Online sales at DSW grew 23% in Q1.
(Page 2 of 2)
How much better? An e-mail newsletter recipient’s average order is $141, nearly 20% higher than the $118 of customers who receive product updates via mail. “We have developed a rapport with them by sending information that was not just a solicitation,” Whitley says. And the boost that Replacements gets comes at very little additional cost, he says. “It would cost us millions to send the newsletter in hard copy, but with the e-mail newsletter, there’s no increase in variable costs,” he says. “And we get great customer feedback.”
E-mail fills retailers’ desire to engage one-on-one with customers, Robertson says. “Retailers in the survey told us that communication and personalization of the message is important to them,” he says.
Another approach is to personalize the web page when the customer visits. That’s an approach that Cabela’s is taking. Cabela’s re-launched its site using Art Technology Group Inc.’s Consumer Commerce Suite and its Scenario Personalization product. By mid year, Cabela’s plans to serve up customized information when a registered customer comes to the site. The right column will feature the personalized material with a low-price impulse-purchase item at the top of the list, Sidner says. Based on Cabela’s experience with customized e-mails, Cabela’s expects the personalized features to drive sales. “With customized e-mail, we see a higher percentage of customers willing to open them, a higher percentage buying and a higher average order,” Sidner says. “We expect to see the same kind of incremental improvement on the web site.”
While simple, such techniques as sending communications to opt-in customers or delivering personalized pages to customers who have registered are good ways to start the CRM process and make an entire project easier to get off the ground, analysts say. “You can fall prey to having too much information; it becomes unmanageable and difficult to start anything,” says John Ripa, product leader for InfoBase eProducts at Acxiom Corp.
Starting simple also addresses the issue of customer data—how does a retailer gather it, make sure it’s correct and keep it fresh? “That’s the toughest part for multi-channel retailers,” says Doug Clare, vice president of global retail for data analytics company Fair, Isaac and Co. Inc. “It’s difficult to assess who the customer is because 90% of customer interactions are anonymous.”
That’s where web purchases, catalog purchases, proprietary credit cards and loyalty programs come into play. Retailers often have a surprising amount of information about customers, but are unable to use it because it does not reside in a single data base. “The data are still mostly siloed,” says Jennifer Kemp, director of global retail for Fair, Isaac. “You need to get a view of the customer data that spans the channels. That’s obvious, but it’s a lot more difficult than it seems.”
While most agree that cross-channel sharing of data is a goal, not all agree that the information needs to be perfect before the retailer can make use of it. Some CRM services providers, such as Acxiom, apply information from other databases to customer information to create a broader picture of the customer. Replacements, for instance, knew its customers, but it didn’t know a lot about them other than their activities with Replacements. It gave a 100,000-customer slice from its data base of 4.5 million to Acxiom to learn such information, on an aggregated basis, as income and lifestyle characteristics, which magazines they subscribe to, what kind of cultural events they attend, how they entertain and so on. “The whole point was to understand what our customers want,” Whitley says.
Replacements previously had to gather such data by surveying customers, a process that was so time-consuming and costly that the company hadn’t engaged in a survey since 1993. That surveyed generated 2,400 responses at a cost of $26,000, vs. $2,600 for the Acxiom analysis, Whitley says.
Others argue that customer data will never be clean and so retailers shouldn’t waste time trying to make it pristine. “The question of data quality goes to the heart of the misunderstanding about what kind of data you need and what you can do with it,” says Stephen Brown, director of product marketing for Ascential Software Corp., which in April acquired data quality company Vality Technology Inc. “Data quality isn’t about data cleansing. And it isn’t about fixing bad or broken data. It’s about understanding the meaning of individual data values and their relationships to one another.”
To that end, Ascential uses statistical analysis and information theory to determine the likelihood that two pieces of information will result in a third piece of information that will allow the retailer to take some action. “We apply artificial intelligence and fuzzy mathematics to navigate these gray areas,” Brown says. “With every piece of corroborating evidence, my picture gets clearer.”
In fact, he argues, anyone who wants to make progress with a CRM program has to live with ambiguous data. “No matter how stringent your standards are, you will not be able to prevent abnormalities from entering the data,” he says. “If you try to prevent them, you make the whole process so onerous that you discourage anyone from doing it.”
The ultimate aim of any CRM program is to increase sales, so whichever approach a retailer takes must lead to action that achieves the retailer’s strategic goals. “There could be hundreds of offers that could be presented to a customer and it’s not always easy to know which ones to present to which customers,” says Jamie Fiorda, product line manager with E.piphany Inc.’s marketing solutions. “Make sure that the analytics you apply to the data are in line with the company’s goals.”
Once the retailer has presented an offer, it can capture the customer’s response to that offer, feed it into the CRM data base and analyze it. “The organization can then learn from that feedback and put new offers out for customers,” Fiorda says. Which the retailer then feeds into the data base. And that keeps the cycle of CRM data going round.