Alibaba’s Tmall Global now features goods from 14,500 overseas brands, 80% of them selling in China for the first time.
A new generation of web analytics tools is helping retailers boost sales by pinpointing where they need to make improvements to the shopping experience.
Not long ago, the operators of e-commerce sites would trumpet major site overhauls with the sort of fanfare that more often greets, say, the year’s new car models rolling off the assembly line in Detroit. A big batch of much-needed site improvements was heralded as a re-launch, frequently accompanied by corporate predictions of Big Things to Come.
But the e-commerce world has turned on its axis a few times since then. Remember Bloomingdales.com? A scant year ago, it was tooting its horn about a site redesign aimed at making a killing online during the upcoming holiday season. By the end of November, when those expectations didn’t materialize, it pulled e-commerce from the site.
Now, web merchants are more circumspect about broadcasting wholesale site redesigns. They’re discovering that small improvements can make big differences, thanks to new tools from a new generation of analytics software that pinpoints web site trouble spots and then validates the effects of change. Where site success was once calculated solely in terms of hits, traffic and other measurements rolled up into server log files, the new analytics tools track shoppers’ paths throughout a site. That yields critical intelligence at the page and even the product level to pinpoint small but specific opportunities for improvement. And the impact of the resulting site alterations is totaled up in higher conversion rates, lower cart abandonment, better uptake on offers-any action for which completion means success.
Conversion rates up 83%
Hewlett-Packard Co.’s web site, for example, has undergone major transformations behind the scenes to integrate Compaq Corp.’s web site following the merger of the two computer companies in May. But it’s been some of the tiniest site alterations that are producing some of the most interesting and immediately measurable results for the combined companies’ web site, HP.com
One of the changes was a site improvement implemented by the former Compaq which was so successful that other parts of the combined site are being evaluated for similar treatment today. With one minor change to its navigation, Compaq increased conversions by 83% and revenues by 25% for accessories to go with the Compaq iPAQ Pocket PC.
To get the lift, Compaq shortened the path between product presentation and the opportunity to buy the products. Analytics software from Keylime Software Inc. flagged an element of navigation that was creating “dispersion” on a page, says Hewlett-Packard director of research and e-testing Seth Romanow. Site visitors who were apparently following the path toward the purchase of iPAQ handheld accessories instead scattered when they reached the accessories page.
Analytics suggested that the navigational pathway was the culprit. Shoppers who were buying iPAQ handhelds would be presented with information about modems, for example, on the iPAQ product page. But when they clicked on a modem to buy it, the site presented shoppers with the whole list of accessories for consideration instead of delivering them directly to the modem, so they could drop it into their shopping cart and then check out.
“By taking that link out and directing people more specifically into the online store for a particular product, we were able to increase conversions and revenue, from a very small change that didn’t create a cost,” Romanow says. Hewlett-Packard is applying the same type of pathway analysis to pages and links throughout its e-commerce site, he adds.
As Hewlett-Packard’s experience underscores, it’s a new game. Server log files until now have been the basis of e-retail metrics. They show numbers but not pathway information, which is key to uncovering web site pile-ups in need of fixing. Now, analytics are providing that level of detail.
In-stock or out?
In January, Finish Line Inc., a retailer of athletic shoes and apparel, signed on with Buystream Inc., a provider of hosted analytics applications. It was the retailer’s first foray beyond log file data into clickstream analysis that follows visitors’ travels through the site. E-commerce director Kent Zimmerman says the tool proved its worth within three months, when Finish Line encountered problems in communicating product availability to customers while integrating a new inventory database that gave it much greater visibility into stock levels across channels.
The problem was not that incorrect information on availability was being given to shoppers, but that it was being presented to them in a way that led some to the wrong conclusions, such as that when a single item they’d placed in the cart was out of stock, all items in the cart were out of stock. “We realized that the wording and the point in time at which we were displaying an inventory availability message to the shopper was creating the wrong impression,” Zimmerman says. By changing the wording and moving it to a more appropriate place in the process, Finish Line reduced its shopping cart abandonment rate significantly, he says, though he won’t disclose numbers.
“There are industry averages on the number of people you’ll lose at any step in the checkout process. If at any given time, you lose more than is normal, that should raise a red flag,” he says. “We saw something that looked like a red flag, and the tool dug into it.”
Zimmerman adds that the reduction in the abandonment rate was immediate, and that the real-time feedback is a major benefit of the new analytics tools. “One of the good things about these analytics packages is that you can measure the effect of changes almost as soon as you make them,” he says. “If you want to make a change that’s going to impact your conversion rate by 0.2%, you have to see the real-time effect. You can’t wait to collect the data for a week and then go back and see if it has any effect, because that’s a very subtle change. We made a small change that we were almost positive would have an effect in reducing our abandonment rate, and the tool showed we were right.”