Integral Ad Science’s UK managing director Niall Hogan says discrepancies around viewability measurement are holding back discussions on media quality.
Over the first half of this year there was a definite shift in attitude from UK advertisers and agencies who are doing far more to understand media quality concerns. Today, the digital industry as a whole is placing much greater emphasis on viewability (whether ad impressions are in or out of view) as a key indicator of campaign performance.
In the US, the ad industry appears to be a step ahead in how it is approaching viewability; 2015 has seen a growing number of advertisers paying only for viewable impressions and getting closer to using viewability as a trading currency. There have even been moves by global advertisers such as Shell and Unilever to define their own standards for viewability to use as the basis for their digital transactions.
For certain advertisers, viewability targets have assumed almost equal importance with traditional indicators of performance. It is clear that globally the buy-side is gearing up to put viewable impressions at the centre of future trading deals. In the UK, however, we still seem some way off using viewability as a trading currency. So what is holding us back?
The main obstacle is discrepancies between different vendors’ viewability reporting. A few years ago, viewability measurement technology vendors were bringing varying measurement methodologies to market and this created wildly different results in viewability reporting from one vendor to another.
Some viewability vendors only employ one measurement method – using the position and size of an ad relative to the screen. Others will combine this with analysing the speed of ad processing when on and off the screen in order to get greater measurement rates.
Today, one of the main reasons for discrepancies is the fact that buy- and sell-side technologies are usually implemented in different ways. Both are measuring and analysing different ad elements, deciding when to start the viewability timing and where the ad is on the page to determine if it’s in view.
Publishers’ measurement pixels are found within the ad container, and activated as soon as the container is available on a loaded web page. This pixel determines where an ad is on the page and if it is viewable. Buy-side measurement pixels are mostly fired from the ad tag when the ad creative loads. This means the count for an ad appearing in view starts at slightly different times. We’ve all experienced ads loading later than the rest of a page’s content. With a difference of just 200 milliseconds set against the Media Ratings Council (MRC) standard of a viewable impression (that an ad is considered in view within one second) that’s a 20 per cent difference. Since the majority of users typically scroll down soon after arriving, a 200 millisecond margin of error can be significant.
With the buy- and sell-side starting the stopwatch at different times, the discrepancy problem is further exacerbated by the fact some vendors do not account for, or remove, fraudulent impressions from their reporting. Even a very basic fraud product will use a blacklist to identify illegal bots and block impressions only once they are seen.
Sophisticated technology should identify fraudulent impressions and remove them from the count to give a more accurate measurement. With fraudulent impressions removed, advertiser and publisher would have reassurance that reporting is more aligned and that any future campaigns based on viewability would be billed fairly.
Subsequently, with the debate that advertisers should only pay for measureable ads in view, the transaction is compromised if the numbers between advertiser and publisher don’t add up. Again, we may be looking at hundredths of a millisecond, but the result is unmeasured ads and discrepancies, and a billing headache for the finance team.
Resolving discrepancies between technology vendors has been the key focus for industry organisation bodies over the past six to nine months. Thanks to the work by verification and certification bodies such as the MRC in the US and the Audit Bureau of Circulations (ABC) in the UK, standards have been created and viewability technologies audited and aligned so that today we are mostly seeing only up to two per cent variance.
Both the MRC and ABC have built audit and accreditation programs with input from all sides of the ad industry. The ABC audit is extensive and consists of three core stages, which are:
1. A detailed ‘declaration on methodology’ is completed, giving the auditor insight into the technology and processes used in measuring viewability. This can involve both written submissions and face-to-face interviews.
2. A series of automated tests are then built to test agreed principles. These tests represent all possible scenarios and outcomes that the viewability technology should conform to in terms of functionality.
3. Raw data logs are provided that show the audit firm un-aggregated impression data and is used to validate and confirm test outcomes.
There is a lot of work being done to ensure discrepancies are minimised and to increase trust in effective measurement of media quality. The MRC recently published guidance on the main reasons why discrepancy was occurring and the ABC has refined the testing principles for its Viewability Verification Accreditation in order for UK vendors to overcome agreed reasons for discrepancy.
To many in the digital advertising space, discrepancies are an unavoidable cost of doing business and likely aren’t going anywhere. But the reality is that they need to. When numbers don’t match it creates a level of mistrust between advertisers and publishers. These discrepancies lead to a power struggle between the buyer and seller and our industry runs best when both sides trust that advertising campaigns are accurately executed.
The need for trust is why engineers at Integral Ad Science are working to eliminate discrepancies between publisher and agency viewability numbers by employing the same underlying measurement methodology. The dual-sided approach is more accurate than only looking at one half of the equation: from the publisher side it determines where the ad is, and from the advertiser side when to start the clock. The work being done ensures publishers have access to the same data as their buy-side partners so that discrepancies and reconciliation can be mostly eradicated, and trust in measuring media quality can be reintroduced.
A world in which discrepancies do not exist might seem like a pipe dream, but the industry needs to prioritise making it happen – whether that’s through technology or industry standards. If so, we’ll achieve a much higher level of trust and efficiency, laying the groundwork for better partnerships and, ultimately, a better media quality ecosystem.