Twitter’s algorithm changes likely mean fewer consumers will see a brand’s tweets.
It takes nothing more than lunchtime at McDonald’s to understand the power of cross-selling and up-selling. The fast-food giant has for years successfully increased order size by offering a complementary product-fries-or an upgrade at the point of purchase (“Can I supersize that drink for you?”).
Translating those concepts to the web, though, is a bit more complicated. When no human sales associate is on hand, technology must take on the task of presenting complementary products and upgrades. Much of that technology has only recently emerged, but retailers are already on board. More retailers than ever are implementing cross-selling and up-selling features on their web sites-62% and 46%, respectively, of 100 top retailers recently surveyed by the Direct Marketing Association and e-retail consultants The E-tailing Group Inc. in Chicago.
By category, online retailers of computers, books and music, consumer electronics and pet supplies are out in front in cross merchandising. 100% of the surveyed sites in those categories had implemented cross-selling features, while department store web sites followed close behind at 86%, along with online retailers of toys and games at 75%. E-retailers of sporting goods lagged far behind; only 25% had cross-selling features.
Up-selling, less established across the board as an online merchandising tool, nevertheless is a feature on 100% of book and music sites, 80% of computer sites, and 75% of sporting goods sites surveyed. Though there’s little aggregate data on exactly to what extent cross-selling and up-selling online boosts retail sales, it makes intuitive sense, says Lauren Freedman, president of The E-tailing Group. “Almost any merchant will tell you that when they merchandise their cross-sells and up-sells, they sell more product and increase their order size significantly,” she says.
As generally defined, cross-selling presents another product to the shopper either at the point of purchase or while an item is still under consideration. The second product is selected for presentation because there’s some affinity between the two products or between the person making the purchase and the product. Up-selling offers another product as well, but with a twist: it’s an enhancement to the product being considered or in the shopping cart. The goal of either practice is to sell more product, and though the strategies differ, relevancy in some dimension, whether between products or between shopper and product, is key to both.
To a large extent, determining relevancy is still the domain of human merchandising expertise on most web sites and it’s an art. “It take a lot of hard work to put up relevant cross-sells and up-sells,” Freedman says. “It’s time consuming to take every product and figure out what’s going to sell with it.” But the flip side is that without relevancy, consumers don’t even look at the other offers, much less buy.
The systems used to come up with alternative product recommendations online range from plain old common-sense merchandising to rules-based programs or analytic software, or a combination of the three. Home and garden retailer Smith & Hawken, for instance, wanted to boost its online sales performance and give its online customers more personalized service. In a three-month head-to-head test just ahead of garden season, the company compared online shopping transactions that offered up-sell and cross-sell product recommendations generated by Net Perceptions Inc.’s analytic software versus those that did not. The Net Perceptions software-driven product recommendations resulted in orders that were on average 16.5% larger; they also increased the number of items in each order by an average of 60%. E-commerce sales at another Net Perceptions client, MusicianFriend.com, the online subsidiary of music gear retailer The Guitar Center, rose 170% after MusicianFriend.com used the analytic software to recommend additional products based on customer tastes and preferences.
Rules-based vs. analytics
The technology underlying Net Perceptions software solutions for cross-selling and up-selling incorporates both ruled-based and analytically-driven personalization, a combination still not widely used among retailers. Rules-based recommendations are the more common approach among online retailers today, according to Kevin Scott, customer applications research analyst at AMR Research Inc. “Rules-based recommendations say, if I buy tan chinos, show me the red sweater. Analytics-based recommendations say, based on what I know about this customer and his resemblance to other customer groups, what product offer is he most likely to accept?” Scott says. “Retail today is still more rules-based. It’s a cheaper scenario and you don’t have to have a Ph.D. in statistics to know whether your models are working.”
Rules-based recommendations may be cheaper than technology, but the pay-off may be that retailers miss opportunities that their merchandisers may not be aware of. “Though a rules-based approach can be effective, it’s still limited to what the marketer can think of,” says Greg Girard, vice-president of retail applications strategies at AMR Research. “You can only write rules you can imagine. Analytics can find links you might not think of.”
Net Perceptions estimates that a smaller retailer can get in the door starting at about $300,000 for analytics software and services to generate cross-sell and up-sell product recommendations. And though that’s a small tab compared to the $2 million and up typically paid by larger retailers with heavier traffic and more volume, it’s still not in the budget at a retailer like eBags. The luggage and travel gear merchant has a policy of doing everything as cheaply as possible on the back end without scrimping on the customer experience. And that goes for online merchandising as well; eBags wrote its own rules-based cross-selling and up-selling applications in-house.