A welcome trend in online advertising is the reduction in the number of ad units per page … welcome to users that is who have conditioned themselves to skip over anything that flashes, dances, or interferes with the content of a page. It’s also a welcome trend for page designers, who can spend months on an elegant page template and palette only to see it trashed by a proliferation of ad units.
The bad news for publishers and advertisers is the reduction in clutter also means a squeeze on inventories. The 90s saw sites such as Forbes.com get "NASCARed" (to use our term at the time) with dozens of micro buttons that permitted publishers to report up to as many as a dozen "impressions" per page. Since traffic was measured with a dull axe in those days, and because advertisers and their agencies weren’t really measuring performance to any exact degree, people could get away with claiming "billions and billions served." Remember, these were the days when an online publishing executive could stand up and claim that "hits" were a relevant measurement, and when no self-respecting online publisher would admit to anything less than millions and millions of those hits, ignoring the reality that a hit counted every image, every server-side include, and parsed a site into its content components.
The standardization of units by the Internet Advertising Bureau saw a reduction in clutter-ads and a big transition towards bigger, but fewer units such as the skyscraper (which we invented at Forbes), leader boards, big boxes, and other more substantive units.
Bigger is certainly better for all parties, but has lead to the challenge du jour for online publishers — inventory expansion. Want insight into the M&A flurry over the past six months? A rationale for the Times buying About? WSJ.com buying Marketwatch? It all comes down to page views. With internet usage beginning to plateau, publishers have to find the page views somewhere — and syndication, RSS, link sharing and other traffic swaps are only going to carry them so far.
Last week Doubleclick published a ten-year history of online advertising by its Director of Research, Rick Bruner. He writes:
"… the growth in the number of unique visitors and page views has slowed to an almost negligible rate compared to years earlier. Among the 20 sites with the most display ad impressions, the total number of page views was up only 5% from Q4 2003 to Q4 2004, according to Nielsen//NetRatings. At the same time, the number of display ad impressions in the last year among major U.S. sites is down (5% down for the top 1,200 ad-supported sites; 13% down for the top 20 ad-supported sites). Part of that decline in overall display ad impressions among the largest sites is due in large part to a reduction of clutter, as sites increasingly feature fewer smaller ad units (such as buttons and half banners)
and standardize on the new larger ad sizes promoted in recent years by the IAB (especially extra-large banner “leader boards,” wide skyscrapers and medium and large rectangles)."
As publishers try to drive traffic to a flattening audience (traffic growth is not infinite), clutter may return. Dave Morgan at Tacoda writes on Click-Z:
"What do we think will happen as increasingly more ad money chases scarcer, slower-growing Web page audiences? Most online ad buyers I informally polled over the past few weeks gave me the same answer: Web pages will become increasingly crowded, expensive, and less desirable."
Bruner disagrees, saying in his retrospective that the de-clutter movement will continue:
"Premium media brands are likely to attempt to further reduce ad clutter to avoid the risk of turning off their audiences. Publishers that depict the sponsorship value of advertising more transparently to consumers and at the same time reduce interruption-overload will benefit from more loyal audiences
and higher ad prices."
Something has to give or something needs to be invented. As targeted advertising meets CPC models, insertion technology has to evolve to truly offer contextual advertising, drawing from a pool of ads and placing them in the right place at the right time. The old model of registering one’s traffic to do crude IP look-up serving based on geography has to move much further into clickstream analysis, parsing, on the fly, the session’s attributes and attempting, in real-time to shift the served inventory towards the bias of the user. Falling back to the old NASCAR model of splattering a page with a higher number of units will drive users away.
Traffic development, syndication (The NYT reported huge increases in traffic thanks to its RSS feeds), and better insertion and targeting are the way to fix the dilemma. Publishers need to jack their CPMs to reflect the growing value of their limited inventories, building a performance case to sponsors and advertisers and not falling over themselves to kneel before the altar of Big Numbers.