(Searchline) Do you know the difference between a tureen and a terrine? Or between a casserole and a cassoulet? The answers will come below, but if you know without being told, you probably don’t need the kind of navigational help that tabletop maker Mikasa & Co. has added to its main Website recently. And good for you.
For the rest of us who can’t tell a demitasse from a demiglace (one’s a coffee cup, the other’s a cooking sauce), Mikasa has simplified its online merchandising by incorporating the new “learning navigation” site search platform from SLI Systems.
The platform is actually SLI’s version of what other site search providers have called “guided navigation”, a mechanism that learns to deliver more relevant listings based on what consumers have searched a site for in the past.
A dynamic Web-hosted site navigation service, SLI Learning Navigation actually builds on the company’s Learning Search product, which helps guide external search—that is, searches done off-site on Google, Yahoo!, and the rest—by tracking the keywords that users plug in most often to get to a merchant’s product pages. Learning Search also aggregates past search behavior to come up with alternate terms that let users refine their searches and get to the product they want more quickly and surely.
Leaning Navigation picks up some of those jobs at the Website’s edge. It lets retailers group their range of products into natural categories or “facets” to show similar products together. That’s not always an easy task, says SLI CEO Shaun Ryan, especially when a retailer has a large number of SKUs and wants to give visitors the option to search by categories such as men’s and women’s fashions, color, size or brand.
The platform serves up results for these filtered or faceted searches in order of items’ clickthrough popularity with previous visitors. In this way site search can help with product merchandizing by learning from and then delivering the products other customers have most often clicked on.
What’s more, the navigational pages created by the SLI platform are optimized with URLs, links and titles that Ryan says will make them more crawlable by search engine spiders. That will tend to make the pages findable in a general search for, say, “men’s brown boots”, and raise their ranking within the relevant organic results.
Designing those navigational pages with search discovery in mind is a unique focus for SLI’s platform, Ryan says. It stems from a drive to keep the navigational page URLs as simple and parameter-free as possible. For example, a shoe Website might offer a page of men’s loafers in black and in EEE widths. But a carelessly built URL might confuse that page with one that filtered products by EEE widths, then black, and then loafers. To a search engine, those pages might seem duplicates of one another, and most engines would then discard one of them.
In his blog, Ryan cites a URL for a Home Depot navigation page that contains miles and miles of parameters: “http://www.homedepot.com/prel80HDUS/EN_US/diy_main/pg_diy.jsp?CNTTYPE=PROD_META&pos=n10&MID=9876&com.broadvision.session. new=Yes&N=2984+6633&CNTKEY=misc%2fsearchResults.jsp”. Navigational pages with URLs this complex are just about invisible to search bots. Instead, merchants have taken to producing “product map” pages specifically for the crawlers that contain all the product links from the page in a stripped-down format. But the SLI Learning Navigation platform creates searchable pages automatically, according to Ryan, going as lightly as possible on the addresses so as not to confuse the spiders.
“It’s important to make it as easy as possible for the search engines to spider these navigational pages for their indexes,” he says. “That involves keeping the parameters as simple as possible. We also show the filtering selections in the title of the page.” That title can then more closely reflect the actual content of the page—something else that search bots usually give high marks for, thus improving a navigational page’s ranking in search results.
Like installing SLI’s other platforms, implementing Learning Navigation is almost a turnkey process because it’s a Web-hosted service. The search and navigation functions are hosted in SLI’s domain using product data feeds supplied by the merchant as often as required. To make sure that data is up to date, the system can be set up to query the merchant about item pricing and availability at regular intervals, up to and including the moment the site search is being conducted. As with SLI’s Learning Search service, merchants are charged according to the number of pages served in a given month.
It was largely that ease of adoption that led Mikasa & Co., the tabletop division of ARC International North America, to opt for SLI Systems’ Learning Search and Learning Navigation platforms. The company was planning to launch its first print catalog last spring and knew that the book would drive traffic to its slightly outdated Website. According to director of Internet operations Stephen Henderson, Mikasa also knew that the off-the-shelf site search it was using as part of its e-commerce platform was not going to be able to meet the needs of those new customers.
“We knew a lot of visitors were going to be searching by item numbers and by details of the catalog, and we had to be able to offer them a good experience finding what they were looking for,” Henderson says. The site search then in place on Mikasa & Co.’s pages was slow and relatively hard to use, so Henderson and his team went looking for a better solution.
“We looked at a number of the search providers out there, but I was impressed with the logic that was built into the SLI platform,” he says. “It felt like exactly what we should choose to make a quick change from the solution we had been on.”
The company’s main site, www.mikasaandcompany.com, contains about 6,000 live SKUs in dinnerware, glassware, and home accessories, but Mikasa’s e-commerce inventory contains about 60,000 items altogether, including a backlist in older and discontinued china patterns. Those items are organized into product categories such as formal china, everday china, stemware, flatware and service pieces, and then further filtered within those categories into some 250 style collections.
Users can search the site’s inventory by those categories and collections. But with the SLI learning navigation platform in place, they can also search by brand (the site offers both Mikasa and Studio Nova), by material (stoneware or bone china, for example) and by price range. In each case, the product pages default to popularity rankings, also users can re-sort to show items from highest price to lowest or vice versa.
Visitors can also search the site for “unique gifts”, and can call up “best sellers,” “top searches,” and clearance items.
“If most of the visitors looking for our Italian Countryside collection are clicking on the dinner plate, then that item will come up first in any search on the term ‘Italian Countryside,’” Henderson says.
Mikasa has seen a lift in the number of search engine referrals to the site, but Henderson can’t positively credit that traffic increase to the optimized pages from the Learning Navigation platform. Mikasa is also deploying SLI’s Site Champion optimization technology, which automates the creation of “related search” keyword links. Those links also serve to make pages more crawlable by search bots, and may be responsible for some of the increased search engine traffic.
And by the way, a tureen is a large soup serving bowl, while a cassoulet is a French stew. And casserole and terrine can refer both to foods (terrine’s a sort of pate) and the dishes they’re prepared in.