Mobile accounted for 25% of Ulta's e-commerce revenue during Q2.
(Page 2 of 3)
It does take more work to design the site initially, because the designer has to think about which elements will display on screens of various sizes, and where they will be placed. There is also an important limitation, especially for a fashion brand like Original Penguin that relies heavily on eye-catching images of models showing off its apparel. He's found that just scaling down an image to fit a smartphone screen often alters the focus of the photo and makes it less exciting. Instead, he must create new images for mobile screens, for example, cropping a photo with four models to show only three, but keeping the center of attention on the most important garment and model.
Justiniano also highlighted another key advantage of responsive design: there is only a single URL for search engines to index. That means all credit for traffic, inbound links, Facebook Likes, reviews and other activity that boosts a site's standing on Google and Bing goes to that single URL, rather than having some of that credit going to a separate mobile web site's address.
Original Penguin was one of many examples cited throughout the conference of how smart design helps with search engine optimization, or SEO, the art and science of moving an e-commerce site up on search engine results pages. This has become such an important consideration for web site designers that IRWD organizers this year for the first time offered a full SEO Day in advance of the main conference.
Speakers during the SEO Day explained that Google Inc. last year made more than 500 changes to the algorithm it uses to rank web sites in response to a search query, many of them designed to ensure that searchers reach a site with credible and original content.
For retailers, that means it's a mistake to use manufacturers' product descriptions, content that many other e-retailers may also use and that search engines may give little credence to, said Seth Dotterer, vice president of marketing and product at Conductor Inc., which specializes in SEO. "If you are just pulling in content from a fire hose that everyone else is drinking from, that doesn't work," he said.
On the other hand, retailers may own valuable content that they don't think to post to a site that sells to consumers. For example, HSN Inc., which sells online and via TV shopping shows, found posting press releases about the company's financial results—documents that Google deems authoritative—benefited search results for HSN.com overall. But be sure to get approval to use that kind of material from everyone involved, including the legal team, finance and public relations, Hugo Guzman, HSN's senior manager of online marketing, advised. "A lot of folks are tempted to run with content, but you need to get buy-in or you are going to run into trouble," Guzman said.
When to break the rules
Checking with top management before posting a financial press release is good advice for all e-retailers, and IRWD speakers provided many other examples of principles that apply to almost all e-commerce sites. For example, make it easy for site visitors to find what they want by providing a large, prominent search box, suggested Bill Albert, executive director of the Design and Usability Center at Bentley University, in a session on identifying and overcoming top usability challenges.
He also advised against using terms that consumers might not understand. A retailer's buyers may think of a pair of petite skinny jeans as "ankle length slim fit denim," but shoppers likely won't search that way. "Avoid jargon and overly stylized terms," Albert said.
How shoppers search depends on who they are, said Albert's co-presenter, Ryan Hennig, vice president of marketing for Miles Kimball, which has brands that cater to older consumers. Those shoppers might call a "bra" an "undergarment," and so the retailer makes sure its search program is built to understand such terms, Hennig said.
That's an example of tailoring design to a specific consumer segment, and sometimes doing that leads an e-retailer to break with convention.
Few retail sites, for example, use brown as the background color on its product pages, but Seventh Generation, a web retailer of natural products, tested brown and found it converted better than a white background. One reason: many of its products come in white containers and don't pop out on a white page. Making that change, along with several others in a recent site redesign, including a clearer Add to Cart button, has boosted its site conversion rate from under 4% to 18%, said Reid Greenberg, director of e-commerce at Seventh Generation.
Test and test again
Greenberg discovered the power of brown by testing, and IRWD speakers reiterated throughout the conference the importance of testing all elements of site design. And they pointed to inexpensive testing tools that can yield big results.
Adrian Salamunovic, co-founder of web-only retailer CanvasPop.com, listed several tools he uses. They include UserVoice.com to get feedback from customers and to track complaints, with pricing starting at $20 per month per agent; UserTesting.com, which recruits consumers to review a site and provides videos of their sessions and written answers to retailer's questions, for $39 per tester; and Baymard.com, which for $150 will check 63 elements on a retailer's checkout pages, such as the correct formatting of fields for collecting credit card and ZIP code information.
Retailer Crate & Barrel, a store and web retailer of contemporary home furnishings, uses Optimizely, a tool designed to let retailers test web page elements without rewriting the underlying code. Optimizely starts at $19 per month for a site with 2,000 monthly visitors and goes up in price based on traffic.
The tool has made it easier for Crate & Barrel to continually test its site, and the results often surprise, said Joan King, the retailer's e-commerce director, during a full-day workshop on site design that preceded the main IRWD conference. For example, the retailer thought adding product ratings to category pages would boost conversion; instead, it reduced it 10% in a test. The idea was shelved.