Life is tough for the modern enterprise. There are many customer touch points, and customers want to be remembered, but not bothered, at each of them. CRM suites are supposed to help companies optimize products and offers to their customers, but sometimes it's all they can do to simply try to keep up with changes in customer preferences amid the
pressures of the marketplace.
Closing the distance between customer data and customer understanding are a flood of analytical CRM tools, being offered by everyone from traditional operational CRM providers to upstarts to long-term data mining specialists. The growing focus on analytical applications as a tool to directly guide customer interactions will grow the overall market for analytics from an already healthy $1 billion now to $2.2 billion by 2004. IDC analyst Henry Morris, who projects similar numbers, notes that the CRM-focused segment of this market has been growing at a rapid clip since 1997 when it accounted for only $90 million of the total analytical applications market.
Corporate number crunching, of course, goes all the way back to the salad days of IBM mainframes. Also, business analysts have relied on tools from traditional analysis vendors for some time to guide marketing and product development systems. That experience has been crucial as growing demand for demonstrable CRM results has pressured virtually every significant CRM
vendor to boast analytical capabilities, or at least interfaces. "Not many [pure analytical vendors] have a good understanding of CRM, so a lot of people are partnering specifically to provide this domain expertise," says Madan Sheina, principal business intelligence analyst for UK-based research service, ComputerWire.
As companies have grown comfortable with using software to streamline sales, service and marketing processes (sometimes referred to as operational CRM) and have now grown accustomed to the idea of extending some of these features and data to partners and suppliers (dubbed collaborative CRM), analytical CRM provides what some describe as the final, crucial link. "Simply having great bandwidth [to the customer] is of no use if you don't know what to say or who you're talking to," says David
Cody, director of marketing for the CustomerCentric Solutions division of SPSS. "Analytical CRM is that piece of the CRM triumvirate that makes [data] usable information within the enterprise and makes customers come back and say, 'Yes, something's different about the company.'"
Analytical tools can help reduce marketing costs. "With every interaction that occurs, I collect information, and that not only helps me understand that campaign, but when collected and put in individual profiles, helps me better assess what might work and might not work [in the future] both individually and as a cluster," says Paul Morris, vice president of the analytics business of Delano Technology, in Toronto. The ROI from analytical CRM processes ranges from eliminating waste in marketing campaigns by cutting down on overall frequency, as well as eliminating customers with an extremely low propensity to respond, to improving overall efficiency of service by harvesting inquiries prioritized by measurements of the customer's value to the company.
Even the hyped "real-time" analytics offered by a number of Web commerce firms can't change the reality that sales
and marketing is largely about rejection, but improvement is generally welcome. "We're still talking about pretty low [response rates], but even to move it up a percent can mean hundreds of thousands of dollars,"
says Nelle Schantz,
a global strategist for business intelligence software vendor, SAS.
The move to analytical CRM is a logical step for many firms, which for some time have viewed building data capacity and the framework for analytics as a must-have corporate investment. According to technology research and consulting firm, Meta Group, the average Global 3000 firm spent in excess of $3 million building and maintaining their data warehouses. Yet more than half were unable to link the spending with any distinct business benefit. "Currently, 80 percent of enterprises are looking to tangibly justify their [data warehouses] or trim expenses," writes Doug Laney in a May 2001 Meta Group report.
"CRM is getting a bit of a black eye, because companies invested in these applications, and companies are suggesting they're not seeing the payback they thought they were going to get," says Mike Schroeck, partner in charge of the customer analytics practice of PricewaterhouseCoopers.
As analytical applications have increasingly come into vogue with enterprise customers, haphazard buying has become a problem. Many companies license an array of
analytical tools, and the ensuing feature overlap may as much as double the necessary cost of ownership for analytics,
according to Laney's report.
In part, this may be fueled by a lack of understanding among the buying community of the appropriate scope of analytical applications. Sometimes the qualified statisticians who provide the ultimate quality control over the implications of CRM analysis are left out of the loop, resulting in unreasonable management expectations. "The concept of profitability is not something you measure on a 10-minute basis," says Chuck Teller, vice president of the enterprise performance management division of PeopleSoft.
"Analytical CRM isn't really for everybody," says SPSS's Cody. "You have to have a product or service where knowing [the customer] makes a difference, and gives you some kind of
advantage that will make you work smarter or better and please customers."
Making analytical insights and results available to a broad user population is crucial. "You want to put this information in the hands of people who are not analysts by job title," says E.piphany vice president of market development Paul Rodwick. E.piphany touts an architecture which it claims offers the ability to ask meaningful questions about data values that may not directly exist in the database (such as a 90-day average account balance) without the need for database skills. Several firms, including Rodwick's, heavily promote their artificial intelligence processes that search for patterns on demand, locating hidden demographic groups disinclined to buy a particular product, for example.
These neural network technologies are often used to recognize anomalous, potentially fraudulent patterns, but they are hardly perfect. Meta Group analyst Aaron Zornes learned this firsthand when his provider automatically disabled his mobile phone because he was dialing Sacramento, hardly an exotic port of call for a Californian.
TeaLeaf co-founder Randi Barshack says that companies need better Web analysis tools than site and transaction logs in order to gain proper insight into the cost of online errors. "It's not just knowing that it cost $5,000 to fix [the Web site], but that it may have cost $10,000 in lost sales and $20,000 in receiving incoming calls in our call center," she says.
"[Web analysis] used to be transactional data: I was looking to sell 200 TVs online, but I only sold four. That's the data people are used to," she says. "So I only sold four TVs, but I want to go back and look at the people that put TVs in the basket but didn't buy." The reason could be anything from erroneous pricing to a poor description to a misreported out-of-stock. The inventory management database may have gleefully reported that the system is backordered 12 weeks, but since a transaction isn't made, the error isn't normally caught by traditional evaluations. "It's about finding the customers with bad experience that didn't write you the nasty letter."
The in-depth evaluation of customer value that analytical solutions provide can save many organizations from making horrible mistakes. Schantz recalls working with a bank, which had once introduced a $25 user fee for its credit card. The fee was indiscriminate across the customer base rather than
targeting unprofitable bank customers, and in protest the bank lost a customer with a $750,000 mortgage and a $1.2
million investment account--foregoing tens of thousands of dollars in management fees and interest over $25 in recurring revenue. "They realized that their least profitable credit cards were their most profitable [overall] bank customers," she says. Even if apocryphal, the story illuminates the real danger of business unit decisions that make no effort to gain real insight into the overall impact of their actions.
Using Web transaction analysis software from San Francisco-based TeaLeaf Technology, Tower Records senior producer Lisa Scovel found that customers were loathe to take advantage of the company's Web-site browsing capabilities other than a keyword search, which among other things robbed Tower of a chance to present merchandising
campaigns as customers clicked through product trees.
At the same time, taking a careful look at the customer path through the site alerted Scovel to other problems, such as a restriction from a forgotten legacy system that prevented international customers from ordering large numbers of items from the Tower UK site. This had a direct negative impact on revenues, since international customers tend to make monstrous purchases to defray shipping costs.
While internal efficiency is often one of the major paybacks from conventional CRM initiatives, analytical specialists are of one voice in saying that they provide the real force behind results that both employees and customers appreciate. "Salespeople should be able to tell you yes, they spend less time with dead prospects. Customers should be able to tell you yes, they're still getting your junk mail, but no it's not 'as junky' as it used to be, or that the Web site is now particularly insightful," Cody says.
Critics have noted that CRM has so far failed to instill true best practices in the areas it seeks to serve, unlike other enterprise technology projects such as ERP. "If you look at CRM, there are no best practices on how you should manage
customers. [CRM vendors] give you tools, but no rules," ComputerWire's Sheina says. As analytical CRM matures, however, it may finally provide the market-tested insight companies are clamoring for. By packaging models that can clearly identify and track customer retention challenges and coherently segment customers into significant behavioral groups, he feels some long-needed order is finally at hand. "[Now] there's enough experience in the field to glean these customer benchmarks, and this is how you should measure your customer base."
SPSS, for example, has focused on delivering reusable
predictive models for customer churn. Paired with the ability to provide relevant information to the field, such as allowing sales representatives to identify those customers most likely to depart, analytics may also provide the final compelling argument for wholehearted participation in a CRM project, far better than a dense quarterly market analysis ever could. "Tell me which of my best customers are most likely to leave, and I'm going to sit down and call them. No sales guy is going to say 'I'm tired of being dictated to by Ph.D.s' when they're
getting that kind of information."