In its second-largest acquisition, Amazon buys the company for $970 million.
(Page 2 of 3)
Vertical engines also claim to deliver more highquality traffic. Greitzer compares the quality of traffic delivered to a retailer from a firsttier shopping engine such as Shopzilla to that which a merchant would get from a major general search engine. “Where it starts to get interesting is if you really dive deep,” he says. “The microniched verticals are a really appealing category. I don’t think it has scale yet, but it’s one that’s interesting to watch. Some search engines we see really try to narrow the content. The people searching on Zafu.com, for instance, are more qualified as jeans buyers than the people searching for jeans on a general engine.”
Besides narrowing content, vertical engines also attract shoppers and facilitate purchases with added services to make choice even easier within their given niche. Zafu.com’s jeans finder feature asks searchers questions about their body type and preferences regarding fit and sizing.
Algorithms built into Zafu’s engine match the shopper’s profile to information on jeans in a database that stores information on the cut of different models of jeans. Information on the price and availability of specific models of jeans at the retailers who carry them is delivered to Zafu’s engine via data feed from affiliate networks such as LinkShare. That allows jeans shoppers who’ve already narrowed the field based on fit to further narrow results by conventional search filtering parameters.
Archetype Solutions, whose technology powers custom clothing applications on LandsEnd.com, also is the company behind Zafu.com and its jeansfinding software. According to CEO Robert Holloway, a Google or a Yahoo tells shoppers-among millions of other things-they can get jeans from this store or that store. But not only is Zafu all about jeans, it also gives searchers more of the information they need to make a purchase than they get on a general search engine. “The fundamental difference is that Zafu delivers a really qualified consumer, where conventional search does not,” Holloway contends.
Zafu doesn’t track through to sale, collecting instead on clickthroughs as well as thirdparty ad sales. But Holloway says retailers tell him traffic is converting well and that Zafu, which started by buying jeans to ascertain their fit, is now deluged by manufacturers and retailers sending them for free to ensure their brands get into Zafu’s search engine listings.
In the six months since Zafu.com launched, more than 1 million women have used its jeansfinding engine, many of them on a cobranded site offered as a service on Shopping.com. Visitor traffic is growing about 25% per month, Holloway says. This summer, Zafu plans to launch similar fitfinding engines for plussize women’s jeans and for bras.
FindGift.com is another vertical search engine that wraps recommendations around basic searching and sorting features. “If people are looking for a digital camera, they already know what they want and they’ll do a keyword search on it,” says Bob Zakrzewski, president and cofounder. “With giftgiving, people don’t have a clue, so you have to invert the search process. Rather than asking them what they want, you have to kind of interview them about the recipient and then start putting intelligent choices in front of them.”
FindGift also makes it easier for shoppers to come back and buy again the next time they need a gift with a saved gifts feature. As shoppers browse gifts and categories, they can save any products that interest them into a file that’s there the next time they return. As a giftgiving occasion draws near, they have a record of the items they’d previously noticed. The system also can see the gifts saved and suggest similar ones from the database.
FindGift.com, with 2,000 product categories and more than 40,000 SKUs from over 1,000 marketers in its engine, takes data feeds from retailers. Human editors look for what catches their eye and the system gets the products into the right categories.
FindGift.com has operated on a PPC basis since 2001 after experimenting earlier with other business models. It’s a flat CPC rate of 11 cents to 20 cents per clickthrough depending on volume. Rather than a bidding system to maintain position within category search results as in Google and Yahoo’s paid search programs, product listings within categories are ranked according to popularity. Each night, FindGift’s engine reviews data from the previous day, recalculates and reorders the listings.
For retailers, Zakrzewski says PPC advertising on FindGift.com is typically more costeffective than advertising on a general search engine or a shopping search engine.
“For a lot of the common gift categories, Shopping.com may range from 15 cents per click on the low side to an average of about 25 cents a click, and upwards of maybe 50 cents a click for jewelry and watches,” he says. On FindGift.com, conversion rates support an average return of $8 to $12 for every dollar merchants spend with the engine, he adds, with more established retailers and brands coming in at the higher end of that scale.
But it’s not always the case that vertical search vehicles are less expensive on a costperclick basis than larger general search engines. “Niche engines could have a higher price because they do have a more targeted audience. So we focus our clients not on the rate they pay but on the return they receive,” Fathom Online’s McMahon says.
Greitzer notes the launch of a number of niche search engines over the past 12 to 18 months, spurred by marketplace opportunity. Not surprisingly, some are more effective than others. CowboyChuck.com’s Brown, for example, though pleased with his experience on FindGift.com, has gotten little or no response from campaigns on other gift engines.
“It’s relatively easy to build a search engine. But it’s hard to build a good search engine,” JupiterResearch’s Heisler says. “The development is complicated, and they tend to reply for the most part on merchants providing some type of feed. So they have to have an application programming interface that works. It’s not too easy overall to create a userfriendly experience, and one that works for both merchants and consumers.”