Category Archives: ad server

What’s the difference between viewthrough and viewability?

Leave a reply



In case you haven’t noticed, digital marketing performance is more important than ever. With advertiser’s increasing need for accountability and measurement tools evolving to meet that need, there is major and growing interest in associated ad analytics or adverlytics. More specifically, display viewability measurement and attribution are hot topics. Yet, before the explosion in ad viewability technologies and an enthusiastic trade group response to standardize (IAB, MRC), viewthrough remains complicated and mysterious.

Even among experienced digital marketers, agencies and analytics professionals, there is often confusion around these terms with many people using the terms wrong or inter-changeably. They each describe important steps in the digital advertising funnel with viewability an important first step towards measuring exposure and viewthrough gauging post-exposure latent response. It makes a lot of sense to use these tracking methods together.
Viewthrough Defined
Display viewthrough measurement has been around since the late 90s and the advent of agency or
3rd party ad serving (3PAS). That means an advertiser’s digital media agency centrally manages the serving of their display ad campaign across multiple Web sites and/or ad networks. Typical tools used for this purpose include Sizmek, Pointroll, Google DoubleClick and Atlas. The beauty of this approach is having one report that consolidates one or multiple campaign performance metrics across multiple media vendors. The advertiser-centric approach and even allows for performance analysis by placement, ad size or creative treatment across media vendors. Prior to this, agencies had to collect the reports from each media vendor and doing cross-dimension analysis required advanced Excel manipulation. Included in the consolidated report was usually impressions, clicks, media cost a calculated click rate and effective CPM (cost per thousand). Potentially average frequency of ad delivery could also be reported.

How Viewthrough Works
As a more advanced option, the ad server could generate a special beacon tag which could be coded on an important digital event on the client’s Web site, e.g. a visit to the home page, playing a game, initiating a particular download and most obviously the rendering of an order thank you page. The latter being a revenue-generating conversion. The way it works is simple enough: the ad server site code that is placed is page-specific and named in a certain way with a certain ID. When that beacon tag fires, the ad server increments the count for that specific ad. The ad server uses clicker cookies to distinguish between clickthrough conversions and viewthrough conversions by the presence of a click cookie that is served to the browser if they click an ad. Generally, a viewthrough can be simply a visit to the target site or it can be an actual ecommerce conversion.
Fig 1. Viewthrough Process
In practice this meant that digital advertisers were no longer limited to the more direct-response oriented clickthrough rate for measuring display campaign performance. With viewthrough they could also analyze the passive impact of display which is more difficult to measure since it is not immediate and is harder to measure than clickthroughs which use appended query string parameters in the landing page clickthrough URL. That said, viewthrough measurement from the ad server is not enough as unscrupulous ad networks learned to do what is called cookie-stuffing to take advantage of affiliate advertising systems.
Site Analytics Technology Matters
Going beyond simplistic, ad server viewthrough measurement is not easy for digital marketers and technology choices matter. The complexities of generating and managing ad server site tags has also made ubiquitous measurement operationally difficult. What’s more ad server counting of viewthrough was an afterthought. The solution to better viewthrough measurement is ingesting ad server data into the site analytics system where site taxonomy, visitor uniqueness, multiple marketing channels and event level tracking infrastructure already exists. Another option are the better algorithmic attribution systems; a couple even integrate display viewability into their solution (ideal).
The reality is that there is are a few site analytics platforms that integrate with the ad servers in the market today. All are not created equal, not independent of media and there are significant differences in the methodologies. The integrations can be complex but well worth it for advertisers spending more than $10MM per year in display media. However, the learnings are significant when it comes to understanding customer’s digital path to the target Web sites and the latent effects of display ads (banner/video/mobile). 
Last, a word of caution: the best way to really understand display media from a viewthrough standpoint is through periodic incrementality analysis, i.e. test and control.
Viewability Defined
Display viewability is a newer capability for digital marketers that has really caught on over the last five years. Essentially, display viewability leverages Web technology that can determine how much of an ad is “in view” of the user’s browser window and for how long. While there is some additional complexity involved, simply put this offers a way for digital advertisers to really understand the quality of the ad inventory that they are buying from Web publishers and ad networks.
Fig 2. Viewability Process
Industry Reaction to Measuring Ad Viewability
Not surprisingly, ad viewability has met with resistance from publishers and even media agencies though it is of obvious interest to advertisers. By better managing viewability, advertisers can ensure they maximize the value of their display spend in terms of both raising awareness and driving a measureable response (clickthrough, viewthrough, engagement, conversions). Unlike viewthrough which languished for years as somewhat of a stepchild metric, The Internet Advertising Bureau (IAB) and the Media Ratings Council (MRC) stepped in early on viewability to create some early standards: 50% or more of an ad for at least 1 second. Advertisers can certainly raise the bar here but there is at least a baseline.
Ad viewability is fairly easy to implement on a display campaign but can often be challenging to operationally integrate into media performance reporting. As a best practice, having an independent analytics team pilot test and benchmark current performance will help gauge how much upside potential there might be.
Rising Tide Lifts All Ships
It stands to reason that you cannot have a clickthrough without an ad first being viewable. If an advertiser were to increase their ad viewability rate from 50% to 75% that would likely mean the number of clicks and therefore clickthrough rate should increase. Similarly, you cannot have legitimate viewthrough without an ad first being viewable. Increasing ad viewability rates should also benefit viewthrough rates as well.

  • Viewthrough – It’s all about response measurement. Advertisers ultimately should be very keen on measuring viewthrough from display media if they are interested in Web site engagement and even more so if an online conversion is possible (offline measurement is also possible).
  • Viewability– Think front-end of the advertising process, i.e. the inputs. The focus is on doing what branding has always sought to do through media campaigns: create an impression, influence behavior or change preferences.

As you can see, while they are certainly related viewthrough and viewability are very different. Measuring and optimizing viewability should be a self-evident way to improve those front-end metrics that will likely also boosts downstream ones. Viewability and viewthrough can and should work together.

IAB Recognizes Viewthrough

Leave a reply
Interactive Advertising Bureau -- Dedicated to the growth of interactive advertising 

From their annual meeting in Pam Springs, CA, the Internet Advertising Bureau published its Defining Cross-Platform Engagement White Paper. While there are plenty of interesting metrics to be found, we’re excited to see a short and sweet definition of viewthrough. Positioned as one of the three categories, Display Viewthrough is included among behavioral metrics.

Definition: Number of Brand site visits that could have been influenced by display media within a particular look-back window.
Measureable Today: Yes, Web Analytics

While this is a small victory for the cause of viewthrough measurement, there is plenty more to do to get the word out about VT. We’d like to advise the IAB that the prior metric included in this section of their study, “Video Completion Viewthrough Rate” ought to get a different name – perhaps “Play-through?” It is bad enough that Viewthrough and Viewability,  are routinely confused but this could actually obfuscate the different kinds of post-impression behavior, e.g. VT Visits, VT Conversions, etc…Also, the following metric listed, “Searched for More Information” could also use a better name – Viewthrough Site Search?

Other considerations:

  • VT does not have to be at a branded site although this is most likely, e.g. related microsites
  • Any high-value task that can be measured in a Web analytics platform could have a VT source
  • Look-back windows are usually set by the ad server not Web analytics systems
  • Like the “could have been influenced” but no mention of how to understand incrementality

Scroll down to Page 10 of the Defining Cross-Platform Engagement White Paper, to read it for yourself.

Facebook Acquisition of Atlas: Sad Day for Digital Advertisers

2 Replies

                                                      The Funeral of Santa Fina”, Domenico Ghirlandaio, 1485

Facebook just announced the other day  that it would acquire Atlas Solutions, a long-time competitor to Google’s DFA in the agency ad server business. As an adverlytics practitioner, many are asking about this and still more discussing the implications. Though the trade press will gush, investors may cheer and Atlas employees may be breathing a sigh of relief, it is truly a sad day for digital advertisers. Why? The choices available for independent third party ad serving (3PAS) just got much thinner. 

The Rationale

To be sure this is a canny move for Baby Google, just as Google’s DoubleClick acquisition engorged the dataplex with rich user-level behavioral data that spanned both both buy-side (DFA) and sell-side (DFP) – it was brilliant. And after several years, the folks at Google decided to invest in the DFA reporting interface and we now have a Google Analytics like wrapper (except it is now green). Not much improvement on core functionality like reach and frequency reporting, but hey – it is better than ReportCentral.

The fact that DFA has maintained such a large market share since the Google acquisition suggests that client’s aren’t paying close attention to ad serving details. The very notion of pushback from advertisers and agencies is so unlikely that now, Google will also tell you how many of their ads were viewable, too. The Facebook deal is banking on it. For that matter, MediaPlex was absorbed into ValueClick way back in 2001 – they also own an ad network, an affiliate platform and Dotomi.

For Facebook, this is a very smart move although it relies on digital marketers being easily distracted and not looking too closely. In this deal, it is not really about “closing the loop” for advertisers. FB could ostensibly do this now with the much heralded unicorn called the Conversion Pixel or View-Tags. What they are “closing the loop” on is their understanding and ability to datamine what many other advertisers campaigns are doing, and the targets of those ad campaigns and those behaviors across client sites via the Atlas Universal Tracking Tag infrastructure. It probably won’t be long before Facebook offers their own Analytics platform or maybe a Tag Management System. Like its hero, FB is often times ethically challenged when it comes to who’s data is it any way. Just recently, it was revealed that Facebook has manipulated their advertisers’ campaign performance reporting.

That said, Atlas as a platform has faded over the years and needs investment to compete. It’s once advanced approach to attribution Engagement Mapping and stream of research from the Atlast Institute was a favorite among the analytics-minded. Yet, under Microsoft this pioneer of ad serving suffered from functional obsolescence as more site-centric and advanced algorithmic measurement has become more available. At the same time, Pointroll and MediaMind pivoted from rich media platforms to full-bodied ad servers. Many digital ad ops people that used Atlas regularly liked it, but later complained about lack of support. The upside is that Atlas won’t be shutting down anytime soon.

However, digital advertisers and agencies should be on notice now more than ever.

Recommendations

  1. Don’t Buy Technology and Media From the Same Vendor – In a world of digital marketing and an endless stream of bright shiny objects, it should give client-side marketers and savvy agencies pause that this is a serious conflict-of-interest that work against them. Since these ad serving tools are often counting ad impressions and clicks that they themselves sold – there is an incentive for self-serving manipulation.
    • For a better idea of how campaigns are performing leverage tools like Omniture, ComScore’s DigitalAnalytix. Coremetrics and WebTrends.
    • For more advanced attribution measurement look at independent tools like Adometry, Visual IQ or C3.
  2. Don’t Share your Behavioral Data – For the pleasure of sharing your valuable behavioral data, you are probably paying $0.04 oto $0.08 CPM to Google (through your agency and this may be marked-up). In this respect, far too many digital media agencies are dropping the ball on data stewardship and going with what is expedient. Ultimately reflects on them, but it also speaks to rampant client-side advertiser ambivalence or worse ignorance. 
    • For those that insist are intentionally looking to harness their user data, at least get something in return for example by joining a data co-op like Akamai ADS (now part of MediaMath). Privacy policy implications may vary.

    Advertisers and their agencies need to understand the very high proce they are paying. Time to look away from the Google, ValueClick and now the Facebook ad stack and consider other choices toute suite. In terms of ad servers left, the remaining major independent ad servers include MediaMind and Pointroll (though it is technically owned by Gannett, a newspaper company this is not as big a data play).

    Final Thought: Where is the FTC?

    Not sure where the FTC will come down on this but they essentially rubber-stamped the Google-DoubleClick acquisition back in 2007. Arguably this deal really does limit choice for digital advertisers but don’t count on the FTC doing much to scrutinize this. That digital advertisers and their agencies don’t value independent ad serving (and free of back-door data siphoning) is their problem and eventually it will sort itself out. To be sure the digital media ecosystem is complex and constantly-changing and federal bureacrtas are more focused on other more important matters. Plus, ad-serving just doesn’t make headlines like nefarious cookie tracking and consumer privacy.