For the past two editions, newspaper techniques has introduced and delved into the evolving AdsML standard that is being created and developed by an international consortium. The global standard, based on the widely accepted XML standard and supporting other electronic exchange standards, is intended to govern how digital information is exchanged for all types of advertising, across all media and through all stages of the lifecycle of an advertisement. For today’s partners exchanging information for the publication of classified ads, the lack of standards in this area breeds complex, time-consuming and expensive workflows, to say the least.
These problems are at the core of what AdsML is intended to alleviate: to improve communication, reduce errors, save time and money, enhance business relationships and create new business opportunities. In case you are not yet familiar with the broad definition of what AdsML is all about, see our past two editions (April, page 8; May, page 42) and/or visit the AdsML web site (www.adsml.org).
Classified challenges
Before saying that AdsML is the quick fix or the Holy Grail for the complexities and issues surrounding classified advertising, let us first define what some of these problems are today. Although there is a number of difficulties, perhaps the overriding issue is that there is no clear standard defining how the core information of these small, multi-categorised ads should be exchanged between the potentially numerous parties.
According to consortium participant Marcel Dumont, director of Rosetta, the Netherlands-based company that creates and develops classified advertising web sites by integrating newspaper (and magazine) ad booking systems with its XMLbased solutions, there have been just a few standardisation attempts in the classified sector. The only available standard is CREST from NAA, for which AdsML will support, but he said it does not have wide support in Europe and has limitations in its use. “For this reason [lack of a standard] all players in the market have been forced to make up their own formats,” Dumont said. “When two or more systems have to interact with each other [in this case] a large amount of time and money has to be spent to get them to work together.”
When you take into consideration the volume of classified ads that a newspaper or niche publication can deal with on a daily basis, sometimes numbering in the thousands, and because this business can account for 40-50 percent of revenues in some markets, the stakes are high. And for now, most publishers are spending a lot of time and money on trying to input, categorise and extract the core information to publish that small classified ad in THEIR publication. When that ad needs to be passed on to an online version or perhaps aggregated with another set of classifieds or sold to another publication, this is where it can get really tricky. “Exchanging that information is difficult as you have to agree on how you transport that information between systems,” Dumont said.
“When you are dealing with a classified ad, at a minimum you have to know where will this ad fit within your newspaper’s classified scheme or ‘classification tree,’ ” said Tony Stewart, director of consulting for RivCom, which is tasked with guiding the specifications process of the AdsML initiative. “Most of the important issues dealing with classified ads pertain to the classifications of these ads. There are categories, sub-categories and on and on. The problem is that everyone who publishes classifieds has their own way of deciding how to arrange their classified pages.”
Stewart said to take the example of a person wanting to place an ad for a car in a newspaper. This advertiser (or seller) will include a lot of information/text for that ad: make of the car, let’s say an older car like a Mustang, the year it was built (say 1966), the colour, the mileage, and so on. Upon receiving this ad, the newspaper may categorise it under its “Classic Sports Cars” section. When the newspaper wants to sell that ad to another publication, there is a good chance the other publication might not offer a “Classic Sports Cars” category. Instead it may offer an “Expensive Cars vs. Low-priced Cars” section. Then that publication must mark up (add metadata) the ad to fit into its classified scheme.
“This is the essence of most of the problems we have in moving the information around – it’s in this area,” Stewart said. “We are talking about computers trying to talk to each other in a seamless fashion. However, I go through the trouble of coding the ad the way that I need it to be published in MY publication, and the things that readers care about in my publication may be different than what they care about in some other publication. In turn, that other publication must figure out where to put that same ad in his or her pages. The problem that we need to solve is how do I mark this ad up in a way that when I send it over to you, you can then more easily infer from looking at the way I marked the ad up just how this ad will fit into YOUR classified scheme.”
Dumont said some of the costs and time-consuming tasks could be reduced further if there was more automation of the language/content checks of the ad within systems. More structured ad order entry also would help, especially with more publishers offering customers the opportunity to input their ads through the web, i.e. these ads then need to be proofed or marked up. But today’s typical labour-intensive scenario of someone at a newspaper taking an ad over the phone or via fax, etc., then either adding metadata or (more often) having a unique system in place downstream to identify the various tags and extract this information so it can be used in an appropriate fashion, is reason enough for more structured ad order entry.
Help Wanted
Obviously, the AdsML consortium faces a big challenge here, one that will not be achieved over night. But Stewart said he thinks there is a realistic solution to be offered by the consortium: “By helping the industry to agree on the types of information and the terms that will be used for that information for the markup of the original contents of a classified ad. We have to find a way to get people to agree on things such as prices, measures, etc., on terms for different types of ads, colours and so on. If we can find a way to standardise what it is that people are trying to extract, what format it should be put into for other players to use, and put into place a mechanism by which we can keep extending the list of what needs to be extracted – because over time more and more things will be discovered – then we will have gone a long way.”
Dumont said for the publisher, the benefits of AdsML regarding classifieds are obvious: Because AdsML would allow information to be abstracted from vendors software, the publisher would then have more freedom to add/change the products it is using, and possibly bring in more revenue as it would be easier to repurpose content for other publications and media.
Also, he said a publisher can expect cost reductions as there is no need to make expensive interchanges between systems (using such EAI [enterprise application integration] solutions from Tibco or Bizztalk).
Good for developers
But Dumont stresses that the vendor community, which he is a part of, will also reap the benefits of a de facto standard for exchanging information between advertising systems. “There is currently a major obstacle in the sale of advertising systems if other systems have to be integrated,” he said. “It [integration] is expensive, takes time, involves a large number of people and is prone to delays.”
Still, a certain degree of integration is probably inevitable, yet this sometimes painful process can be improved. David Jones, managing director of Vio Worldwide and a member of the consortium, said the implementation of AdsML for suppliers should cause a positive domino effect. “Using standards like AdsML will help developers build an interconnection once between two systems and then reuse it many times. That will reduce the labour component in implementing integration projects, and therefore reduce the capital cost of innovative automation.
Download newspaper techniques article “Classifieds present test for AdsML”