comScore released its vCE Study Charter this week. The study focused on ad delivery based on key criteria, including whether or not the ads were delivered in-view, to the right audience, in the right geography, in brand safe environments and absent of fraud. If you have time to read it all, you can download it in full here. Too busy closing business? Or attending the latest industry event? ExchangeWire has summarised the juicy bits from the report below.

According to the Interactive Advertising Bureau (IAB), display-related advertising spending in the United States reached $10bn in 2010 and has grown at 20%+ rate since then, far exceeding the growth of traditional media. Until now, digital advertising measurement has not kept pace with the complexity of these changes.

Addressing this industry-wide call-to-action, the IAB, the American Association of Advertising Agencies (4As) and the Association of National Advertisers (ANA) jointly launched an initiative called, Making Measurement Make Sense (3MS). 3MS has published guidelines and is conducting research to help address issues surrounding ad delivery, measurement and validation.

Validating Campaign Measurement

In January 2012, comScore released a breakthrough innovation to the marketplace that addresses many of the guidelines outlined in the 3MS initiative as well as some additional industry issues relating specifically to ad delivery validation. This solution, validated Campaign Essentials™ (vCE™), provides an unduplicated accounting of impressions delivered across a variety of dimensions.

The vCE Charter Study

To better understand the issues associated with display ad delivery and validation, and to test-drive vCE, twelve leading marketers participated in a U.S.-based charter study, called the vCE Charter Study. The eye-opening findings help to pave the way for a more accurate measure of campaign delivery that relies on validated impressions.

It should also be noted that because vCE Charter Study participants included major branded advertisers, who inherently buy more premium inventory than the average online marketer, the study findings are not necessarily representative of the overall online advertising market. In fact, because these advertisers generally engage in high-end, premium campaigns, the findings may represent “best-case scenarios”, rather than the norm.

In-view by Campaign & Site

Across all campaigns in the vCE Charter Study, the average in-view rate was 69%. This indicates that, on average, 3 out of 10 ads were not seen, and were therefore wasted.

For one site, only 7% of the delivered ads were in-view, meaning 93% of all ads served on that site never had the opportunity to be seen and were therefore completely wasted.

Since only 1 out of every 14 ads on the site had the opportunity to be seen, if a marketer paid $1.00 CPM to deliver advertising on that site, the effective CPM would have been $14.00. While a site with a $1.00 CPM may seem like a bargain, when waste levels on the site are as high as 93%, it can effectively become one of the most expensive placements in a media plan.

In-view by Placement

Even within a given site, in-view rates can vary significantly by placement.

The placements appeared to fall into three distinct levels of in-view:
The largest number of placements delivered more than 80% of the advertisements in-view—well above the vCE Charter Study average of 69%. Such placements could be considered high-viewability inventory.
Approximately one-third of the placements delivered advertisements between 66% to 75% in-view, which indicates they were on-par with the vCE Charter Study average.

A small number of placements, however, dragged down the site’s average, given their very low in-view rates.

In-view by Relative Size of Site

An important question relating to viewability is how in-view rates vary based on the size of a site. Using comScore Media Metrix® rankings within specific content categories as a proxy for site size, average in-view rates were calculated based on Top 50, Top 100, Top 500 and long-tail sites. The difference in in-view rates between Top 50 sites versus the longtail sites was a full 16-percentage points.

In-view by Content Type

In-view rates also showed variation by content type. For example, Coupon sites delivered relatively strong in-view rates (89%), whereas Pet sites (27%) struggled, delivering slightly more than a quarter of ads in-view. This variation across categories might, in part, reflect the common layouts among sites in a similar genre.

In-view by Ad Size

The most common ad size used in the vCE Charter Study was the Classic Leaderboard (728x90), followed by the Medium Rectangle (300x250), and then the Wide Skyscraper (160x600). The Classic Leaderboard delivered the strongest in-view rates (74%), but there was significant variance across all sites with a range of 7% to 93% using this size. The Medium Rectangle format (300x250) delivered 69% of its ads in-view, and the Wide Skyscraper (160x600) delivered the lowest portion of ads in-view (66%).

In-view by Position on Page

When discussing viewability, there is a common misperception that ads delivered ‘above-the-fold’ are seen, while ads delivered ‘below-the-fold’ are not.

Surprisingly, the findings demonstrate that some ads delivered ‘above-the-fold’ were not seen because users quickly scrolled past them before the ad had a chance to load, and many ads placed ‘below-the-fold’ delivered a high opportunity to be seen.

The implications of these findings are far-reaching. Publishers, for example, should monetise all ads on their site that deliver an opportunity to be seen, regardless of where the ad is placed on the site. This might mean that inventory ‘below-the-fold’ can be priced as premium as long as the publisher can prove it was viewed.

In-view & Cost

Finally, comScore explored the relationship between the cost of the ad and the in-view rate. The analysis showed there is virtually no correlation between the CPM paid for the ad and whether it was in-view (correlation coefficient = 0.19). Understanding the actual delivery by both site and placement is critical for marketers seeking to value media based on its ability to reach a real user.

Audience

Defining Target Audience

The comScore vCE Charter Study evaluated audience delivery within both traditional demographics and behavioural segments. Behavioral segments are comprised of the heaviest consumers (top 50%) of topic-specific web content, while age-range is a common example of a traditional demographic.

Audience Targeting by Traditional Demographics

Campaigns with a target audience that included one demographic variable delivered impressions to the target an average of 70% of the time. In cases where there were two variables, the accuracy of targeting decreased to an average of 48%, and with three variables, the average was 11%.

Audience Targeting by Behavioural Attributes

Across all campaigns, the average campaign reached its behavioral audience target 36% of the time, with a wide range from 23% to 67%.

It is also important to note that, in some campaigns, the behavioral attribute target actually did a much better job at delivering on-target impressions than the demographic group, suggesting that using demographics alone to evaluate the success of campaign delivery is not sufficient.

Audience Targeting & Cost

Using available CPM data, the correlation between CPMs and the accuracy of demographic targeting (primary audience only) was analyzed as part of this research. The findings revealed a very small correlation (correlation coefficient = 0.18), suggesting that there is little or no relationship between the amount paid for an ad and its ability to reach the desired demographic target audience.

Another very real issue is the accuracy of cookie-based targeting data. Unless cookie-based audiences are verified against a credible third-party source, it is possible that they are missing the mark. In the vCE Charter Study, demographically cookie-targeted ad placements reach their desired demographic 14% to 96% of the time. This indicates a wide variation on the quality of demographic cookie data.

Geography

Defining Geographic Targets

Controlling geographic distribution of advertising on the internet can be challenging. For marketers trying to maximise every dollar of their advertising budget, it is critical their ads are delivered in the desired market where their products are actually sold.

Geographic Targeting: Overall & by Campaign

All campaigns in the vCE Charter Study had a geographic target of the U.S., and in total, about 4% of impressions were delivered outside of the U.S.

Delivery of ads outside a given geographic target often occurs for two primary reasons:
The first reason is simple communication error, when the requirement does not appear on the insertion order (IO), which authorises the purchase of impressions from the site and determines the characteristics of the ads to be served. Such misfires can be easily remedied by ensuring geographic requirements are a standard part of IO contract templates.

The second reason is due to human error. To target an ad to a given geography, the requirement must be programmed in the ad server that is delivering the ad. Occasionally this step is missed by the publisher, or in rare cases, the wrong geography is inadvertently selected.

Brand Safety

Defining Brand Safety

When brands spend money on advertising, they need assurance that their ads will not run next to content that is at odds with the brand they are trying to build or the equity they have already established.

Brand Safety on Adult-Content & Hate Sites

vCE Charter Study used a standard definition of “objectionable content”, based on historical data of sites/categories most commonly identified as being ‘not brand safe’ by leading advertisers. The measurement was applied to all campaigns.

Categories Deemed “Not Brand Safe” for Purposes of vCE Charter Study were:

- Piracy and Copyright Theft
- Anonymizer
- Child Abuse Images
- Criminal Skills
- Hacking
- Illegal Drugs
- Marijuana
- Spam URLs
- Botnet
- Command Control Centers
- Comprised and Links to Malware
- Malware Call-Home
- Malware Distribution Point
- Phishing/Fraud
- Spyware and Questionable Software
- Peer-to-Peer
- Torrent Repository
- Hate Speech
- Pay to Surf
- Nudity
- Pornography
- Sex and Erotic
- Content Server
- Private IP Address
- Redirect

To the surprise of many advertisers in the vCE Charter Study, 72% of the campaigns had at least some impressions served in this type of inappropriate content, which spanned a total of 980 sites. Even though the actual percentage of impressions was less than .01%, the study showed that 92,000 people saw these ads.

With the increasing use of social media, a snapshot of a marketer’s ad in an inappropriate environment can quickly go viral, exposing many more people to the unintended, but negative, association of a brand and inappropriate content. With 92,000 people being exposed across all vCE Charter Study campaigns, the advertisers’ concerns are justified.

Fraud

Defining Fraud

The vCE Charter Study specifically measured two aspects of inappropriate delivery:

1. The incidence of ad delivery via non-human spiders and bots identified by the IAB
2. The incidence of ad delivery on sites with clear illegitimate and intentional fraud

The inherent complexity in the online advertising landscape results in a lack of control or visibility into online ad delivery. The complicated daisy chain of ad delivery can involve up to 20 different players, and quite often neither the buyer nor the seller has insight into each step in the process.

The term “fraud” as it relates to online advertising encompasses a variety of impression-delivery scenarios. In some cases, there is direct fraud, which is deliberate and completely illegitimate, while other types of fraud are an unintentional by-product of legitimate business practices.

List of Non-human Spiders & Bots Identified by the IAB

To help members of the online advertising ecosystem better understand and avoid issues relating to fraud, the IAB maintains a list of all known non-human spiders and bots. All IAB-accredited ad servers are required to filter out these known sources of non-human ad impressions. An analysis of vCE Charter Study campaigns showed that the average campaign in the study had 0.16% of total impressions being delivered via these spiders and bots, with a range of 0.03% to 0.49%.

Sites with Intentionally Fraudulent & Illegitimate Activity

Analysis revealed more than 200 sites that were guilty of this type of fraudulent delivery.

Implications: Putting all of the Pieces Together

The vCE Charter Study demonstrates that each dimension of ad delivery has a significant impact on whether or not an ad has an opportunity to achieve its intended objective, and should therefore be a central component of ad delivery validation measurement.

In order to achieve this un-duplicated accounting of delivered impressions, advertisers require a simple solution that eliminates all of the wasted time and error associated with merging disparate data sources. The vCE Charter Study demonstrates that the technology now exists to identify and correct the source of sub-optimal performance, and that the opportunity to do so is substantial.