In the past six months both the blogsphere and the industry trade press have been stumbling over each other to write the obituary of the online display ad.
Naturally, I have an opinion.
In the most recent IAB report for 2008, based on spending, Display (collectively: rich media, digital video and banners) accounted for 31% of the online advertising market. And what was Display’s share in 2007? Again, the same 31%. This should not be interpreted that online display advertising spending has remained flat. Online marketing budgets overall grew by 10.6%. So we can extrapolate that spending on Display advertising also grew by 10.6%, merely maintaining a consistent portion of a growing pie.
So where does all the hoopla over the death of display ads come from? Most loudly from the voices of those who are competing for the same marketing dollars— search marketers and social media consultants wishing to woo those dollars over to their own budgets (Don’t get me wrong, I’m a believer. But I call ‘em like I see ‘em). So if not a reflection of trends in actual spending, what data are these dire predictions being based upon? “Click-through,” of course. Click-through performance is in decline, and therefore online display advertising is a failure.
Click Here Now!
Click-through is a very shallow metric for measuring the performance of online display advertising. It is largely measured because it is easy to measure, not because it is the best yardstick of performance. Fear not. The advertising trade press has declared the demise of click-through… in 1999. And again in 2000. And 2001. And every year up to today. And yet still, click-through remains the primary metric for judging the success of online display ads. This is mostly due to laziness. Other metrics are more difficult to measure and more complex to analyze. Click-through is easy. It is very hard to compete with easy.
Nevertheless, the consensus is that Online Display Advertising is broken. While the metric may be shallow, and the detractors may have an agenda, I won’t dispute that there is a problem that needs fixing.
DejaVu
We’ve been here before. The last time online advertising went through major upheaval, as the dot-com bubble began to burst, solutions for the declining performance of display advertising were also sought. A flurry of activity on the part of publishers, whose revenues were dropping from declining ad sales, led to the adoption of new display ad standards. In 2000 WIRED introduced the “Leaderboard” at 728x90 pixels and Cnet introduced the “MPU” (Messaging Plus Unit) at 336x280 pixels. These two unit, together with six others, comprised a list of eight standards that the Interactive Advertising Bureau (IAB) adopted in February of 2001. Early on, a competing standard to the 336x280 emmerged. As other sites began to accept the 336x280 unit, it became common practice to scale the unit down to 300x250 (exact same aspect ratio) until this slightly smaller bastard unit eventually superseded the original as the more common size (they are both now listed as AIB standards). Today the MPU, the Leaderboard and a third unit, the 160x600 pixel “Skyscraper” constitute 90% of all online display ads sold (anonymous source at Havas). When adopted in 2001 the initial download size for an MPU was 40k. Today the IAB’s recommended initial download size is still 40k.
Now let’s look at some numbers.
I wouldn’t be the first to call for larger ad sizes. But this is not nearly as significant as the lesser mentioned variable— K-Size. Larger media placements are worth nothing if not allocated enough “K” to do something with them. Lack of needed k-size kills more creative concepts than a client with a hangover. And if your concept does survive the k-size crunch it gets watered down like a cheap drink. A media spec with a low k-size will destroy production quality. Yet it rarely gets mentioned in this debate. K-size can be the difference between two or three static frames or smooth flowing animations, it’s the difference between crisp photographs or smudged and rutty image compression.
The chart above only tells part of the story. It would be one thing if the broadband pipe had opened wide but publishers had kept page-load low, optimizing for faster downloads. But this is precisely not the case. A quick perusal of major media outlet home-pages show page loads between 650k-900k. By now some may point out that there are sites that offer a larger K-size spec (though not typically). They are killed by the process.
The Real World
Most amazing about this debate is that it is taking place largely at the exclusion of those who actually make online display ads. So I’d like to discuss now how this process works, how it can work, and what industry changes are needed to make it work right.
The first question one may ask is, why I choose to single out the MPU. Those who work in media buying may be aware that this unit accounts for the smallest of the three largest units mentioned above, that together account for the bulk of online media inventory. Because for the creative department, this unit is the original. By this, I mean, this is the unit that creative is concepted against. In most cases, all other units are considered “resizes.” This part of the process, presenting internally, and pitching ideas to the client, is not unlike the process used to develop creative for any other media channel. What is different is that the concepts have to be executable in 40k. True, some publishes may offer more K, but it is rarely able to be developed against. This is a result of the way online media is purchased. Let’s have a look, shall we?
For media, that plan is fine. However, in the creative and production budget, that is one line item: a 300x250 banner, to be trafficked to six media outlets. There’s a 120k unit in there, but there will never be a 120k unit trafficked. Because a publisher that only accepts 40k cannot receive a 120k ad, but a publisher that has a 120k spec will accept a 40k ad. And the client isn’t paying the production costs of making multiple versions of what they view as one ad. So the agency can either develop six different versions, degrading in quality/functionality with each drop in K-size and eat the production costs (which they won’t). Or they can just develop one unit at the lowest common denominator spec and traffic it to all the different sites. This is the industry-wide practice. This is why, regardless of the publishers’ individual specs, in nearly all cases only 40k units ever get produced.
There was a time when all of this media was managed in-house, by the advertising agencies and digital boutiques who developed the creative. But over the past decade, as online media budgets grew enough to merit attention, the major holding companies spun the online media buying departments off from their individual agencies, and each consolidated them into one of their dedicated media buying subsidiaries. I can see how this seemed like a reasonable strategy at the time. The economies of scale work great for broadcast. From a media perspective, a 30 second spot, is a 30 second spot, is a 30 second spot. As a unit of media, they can be shuffled interchangeably. This is what is attempted with the IAB standards established in 2001. But those standards were established when about 90% of the US internet audience was on dial up, mostly on 800x600 monitors (followed by 640x480!). Even when some sites offer more bandwidth, all it takes is one site in the media buy with a lousy spec to ruin a campaign. One irony of this process is that the more the client spends on media, the the greater the chances that all of the media will be dumbed down. A smaller client will have a smaller media buy, possibly limited to only a couple of outlets. If there is a buy with a large k-size, they will be more likely to be able to take advantage of it in their creative development and production.
When media buying was done in house, there was a way to handle this. When a particular publisher had a more restrictive ad spec than others in the buy, a member of the creative team could walk down the hall and ask for help from their account’s media planner, who would then set up a call with the publisher. Knowing that the agency controlled the media dollars, publisher were much more cooperative. The decision would be escalated to a director level account manager on the publisher’s side who had the authority to overrule the standard ad spec.
Today the creative agencies no longer control these media dollars. While there is pressure from media buyers (and clients) for larger ad sizes, there is little pressure for more bandwidth/k-size. The full ramifications of k-size seem to be seldom understood by media buyers or clients, who are a few steps further removed from the actual process of making the ads.
On the exceedingly rare occasion that a call can be coordinated with a publisher to discuss accommodating a creative concept that is outside of the existing media spec, the agency now lacks any leverage at all over the publisher. Hence, they will no longer defer to a more Sr. level member of their ad sales team to make the decision. Instead they defer to the Web Master, or a Sr. member of the site development team. They have a different set of motivating criteria. While account management was previously motivated by the fear of loosing a piece of business, whether implied or explicit (“We’re already developing this ad unit to a more generous media spec for another site. If your site cannot accommodate this spec, we will have to remove you from the media plan.” I’ve heard an in-agency media planner say this to a publisher. It works.). The site-dev team at the publisher does not think this way. The publisher knows the agency has no other option but to grovel and beg. If it were any other case, they would not get deferred to a dev team. The development team will blame it on testing. The stereotypical response is, “Going outside of our existing [iron-clad, carved in stone] media spec would require additional testing. I’m afraid we have to say ‘no’.” The development team does not look at such a request in the context of their company’s revenue. To them, this is a testing and QA question, and frankly viewed as an annoying diversion from their main responsibility— building the publisher’s own website. When asked, the answer is “No.” Everytime “No,” always “No”.
Some might ask, why doesn’t the agency bring in their media counterpart to advocate for them on a joint conference call between agency, media and publisher? That only works on paper. In this scenario the creative department, working through account management, sets up a call with the media firm to coordinate setting up another call with the publisher. And the client will want to be in on both the agency/media call, and the subsequent agency/media/publisher call as well. That means coordinating multiple schedules between multiple individuals at multiple companies… twice, just to get the conversation started. All the while moving against the fast paced schedule of online campaigns that generally go from brief to traffic at about 10 times shorter schedule than they do for broadcast or print. The ball just moves too fast for that much bureaucracy. Furthermore, being on the same team or not, the media firm always views this sort of thing as an encroachment upon their “turf” by the agency side.
A good argument could be made that the networks need to foster tighter relationships between the agencies and media-buyers in their network. Deeper relationships than a monthly or quarterly director-level status meeting, and involve people that actually do the work, so they’re able to be more agile. But my money is against it— against the idea that they would ever do so, and against the notion that it would be successful if they did.
This is the way the industry actually works. Yes, it is broken.
My recommended solution would be to move online media planning back into the agencies. This is, of course, easier said than done. Billions of dollars are at stake, and the media buying firms view online as a long-term, high-growth piece of their business. They would not be cooperative about such a restructuring. But if agencies or perhaps some forward thinking clients decided to experiment on their own, it’s hard to fight against success. Perhaps it could be a model to emulate.
And it just might help save the publishers from themselves.
I have a lot more to say on this subject, including some recommendations on solving some of these problems, but this article has grown to a length that I feel is stretching the short attention spans of likely readers. So I’ll stop here, and write a follow-up, once I have some comments.
Monday, August 3, 2009 at 7:46PM 
next several months multiple vendors will be introducing new models, and Google 
