“Data Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.
Today’s column is the second installment in a two-part overview of the technology issues associated with ad viewability (Part I is here). It is by Jeremy Stanley, Chief Technology Officer, Collective & Co-Chair, IAB 3MS/Viewable Impressions Implementation & Implications Working Group.
Whether your goals as a marketer are lower funnel and direct response oriented, or upper funnel and brand building oriented, viewability data – when measured accurately and used appropriately – has great potential to improve the performance of digital advertising campaigns. While the Media Rating Council (MRC) has recommended in its Viewable Impression Advisory that it is too early to transact on viewability as a currency throughout the digital ecosystem, that does not mean that advertisers should ignore this powerful capability. Rather, they should be measuring viewability today, with the best technology vendor they can identify, and intelligently using that data to improve their return on advertising spend.
In part 1 of this series (A Viewability Technology Primer, Part 1: Promises & Pitfalls) I covered the six different technologies in market today for measuring viewability: page geometry, panel, behavioral proxy, browser exploits, publisher API, and browser monitoring. We concluded that there is significant variability in the quality and completeness of the viewability data provided by these different approaches.
Given these technological challenges, what is a marketer to do?
First, you must choose your vendor wisely. Begin by being a healthy skeptic – there is a tremendous amount of hype and marketing ‘noise’ being produced by some technology vendors in the race to secure viewability measurement contracts. The most sophisticated vendors are intimately familiar with each of the above approaches, and can provide compelling demonstrations of their technology in real-world circumstances.
Dig deep with them to understand exactly how they are measuring viewability. Ask the tough technology and implementation questions such as:
- How the technology handles nested cross-domain iframes
- How often they can directly observe viewability by browser and device
- How they handle impressions where they cannot directly observe viewability
Finally, just as the MRC recommended in its recent advisory, demand transparency in the data reported – identifying not only how often ads are being viewed, but how often the viewability data is being directly collected versus inferred or estimated. Be sure to request ‘cuts’ of these statistics that show variation across individual publishers and browsers to validate what the vendor has communicated about limitations in their implementation.
Once you have selected vendor, and are being supplied with transparent and high quality (if not necessarily perfect) viewability data, the next step is to incorporate this data into your performance measurement strategy.
I firmly believe viewability is a powerful tool for both brand and direct response advertisers, but the most appropriate use depends greatly on the broader objectives of the advertiser.
You are a brand or direct response advertiser measuring click through or engagement rate:
While click-through rate (CTR) has many faults, and should never be the sole basis of measurement for a campaign, it can provide meaningful insight into how successful a creative is. Engagement rate (ER) tracking, either by identifying when a mouse is hovering over an advertisement or tracking specific deliberate user interactions with a creative can be even more informative.
Viewability data can significantly enhance both of these metrics. Consider evaluating vCTR or vER performance (the leading “v” denotes viewability), wherein the number of clicks or engagements are divided by just the number of impressions that were viewed. These are more meaningful measures of how often the user is interacting with your creative, and thus provide a more stable, consistent and interpretable view of creative performance over time and across sources of media.
You are a direct response advertiser measuring attributed conversions:
Attribution is a complex topic, fraught with debate about causality, complexity and cross-media dynamics. Fortunately, whatever your approach is to attribution, viewability can play a powerful role.
Initially, you can measure the cost per view (CPV) rate in addition to the cost per action (CPA) rate for your campaigns. Media partners who are serving advertisements on very inexpensive exchange-bought impressions in order to maximize their chance of receiving last-impression attribution credit can be penalized significantly for also delivering very high CPV rates.
Ultimately, the viewability data should be directly integrated into your attribution model, such that credit for conversions is only given to impressions that were viewed – as other impressions cannot have had any impact on a user.
You are a brand advertiser measuring targeted reach and frequency:
In today’s world of abundant supply and interconnected platforms, meeting your reach and frequency goals for a target audience is easier than ever. But absent a more concrete goal, brand advertisers often seek to continually reduce their cost-per-thousand (CPM) rates. This can inadvertently steer a campaign towards remnant inventory with very low viewability rates. Combat this directly by incorporating viewability rates as a performance objective or seeking to minimize your CPV rather than your CPM.
Further, if viewability is integrated into an ad serving or real-time bidding platform, then frequency can be managed on a viewable impression basis rather than a served impression basis. This prevents a brand advertiser from missing important opportunities to influence their target audience simply because they have met their frequency cap with served (but not necessarily viewed) impressions for a given user.
You are a brand advertiser measuring awareness, favorability or intent lift:
Brand advertisers who are seeking to directly influence the attitudes and ultimately behaviors of consumers may have the most to gain from viewability. Ultimately, the longer a user is exposed to a creative, the greater the chance that it will have a meaningful impact.
These brands should optimize directly to both viewability rates and dwell times. Be sure to cap the dwell time at a reasonable duration (e.g., 2 minutes) to prevent ‘run-away views,’ where a user has left the device, from unduly influencing the performance measure.
The message I want to leave you with is that viewability, if accurately measured, holds enormous promise no matter what the objectives of the advertiser are. The greatest value will be achieved when the data is available at an impression level and integrated directly into other measurement systems (such as attribution and ad-serving platforms). Only then can the most meaningful metrics that inherently account for viewability be computed and optimized against.
Advertisers must also not forget that while viewability is an important component, it should never be the sole goal of a campaign. Reaching the right audience, doing so in a brand safe environment, and serving a compelling and personalized creative will continue to be strong value drivers.
Given the fundamental technology challenges we still face and the crowded and confusing vendor landscape, I imagine that many advertisers will continue to tread lightly here. However, the technology holds great promise. Those who can identify and integrate the best measurement technology and then adapt how they execute and evaluate their campaigns will reap great rewards.