How web analytics keep up with marketing at Ice.com

As retailers’ understanding of online marketing changes, so does their need for the tools that measure marketing success. Ice.com should know—in its five years of existence, it’s used four different web analytics packages in an ongoing effort to match what it wants to measure with its analytics package’s ability to measure it. That’s a lot of switching at a time when the average company has purchased 1.8 analytics packages, according to Forrester Research’s estimate, but Pinny Gniwisch, vice president of marketing, says each switch reflected Ice’s evolving needs as a marketer.

“At every stage of the company there’s a different emphasis on what’s important,” he says. “In the beginning it was about getting eyeballs. Then other things became important, like data capture, building up your e-mail list, search marketing. Every year there’s something else that’s the next big thing your company is going to tackle.”

Nine months ago, Ice.com switched to a software solution from provider Visual Sciences LLC as part of an effort to focus on conversions. “We had all this traffic coming to the site. We needed accurate data and we needed to be able to understand why, where and how visitors don`t convert,” says Gniwisch. Ice.com was looking for the ability to do A/B testing in simultaneous segments so as to push out winning promotions and messaging more quickly, and the ability to see campaign conversion and clickstream data in the same interface. Visual Sciences got the nod for both, plus its ability to take the analysis down to a very granular level.

For example, the online jewelry business experiences heightened fraud at certain times of the year, and though an internal fraud detection team at Ice.com catches a large number of those orders before they ship, it’s not before they’re initially recorded as conversion. “On the analysis, you’re converting at 3%, but if you could take out the fraud, it’s 1.5%,” says Gniwisch. “It becomes a question of going further with a campaign, or cutting the campaign. We needed the ability to get the data that represented real sales and push it back into the engine to churn out the real conversion numbers. We’re able to feed that data back into Visual Sciences to churn out another column of the actual conversion data.”

Taking the data down to that level of accuracy and the ability to do simultaneous segmented testing makes a significant difference in decisions on marketing spending, says Gniwisch. “Let’s say we’re paying $25,000 for a front page on MSN. Every point of conversion is going to make a huge difference because of the number of clicks you get,” he says. “On a campaign like that we have seen our conversion rate go higher based on the testing we have done.”