of media quality and effectiveness.

But as they say, the whole is only as good as the sum of its parts. What does it take to authenticate each of the components of media quality – and what should advertisers be looking for when they evaluate measurement solutions?

Viewability

“Does the ad have the opportunity to be seen?”

Source: DV 2019 Global Insights Report

Industry standards for viewability (50% of a display ad must be in view on a screen for one second – two seconds for a video ad – before an advertiser is charged) have been established by the Media Ratings Council (MRC) and thresholds for how much of a campaign should be viewable have been set by the IAB (70%). While viewability rates on advertiser campaigns are on the rise, there is still much room for improvement. The OM SDK (Open Measurement Software Development Kit), launched in 2018, has helped facilitate viewability and verification measurement on mobile apps, but to grow and scale, the industry needs to continue to work together to create consistent standards across devices and formats.

Evaluating Effective Viewability Solutions – Key Questions:

• What technology is used to “see through” to the measured environment?
• Is the solution applicable across devices and formats?
• Does the viewability measurement provider utilize the OM SDK?
• What other measures – in addition to viewability – are offered to gauge engagement and / or exposure?
• Is it possible to employ custom thresholds?

Fraud 

“Is the ad served to a real human, and not a bot?

$5.8 – $19B estimated loss to fraud in 2019.
Source: $5.8B estimate from ANA, $19B estimate from Juniper Research

There are different drivers and types of fraud and SIVT (Sophisticated Invalid Traffic) in different environments. The majority (55%) of mobile app fraud is classified as app fraud, while on CTV/OTT most cases are bot fraud (86%). Fraud targets the money: premium inventory and emerging channels. In the last 12 months, DV has discovered that the number of CTV and mobile apps classified as fraudulent has more than doubled.

Evaluating Effective Fraud / SIVT Solutions – Key Questions:

• Are all types of fraud / SIVT and their variants measured?
• Does the service cover all device types and work across all formats?
• Is fraud detected before (avoidance) and after (blocking) the bid?
• What is the methodology for detecting fraud?
• What kind of AI and machine learning models are used for detection?
• Are fraudulent impressions accurately ID’ed to ensure scale isn’t harmed as a result of “false-positives”?
• Is the service fast enough to ensure fraudulent impressions are avoided / blocked?
• What fraud schemes has this provider recently detected?

Brand Safety

“Is the content that the ad is served against aligned with the brand’s values and within the right geography?”

87% of consumers feel that brands bear responsibility for ensuring their ads run adjacent to content that is safe.
Source: DV / Harris Poll (June 2019)

Delivering ads adjacent to risky content is a strong threat to digital ad budgets in 2019. Brand safety violations have a material, adverse impact on consumer sentiment – with real commercial implications. The number of brand safety incidents are also on a steep rise, especially on newer and emerging channels – compared to the year before, we recorded a 194% increase in mobile app brand safety incident rate in 2019 (DV 2019 Global Insights Report).

What is “safe” or suitable context may vary widely from brand to brand. A travel agency would typically wish content on travel mishaps and accidents to be avoided for their ads; a travel insurance provider may find the same content suitable and even target it. But making this distinction given the sheer volume of content on the Internet requires a deeper understanding of the context in each case.

Evaluating Effective Brand Safety Solutions – Key Questions:

• How many and what types of categories are available for protection?
• What languages are covered?
• Is it possible to set white lists and blocklists?
• How can fake or inflammatory content can be avoided?
• What classification methodology utilized? Is it just keyword protection?
• Is there protection against breaking news that has not yet been analyzed or classified?
• What is the protection coverage across devices, including mobile apps?
• Are all buying platforms covered, including social?
• Are you able to customize the brand safety controls based on issues specific to your brand?

Higher standards drive effectiveness

Many brands have solutions in place to measure digital advertising effectiveness. For example, some platforms offer fraud-free guarantees or segments that target properties that are historically viewable. However, these solutions only address one aspect of quality.

We believe it’s time to demand more – and that is what we advocate with our Authentic Impression. When every impression can verified as fully viewed, by a real person, in a brand-safe environment, in the correct geography, it will benefit not only the brands that pay for the impression but all parties that support the online advertising ecosystem.

In the coming posts we’ll dive deeper into brand safety, understanding more of the advanced methodologies involved in protecting brands and detecting suitable content and context for advertising. Stay tuned!