If a tree falls in a forest and no one is around to hear it, does it make a sound? [UPDATED]


UPDATE 4/26/13 via Marketing Land: Google has just obtained accreditation by the MRC for its viewability measurement solution ‘Active View’; advertisers will be charged only for impressions that meet the IAB’s proposed viewability standard of at least 50% on screen for one second or longer. Active View reporting will be available on the Google Display Network (GDN) and in DoubleClick for Advertisers and DoubleClick for Publishers later this year. NO word as yet on how limited it will be given the ongoing issues with IFRAME still to be resolved by the IAB’s SafeFrames initiative. In a parallel post Marketing Land  has a nice summary of Googles announcement about its ‘Active View’ product and that viewable ads garner a 21% higher CTR as per their research. Of course one should expect a higher response when far fewer ads are invisible  in addition to Google’s enthusiastic embrace of the issue given how much of their revenue comes from clickstream activity across its display network.


In case you haven’t heard of Marketing Land, its a year old companion to Search Engine Land, the eponymous search marketing resource started by Danny Sullivan, early stage evangelist of all things search engine. I highly recommend adding it to your reading list, it covers a fairly broad array of digital marketing topics and produces a healthy amount of original content that’s to the point and relevant.  This week they share news of how TripAdvisor, an under the radar social network behemoth, is raising the bar in the debate about online ad effectiveness.

You will likely recall the issue of ad ‘view-ability’ pop up last year on the back of a comScore study titled ‘Changing How the World Sees Digital Advertising’, a copy of which you can find here. There was also a similar study done by the Media Rating Council (pdf download ) mid year.  Both measured over 2 billion impressions served by between 12-17 advertisers over a two month period and both found a view-ability variance of between 7-100%. ComScore dug up a virtual manila folder worth of somewhat alarming insights, including:

Ad view-ability varied dramatically by site. On one property, 100 percent of ads were in view. On another, only 7 percent could be seen and the rest were wasted.Among ad sizes, the 728×90 Leaderboard scored the best in-view rates, with 74 percent of ads visible, although websites varied widely in performance, from 7 percent to 93 percent view-ability for properties using this size. The 300×250 Medium Rectangle format delivered in-view rates of 69 percent, and the 160×600 Wide Skyscraper delivered the lowest at 66 percent 

There was almost no correlation between the price an advertiser paid and whether or not an ad was in-view72% of campaigns had at least some ads served into “non-safe” environments, the definition of which can include adult sites, sites with hate speech, illegal drug websites, spam URLs, and sites pushing questionable software, among others. (The actual percentage of impressions in those environments was very small – less than .01 percent. Even so, 92,000 people saw these ads, creating risks for the brand)

(Only) 37% of impressions were delivered to audiences with behavioral profiles that matched the brand’s category focus. The more targeting attributes an advertiser layered in, the lower the delivery rate.

Sounds like site selection needs to be taken more seriously and behavioral targeting a little less;  15 yrs ago when ad networks first started appearing it was like pulling teeth to get a site list and while the prices were lower on ‘blind buys’ there was a good reason for that. While we all want more than what we pay for when buying media i.e. ‘a good deal’ almost always you get what you pay for. Remnant/bonus/value add/free whatever you want to call it is ad space that nobody else wants to pay for, like the sale rack at Old Navy. Sure it brings down the effective cost of a buy or a plan but if the publisher cant sell it there is no opportunity cost to them and you just pay more for what they can sell.


When we talk about ‘view-ability’ we refer to the ability or ‘opportunity to view’ that a site visitor has to see an ad in practical terms ( the official criteria is 50% of pixels visible for 1 second). Some ads are obscured from view due to placement at the bottom of a long page or on a page that garners minimal traffic (About, Careers, Contact us and less ethical destinations like mirror pages and black hat SEO stuff). This is an important consideration for media buyers when they are paying for all these ad impressions; if 30% of your media placements are automatically non view-able, which a recent study proclaims as indicative of the internet average, then your starting the race from the back of the grid with a three lap penalty.


0-60 in less than a minute

This is not to be confused with TV & Radio, where the opportunity to view is 100% as long as it airs correctly. Of course if we were to say conservatively that 90% of the time nobody was paying attention to TV ads because they were using these programming breaks to look at their iphone, make a cup of tea,  do some arm curls, vacuum up Dorito debris, stuff like that basically. Just because viewers choose not to look at TV ads en mass does not mean they are denied the opportunity to as when an internet ad is served onto a page a reasonable person (Man on the Clapham Omnibus in UK, Joe Six Pack in America and an Aussie in Australia) would almost certainly not be aware of nor visit. If 90% of TV ads were watched the money saved could end world poverty.

Reading through both reports comScore has done a superior job with  its findings although the 3MS (Making Measurement Make Sense) advisory board backed by the IAB, the ANA the 4A’s, the NAA, the OPA  and the OPP (nice to see everybody on the same page) choose’s to go with its own MRC backed study. The biggest roadblock in measuring view-ability is between 60-80% of online ads are served via a third party ad server  using an IFRAME and you cant track their visibility due to prescient provision added to internet law by some of its founding fathers called the ‘Same Origin Policy’.

This general rule  is designed to protect websites from being altered or snooped on by a nefarious third party. It essentially prevents advertisers and other entities from gaining read-write access to alter and communicate with another sites source code-just like the ancient format known as CD’s help ensure your kid brother doesn’t record a Wiggles podcast over your signed Winger Greatest Hits CD and worse participate in home taping!  If your too young to know what a CD is just assume its a rule designed to stop hacking.

A post authored by two notable web leaders on Exchange Wire posits the challenge thusly:

For the security-conscious publisher, iframes provide an impassable barrier between the host page and the ad. This protects both publishers and users from ad scripts. Unfortunately iframes also prevent rich interactions. More troublingly, iframes limit what the ad can find out about the host page—for example, geometric position of the ad, whether the ad is viewable within the browser viewport, site metadata, site name, etc.

Good news is the IAB is working on a work around called ‘Safe Frames’ that will enable better measurement of the 80% of ads served through IFRAMES.  Theres  even a neat infographic, JavaScript to the Rescue! If your not familiar with IFRAMES (‘Inline Frame’) its a piece of HTML code loaded onto a website that calls content from a remote location much like when you use an app on your smartphone. Apps do not have the ability to change the way your phone looks or acts, just to provide a secure access tunnel to content hosted elsewhere like on Twitters database. If you are stuck on a desert island and would like to know more about this topic, I thoroughly recommend this uniquely user friendly post over at DWUser

Back over to Exchange Wire for more info about this most enthralling SafeFrames initiative:

SafeFrames are an attempt to provide advertisers both with transparency and rich interactions, whilst at the same time ensuring publisher and user security (and, indeed, advertiser security).

External content is isolated on a trusted third-party domain, thereby providing publishers the same protection as provided by traditional iframes. Communication between the host page and the external content is available via a defined API. This allows for rich ad interaction and it allows for certain information to be made available from the host page to the advertiser.

Unfortunately there  will always be trust limitations even with this workaround;  to use an archaic turn of phrase, Garbage in: Garbage out. Here’s the problem- publishers still, rightly, control what information is made available for the IFRAME to collect hence the possibility still exists to game the system by pretending ad placements were view-able among other things. Since we typically don’t concern ourselves with scenarios like this, we take data at face value which leads to Garbage In: Gospel out. The risk/reward ratio precludes established publishers (in mature markets) from fiddling the numbers but when your dealing with a network of low profile sites in a network you take your chances because they will be.


Still not seeing anything

Still not seeing anything

Lets now consider what Trip Advisor brings to the table within the context of the current status quo on ad view-ability  While the various industry bodies, metrics merchants and other third parties pivoting towards this opportunity discuss the challenge (3MS is aiming to introduce digital GRP guidelines next month along with the SafeFrame later this year) TripAdvisor has trialed technology that  determines how far a browser has scrolled  below the fold  in order to verify if an ad has passed the ‘in view’ test of 50% of pixels visible for 1 second (I hope this is a temporary lowest common denominator benchmark to prevent  publishers from heading to the exits.  Seeing half an ad for 1 second is unlikely to trigger any conscious recall or offer the opportunity to click on it. But it is a start) Trip Advisor calls its new invention ‘Delayed Ad Call’ and touts some mutual benefits:

TripAdvisor’s VP Global Display Sales, Martin Verdon-Roe:

It means that not only are our ads more trustworthy but the performance for the clients has also improved. CTR is up and we have also implemented a tracking technology for “ view-ability” and that is [now] showing that the ads below the fold are now just as valuable as those above. In fact the ones below are now producing a stronger view-ability metric compared to the 728×90 which is at the top of the page.This then leads us down new avenues and opportunities. E.g. can we find areas below the fold with high user dwell time. We can put advertising placements there in the confidence that we will get great user exposures but only when the users actually get there. This helps the advertisers performance and allows the sales team to find new impressions on TA that otherwise would have been unsellable.

Trip Advisor is smart to promote this kind of capability; it draws attention to their property and will open a few doors for their sales team. While they have hurdled the IFRAME issue this solution is entirely publisher centric; I doubt it could ever be optimized via a third party ad server, that would contravene industry rules governing IFRAME behaviour noted above. Plus the publisher is looking to shift below the fold inventory at a higher price using this ‘bespoke’ (Hello London readers) device- as a media buyer what incremental dollar value to you place on this all things considered? For now its gimmick worth watching and potentially the thin of a wedge; publishers taking it upon themselves to offer more transparency in return for higher prices.

The good news is something is being done to raise the credibility of the beleaguered banner ad; rumors of their demise are greatly exaggerated, much like a prize fighter staggering around the ring taking hit after hit but never going down. The biggest hurdle for marketers to investing in online media 15 years ago was the small screen size and the even smaller ad sizes.  Objections  like “I don’t want my brand in some tiny little box, the logo is too close to the text” or “Nobody clicks on those stupid things, I know I don’t, neither does anybody I know, its a waste of money” and “Look the CEO says his 15 yr old son clicks on some of those internet ads and he’s convinced that’s why we should not go near them” were an everyday challenge 15 years ago. Of course in time everybody got used to the ‘smaller screen’ of a PC, then a laptop and a tablet and are now grappling with the value proposition of display media on a smartphone. Perhaps the next frontier is smart watches or even miniaturized devices.


APPENDIX: Some Historical Context on Ad Tracking- If you are reading this you probably are already aware of this issue and the exciting world of campaign implementation and ad tracking. I chose not to include the excerpt below in the main post for brevity reasons. My first job in advertising was a traffic coordinator using the original ‘mom and pop’ looking DART interface when flash was still in its infancy. If you are curious here is some background on the original ad measurement issue, the discrepancy in reporting between publishers and advertisers tracking systems.

One of the biggest battles back when third party ad serving was still new in the 90′s was over which set of numbers would we the agency pay on- the publishers internal impression and/or click count or what DART counted in its system. We both relied on a tracking mechanic that measured activity at two different points on the ad and click request timeline which always led to discrepancies. Believe it or not during the ~3-7 seconds (or 30 seconds if you were using AOL) between you typing in a URL and the web site page appearing in your browser ( known as page load time) there is a drop off in traffic of between ~10-30%. The cause is netizens closing or refreshing their browser, shutting off their computer or just deciding Yahoo is not such a cool place on the internet to go visit.

The problem is the publisher counts an ad impression when somebody types in yahoo.com and hits enter (which loads a webpage including banner ads) whereas the ad server counts an ad impression when it receives a request from the publisher to serve the ad remotely (known as an image call) via a redirect tag often called an IFRAME.  Same problem when somebody clicks on a banner ad; the click is immediately recorded by the publisher but the remote ad server (DART) waits until it receives the request to redirect the  users browser to a new web page.

All this adds up to a higher ad an click count for the publishers, which was a non issue until DART came along in 1998 and added a second set of numbers. The argument over which method was more accurate and/or to be billed on raged for years until it just became standard to go on the lower count from ad servers like DART since thats considered the most objective count.

Quick note about North Asia (Japan, China & S Korea): Its still uncommon to be permitted to serve ads via a third party ad server like DART or ATLAS because it would represent a significant drop in revenue for publishers, who are far more powerful than agencies and marketers in these markets. In the case of China  another key factor for portals is the concern that ad content could trigger the ire of government censors.





You can be the first one to leave a comment.


Leave a Comment