Earlier this year SearchLine wrote about an effort among a group of online retailers, e-commerce firms and Internet shopping engines to standardize procedures for submitting data feeds of products to online shopping engines and marketing around SKUs. Here’s an update on the progress of that effort.
First, the group, started largely as an ad hoc discussion at the Shop.org annual meeting in Atlanta last January, now has a name and a Web site: Online Retail Datafeed Standardization (ORDS) at www.onlinefeeds.org.
Just as importantly, the group has found a technological leader in the Association for Retail Technology Standards (ARTS), a division of the National Retail Federation, which sponsors the Shop.org online retail industry group. Richard Mader, executive director of ARTS, says his group’s role will be in part to figure out where existing standards from other retail processes can be usefully applied to the data feed problem, and where entirely new standards need to be generated.
“We’re in the business of creating standards,” Mader says. “This group, many of whose members are also in the NRF, was looking for technical standards that would send data from retailers to comparative shopping engines and get back data relative to simple things like error messages and analytical trends, such as keywords.” So ARTS proposed that it had some standards developed for other parts of the retail industry that could be adapted to suit the needs of both shopping engines and the retailers who want to market through them, such as processes for handling promotions and item categorization.
One important consideration was the need to make sure any standards are platform-independent and vendor-neutral. The 184 retailer and vendor members of ARTS cut across a wide range of product categories and software solutions, with 40% based outside the U.S.—seemingly diverse and neutral enough to develop a set of tools that would have credibility with most retailers.
In late February, a core group of marketers, agencies and search engines came together at a regular ARTS meeting in Menlo Park CA to discuss forming an official industry group, and the best way to proceed. Arthur Coleman, CEO of multichannel marketing consulting firm r:eThought, presented the results of a study comparing the data feed practices from top comparison shopping engines, including Froogle, Google, Yahoo!, MSN, Nextag, Become.com, Shopping.com, Shop.com and PriceGrabber.
Coleman reported finding significant practical differences, starting with the fact that some engines give no feedback at all to merchants on errors that occurred in data transmission, while others give aggregate counts of successes and failures—not very helpful in finding and correcting incorrect feeds for specific product SKUs.
Another problem was the wide range of different taxonomies, categories that merchants and shopping engines use to categorize merchandise in their offering. The forms that merchants must fill out to send their product data to the engines also differ widely, increasing the chance of error and the work that marketers must perform to get items listed. For example, Coleman noted, Yahoo! Shopping has two mandatory fields, 13 optional ones and 40 possible setting combinations for each SKU; Shopping.com has one mandatory field of 21 characters with three options; and Shop.com has four fields of 255 characters each. Other metadata can also be handled differently from engine to engine—things like global shipping and payment settings, feed file requirements, and submission data.
“The NRF for years has published guidelines for the industry on merchandise structures, so we have a good starting point for regularizing that process,” Mader says. “We can look at what the engines have done, look at support that the NRF gave them two years ago, and help with this effort. You’re never going to get a total standard for taxonomy, but you can put a pretty good guideline in place. And if you do a good job, people will follow it.”
Mader says the standards will be persuasive if they are seen to be vendor-neutral, improve the return on investment for marketing on shopping engines, and win the acceptance of at least a portion of the large retailers that make up ARTS’ membership rolls.
One issue that came up among participants at the Menlo Park meeting was whether to require merchants to offer their data feeds in extensible markup language (XML), a text format originally developed for large-scale electronic publishing. Many merchants now send their feeds in the form of highlighted spread sheets and data streams delimited by commas—an old holdover that’s relatively easy to program, but also likely to contain undetected errors that can get their product data left out of a shopping engine index.
“ARTS does not build data messages in any format but XML,” Mader says he told the group. “That’s what all the big guys are using now. Yes, there is a learning curve, but we can make that simple and provide some education. So if the marketers don’t adopt XML now for their data feeds, they’re going to wind up coming back and adopting it a few years from now.”
Mader adds that very simple XML tools exist to help merchants translate their product offerings into this flexible, standardized data format, and that once they’ve made the transition, transmitting the data to the comparison shopping engines can be a breeze. “We’ve built schemas where you can actually send your XML feed via e-mail,” he says.
At the moment, Mader says, ORDS has a call out for a work team of representatives from a handful of retailer, agency and search engine representatives who will pull together a set of draft standards. He hopes this will be accomplished by the next formal ARTS meeting in the second week of May and that the full ARTS membership will be able to start offering comment on those drafts soon afterward. The aim, he says, is to get a technical standard accepted and in publication within the September-October time frame—in time for the ramp-up to the holiday online shopping season.
Alan Rimm-Kaufman, head of marketing consulting firm The Rimm-Kaufman Group and one of the original drivers of the standardization initiative, said the Menlo Park meeting encouraged him about the effort’s prospects for successful acceptance by all the parties: retailers, engines and agencies like his own.
“If I had a criticism, it would be that the retailer participation was low,” he says. “The meeting took place outside of a retailer convention, and I think merchants were underrepresented. I’d like to see the next one of these meetings take place in the context of a Shop.org meeting.” Rimm-Kaufman says he plans to suggest holding such a retailer-focused gathering on standards at an upcoming Shop.org meeting.
Overall, though, he’s enthusiastic that the standards effort has gotten off to a strong start and is headed for a successful result. “The more I talk to search engines and clients, the more I realize the need for better vertical integration,” he says. “The comparison shopping engines want richer data, and the retailers want to provide it, right down to store-level, SKU-level product availability.”
More and more, Rimm-Kaufman says, general search engines use search queries as a way to almost intuit what users are in the market for; they offer users not only their search results but related ads, multimedia content and shopping suggestions built on those keywords. “Retailers want to figure in that offering too, and for that reason they think it’s important to give the engines as much product content as they can,” he says.