Tearing through systems to scrub dirty data and gaining a cultural understanding of names across the globe is no easy task.
For the rest of the July 2005 issue of CRM magazine please click here
Shakespeare may have been right that a rose by any other name would smell as sweet, but that doesn't apply to your database. Using the wrong name flat out stinks, and could cause you to lose a customer and money. Names present lots of problems, especially as companies strive to break down the walls and share information throughout departments. Emerging technology, however, can solve database issues with varying degrees of complexity, from dirty data and duplicate entries to life-stage changes, even to pinning down ethnicity. "Know thy customer is an imperative," says Lou Agosta, an independent technology consultant. "The real question is, how do you know? You can't do it with three-by-five cards--automation is necessary. What I can promise you is blood, sweat, and tears. There's a lot of work to be done."
For starters, businesses should make sure that mistakes are avoided from the beginning, and should understand that maintaining a clean database is not a one-time project, but a task that regularly performed will pay off in the long term. Data must be cleansed. "Sixty percent of data quality problems arrive from data entry errors," says Chandos Quill, vice president of marketing communications for Experian. Data hygiene tools like Experian's assure information are accurate and high quality. Its Customer Data Integration technology, Truvue, assigns a unique ID to a customer. The ID never changes as Experian tracks that person's history over several years. The technology links this information and delivers it to all points of customer and prospect information. It consolidates duplicate records, and eliminates inaccurate ones.
However, just because data is up to date and accurate doesn't mean it will stay that way. A lot of executives see information cleansing as a one-time-only activity, but it should be an ongoing and automated process, according to Jill Dyche, a partner at Baseline Consulting. "The fact of business these days is there's always new data coming in." Customers are constantly moving, getting married and divorced, and changing their names. Because of these common life changes, every month 2 percent to 5 percent of customer data will be wrong, according to Tony Fisher, president and general manager of DataFlux. And at only 2 percent per month, more than 20 percent of customer data will be dirty within one year. That's why like teeth, data needs to be regularly brushed clean.
Data cleansing tools can help--Dyche cites Firstlogic's offerings because of the products' price points, and notes that Experian, DataFlux (a division of SAS), and Ascential (recently acquired by IBM) are leaders in the data scrubbing arena. "There's an assumption that throwing people at the problem is cheaper. Nine times out of ten that's not true," she says. "Use the human for more high-impact work, as opposed to the brute data cleanup."
So who's in charge of keeping track of all this information? Typically it's the marketing team, but the best organizations understand that the need is repeated across all departments. Some companies even go as far as to hire data-quality strategic architects dedicated to the challenges of understanding the data and employing the right tools and training to deal with it.
Emerging technology can help follow individual life-stage changes like marriage and divorce. Jeff George, senior director for data services at Harte-Hanks, emphasizes the massive leap he sees in direct marketing technology used for such purposes. "It used to be that if you had a merge-purge that was customized, that was a competitive advantage," he says. "Now it's survival. Smart marketing is essential." Smart marketing also means having the right data and using it correctly. Botch that and marketing efforts can be downright offensive to customers and embarrassing for the company. Group 1 Software's Todd Schwartzrock, director of product management, enterprise data quality, cites an example of a Big 3 automaker with some 50 million records. A male employee there named Jan-William complained that he receives correspondence addressing him as a female. The mistake could be based on a CSR's perception of the name or on database technology that identified him as such. Either way, this is a horrible impression to make. Arguing from Jan-William's point of view, Schwartzrock says, "You want me to remain a customer and you don't understand that I'm male?" Good luck making money off Jan-William.
A SINGLE VIEW
"The good news is, people are consolidating and integrating their customers' data better than ever, getting one view of the same person over [multiple channels]," Dyche says. "Parsing and name matching is done much better in the automated way. The name-cleansing pieces of these systems is in a lot of ways the hard part. If you can master that, everything becomes easier."
Agosta says when his colleagues ask enterprises how many sources of data they have, the average company acknowledges at least a dozen. Some say they stopped counting. "It is almost without question [that] most organizations have a significant data quality issue with trying to have that one view of the customer," Schwartzrock says. "Understanding each of the stakeholders is critical to making an organization efficient. These are challenges we see every day. Most businesses can do better than they are today by employing data quality tools."
Robb Eklund, vice president of CRM product marketing for Oracle, agrees. "When you have various name repositories across an enterprise, the challenge is ensuring that every part of the organization acts as a steward of information. When the marketing department determines that someone prefers to be called Miss, or that a customer's name has a unique pronunciation, they need to be able to push that information across the enterprise so that everyone from customer service to accounts receivable is made aware. Too often, this valuable data rests in whatever silo it originated from." Tools like the Oracle Customer Data Hub allow companies that run disparate systems to get a single source of the truth.
When companies start to integrate and consolidate information from various parts of an organization into a single view, several sources of data are in conflict, according to DataFlux's Fisher. One is right, one is wrong, but how do you determine which is which? "We generate match codes to come up with the most baseline definition of who this person is," Fisher says. "You still might have it wrong, but you've put together what you consider to be the most accurate information."
Two years ago 90 percent of problems were tactical, Fisher says, but now it's more 50-50: "Half of our engagements are more strategic. The problem is more about identifying the customer over multiple lines of communication and departments. It's about customer identification to provide better service." Even from a marketing perspective, the impetus for removing duplicate and often inconsistent customer information, a process known as deduping, is changing. "Deduping isn't really about deduping anymore," George says. Before it was a cost-driven decision. "Now it's about building a complete view of the customer, creating relevant messaging, and ultimately, increasing ROI."
As channels and relationships expand there are more sources of data. The solution, according to George, is to standardize the data internally. "A lot is, roll up the sleeves and let the software do its magic," Agosta says; however, "this is not a silver bullet, but more of a tool kit." It's not just about the tools, either, George points out. "It's the software, certainly, but it's also the data to back it up. Fill in the holes and use software to identify and build a view of the person. You do it with hard rules and soft rules. There's no off-the-shelf solution."
Dealing with English-language names can be difficult enough, but start throwing in foreign names that are Americanized, or trying to understand personal titles used in different cultures as companies leap into globalization, and things get even messier.
Technology is helping to minimize these uncomfortable mistakes as well. Jack Hermansen, CEO of Language Analysis Systems (LAS), cited some cultural examples LAS technology can reveal using its name reference library, which sorts through roughly a million names per second. The bearer of the name Haj has made a pilgrimage to Mecca; Efendi means teacher in Farsi and is a title of respect; and Md. is the abbreviation for Mohammad in some cultures, which English language can read as Dr. "It's almost impossible for one human being to know all these things," Hermansen says. "That's why computers are helpful. Data administrators typically don't get a lot of support from management to get more money for their data. But now they can have metrics to show they are getting some return on investment."
Tools from companies like Geoscape can break names down to specific ethnicities based on street blocks; TeleTech provides call center interpreters in 150 different languages to make clients across 44 countries who don't speak English more comfortable with their transactions. Companies like Siebel and Oracle are also starting to show interest in globalization tools. "The globalization of business is so clear to executives these days," Schwartzrock says. "The world is flat--it's a level playing field. You need to be able to speak effectively to customers no matter where they live."
Many marketers don't understand that they Americanize everything, says Tom MacDonald, TeleTech's executive director of real-time translation service, InCulture. "They tend to think in the way we speak, [and] spend a lot of money [marketing to the wrong] people." Most marketers will just buy a list and market based on surnames. People can be very proud of their cultures, and there are differences between Latino-Americans and Mexican-Americans, for example. MacDonald can relate to this firsthand: It irks him when people misspell his name without the first "a." Mac is a prefix for a Scottish name, and he and his father are very proud of that heritage. Mc is an Irish prefix. In other languages some errors can be insulting, he says: "To me, it says you didn't take the time to understand. Something like this goes beyond business and gets very emotional."
A business may call a customer by any other name than her own and she may still purchase its products, but that customer, without a doubt, has noticed the error. If that same organization gets the customer's name right and can track her life changes over a period of years, it can offer her products relevant to her family and/or culture. She'll feel that the company recognizes her, and she will have more reasons for continuing a long, rewarding, delightful relationship.
Contact Senior Editor Alexandra DeFelice at adefelice@destinationCRM.com
|Learn more about the companies mentioned in this article in the destinationCRM Buyer's Guide:
Translation software makes inroads where rapid and accurate customer communication is literally life-or-death.
With incomplete information, some hazards can be deadly.
Sponsored By: Marketo and Real Magnet
Sponsored By: Jacada, Avaya, Confirmit, inMoment and BoldChat
Sponsored By: Genesys, Avaya, Verint, and Aspect
Sponsored By: Informatica