Does the idea that your competitors are working on a dynamic merchandising strategy have you thinking you should be doing the same? But have thoughts of the likely IT hassles and the sheer complexity of the task stopped you from doing much about it? If so, you’re not alone.
Most online retailers say dynamic merchandising — in which the Website automatically makes personalized product recommendations — is important in part because they believe their competitors are working on it, yet few have implemented a full-bore dynamic merchandising strategy themselves, according to Darien, CT-based JupiterResearch.
Instead, some companies offer product suggestions based on the purchases of other people who bought the same item; Amazon.com was one of the first to adopt this practice. Other online merchants make product suggestions based on the customer’s past purchase behavior and current site behavior, such as what is in the shopping cart or has been searched. Handbag and luggage merchant eBags, for example, has for the past month been making automatic cross-sell recommendations based on what the customer has put in his shopping cart. When the customer clicks “add to cart,” the site responds with a “May we also suggest?” page giving the buyer a series of related offers. Displaying products the customer last viewed or bought and making product suggestions via e-mail based on past purchases are other examples of dynamic merchandising.
Yet none of them constitutes a so-called black-box or fully automated approach, according to JupiterResearch. “Other than the way Amazon.com does it, which is predominantly through collaborative filtering, there has not been widespread adoption of dynamic merchandising among retailers,” JupiterResearch analyst Patti Freeman Evans says.
When asked why they have yet to fully implement dynamic or automated merchandising, merchants typically cite the usual reasons: lack of IT and fiscal resources and the uncertainty of a return on investment. An additional stumbling block is the difficulty in setting up and maintaining the algorithms to produce appropriate site suggestions.
Moreover, until very recently it was rare that marketers’ online units had any control over the merchandising mixes on their pages, says Matthew Roche, chief executive of Offermatica, a San Francisco company that offers online sales testing and optimization services.
“Up to now, if you’re a multichannel merchant, the rule has been ‘take what you sell offline and sell it online,’” Roche says. “And if you don’t have control over assortment, you’re not a merchant.” He adds that too many marketers put resources into the “personalization rat hole to try to create a one-to-one experience, which is a load of bull.”
That said, Roche does not believe automated merchandising is a pipe dream. “We’re in a real-time market where the market is changing constantly,” he says. “Having merchandising techniques that respond faster than your weekly merchandising meetings works.”
For example, during the week leading up to the final episode of the television show Friends, a Friends-branded board game became a popular seller on the Website of one of Offermatica’s clients. “Because people were buying it, it crawled up the [site’s] best-seller list. When the Friends final episode happened, immediately there was no more interest in the game, and it crawled down the best-seller list and disappeared,” Roche says. “They [the merchant] did not know this had happened until the end of the month when they saw their sales.”
Dipping a toe
Anecdotal evidence trickling out of the online retail industry indicates that more marketers are finding that dynamic merchandising holds promise even if implemented only in manageable bits. Many merchants have recently invested in technology that includes dynamic merchandising functions, says Evans, but they have yet to take full advantage of them.
One case in point is Manchester, VT-based outdoor gear and apparel merchant The Orvis Co. After three months of testing, the company has begun making merchandising changes on key home and shopping pages on a daily and weekly basis — instead of monthly — using information gleaned by two staff “merchant analysts” about what is selling and what customers want to see. To help customers find the items they want, the site offers a search tool from Endeca, which also provides personalization capabilities. But Brad Wolansky, Orvis’s director of e-commerce, admits that the company isn’t taking full advantage of the technology.
“Delivering an experience based on what [customers have] done before or what we think they’ve done before or what they’ve conveyed to us their interest…is not like sending a spaceship to Jupiter,” Wolansky says. “It is in the realm of our current capabilities, but it requires some active analysis so that we know what we’re doing. It also requires having the folks in place to administer a very rapidly moving set of parameters. The issue isn’t so much ‘How much would it cost to get started?’ The issue is ‘How much would it cost to maintain?’”
Outdoor apparel manufacturer/marketer Patagonia is another Endeca customer that has tested dynamic merchandising and seen enough positive results to conclude that it’s worth pursuing but has yet to use all the tools at its disposal.
“In any situation where we have done some dynamic merchandising, we have seen increased sales and increased average order sizes,” says Kevin Churchill, director of merchandising for the Ventura, CA-based company. “Our site is set up so well that if you’re looking for a particular product, you can streamline straight to it, but what it’s not doing that the catalog does so well is offer all the products around it and allow the customer to shop.”
To that end, Patagonia is overhauling its site and will unveil a new one in late spring. “That’s where we’re really going to leverage the tools that we have,” Churchill says.
While other merchants dip their toes in the water, San Francisco-based apparel giant Gap looks to have jumped in headfirst. Gap took the Websites for its three retail brands — BananaRepublic.com, OldNavy.com, and Gap.com — offline on Aug. 24 for several days for major overhauls. The sites now include features such as real-time inventory so that shoppers won’t be shown items that aren’t available, an automatic upsell and cross-sell feature that makes recommendations according to what is in the customer’s shopping cart, and an “outfitting” feature that allows shoppers to click on an outfit and see all the items that make up the outfit on one screen. No word from Gap yet as to how the effort is panning out.
Getting ready
If Orvis and Patagonia seem to be on a path toward higher average order sizes fueled by personalized Website experiences, it is in part because they both have a parametric search function, which allows consumers to drill down to specific products based on attributes they specify. A good onsite search function is imperative for any merchant planning dynamic merchandising based on site behavior. After all, if the site’s search function is failing to give customers the results they want, any “related offers” will be off target as well.
“It’s called guided navigation,” Wolansky says of Orvis’s search function. “Instead of throwing 40 thumbnails at them, we’re guiding them to where they want to be. We’re basically helping them narrow down their search and guiding them along the way. In that process, we can make choices about what we want to show them. We’re helping them organize their thoughts.”
JupiterResearch’s Evans says that before online merchants consider dynamic merchandising, they should focus on the basics, one of which is having a parametric search function: “Helping customers narrow down based on what criteria are important to them is a key factor in helping customers make purchasing decisions.”
If you need to implement parametric search, Wolansky recommends going to a vendor rather than building the search function inhouse. “I continually hear about people who try to do search inhouse, and I can’t fathom that at this point,” he says. “The complications involved in building a search tool are myriad.”
Besides the proper search tool, multiple product images are of paramount importance when introducing dynamic merchandising. “Consumers want to see the product. They want to be able to evaluate it,” Evans says.
Evans also advises honing your site navigation before venturing further. “Make sure your navigation structure is clear and organized based on how a consumer thinks about it, not how you think about it,” she says. For example, where the industry may refer to products for the home such as towels and washcloths as “domestics,” most consumers have no idea what that means. “Often something is very clear in the retailer’s head, but consumers don’t understand it,” Evans says. “They [retailers] need to watch how their consumers are behaving to understand if their navigation is clear.”