Business-to-business marketers are plagued by data problems. Business data is complex and fast-changing. Customers interact with us through a variety of channels, and often provide us with conflicting information. Our legacy databases are not as robust as we need. New tools and technologies emerge and must be evaluated. It’s a never-ending battle. To shed some light on B2B data problems, Bernice Grossman and I compiled a working list of problems and solutions. Here are some of the thorniest.
1) Data entered by our sales people ends up as mush in our database. They don’t follow the rules; or there are no rules. That may be okay for the rep, but it’s not okay for the company.
Here’s the best practice: Create a centralized data input group. Train and motivate them well. Give them objective rules to follow. Develop a simple method for testing the accuracy from this group as an ongoing practice. If this group cannot follow the rules, then the rules should be re-evaluated.
Then, develop a very simple process by which reps pass their data to this group. Dedicate particular group members to certain reps, so the input person builds experience about rep’s behavior and communication style. The bonus: these two parties will team, build a valuable relationship, work together well, and improve data quality.
Consider enabling the data input group with a real-time interface with a database services provider to prompt the standard company name and address. This can be an expensive, but very helpful, tool.
2) How do I match and de-duplicate customer records effectively?
Some approaches to consider:
- Establish—and enforce—data governing rules to improve data entry, which will keep your matching problems under some semblance of control.
- Find a solid software vendor with a tool specifically designed to parse, cleanse, and otherwise do the matching for you. Test a few vendors to find the one that works best with your data.
- Create a custom matching algorithm. As a place to start, ask several match/merge companies to show you examples of the results of their algorithm against your data.
3) When data elements conflict in my house file, how do I decide which is the “truth”?
The short answer is: by date. The most recent data is the one you should default to.
But also keep in mind when importing data to enhance your records that appended data will always have its limitations, and is best viewed as directional, versus real “truth.” Be careful not to build targeting or segmentation processes that are primarily dependent on appended data.
You could consider conducting an audit to validate the quality of your various append sources. (This is usually done by telephone, and it’s not cheap.) Then you can add a score to each appended element, based on its source, to manage the risk of relying on any particular element.
4) Which corporate address should I put in my database? There’s the legal address and the financial (banking) address, which may be different. Or there may be a street address and a P.O. box address. Equifax and D&B often supply the financial address. The address to receive proxies is different from the address to receive advertising mail. How should I sort all this out?
As a marketer, your concern is delivery. You care about a bill to and a ship to. Focus on the address where mail and packages are delivered.
5) Measuring the impact of each touch in our omni-channel world is driving us nuts. Any ideas?
The attribution problem has heated up recently, fueled by the rise of digital marketing. But it’s really nothing new. The traditional attribution methods of assigning the credit have long been either the first touch (the inquiry source medium) or the last touch (the channel through which the lead was either qualified or converted to a sale). Marketers are in general agreement today as to the deficiencies of either of these traditional methods.
Digital marketers are experimenting with various approaches to the attribution problem, like weighting touches based on stage or role in the buying process, or by the type of touch—attending a two-hour seminar being weighted more heavily than a content download.
6) How should I handle unstructured data, like social media content. All this “big data” stuff is getting bigger, and meaner, every day.
User-generated social media content may offer valuable insights into customer needs and issues. But marketers first must think through how they will use the information to drive business results. First you must develop a use case. Then you must develop a way to attribute the information to a record. For example, one method to allow the match is collecting multiple cookies to find an email address or other identifier. There may be situations where you want to track sentiment without attributing it to a particular customer but to a group, like large companies versus small. In either case, we suggest that you test the value of the data before you put a lot of time and money into capturing it in your marketing database.
You can find more thorny data issues and solutions in our new white paper, available for free download. Please submit your issues in the comments section here, and we’ll be happy to suggest some solutions.
About the author: Ruth P. Stevens consults on customer acquisition and retention, and teaches marketing at companies and business schools in the U.S. and abroad. Crain’s BtoB magazine named Ruth one of the 100 Most Influential People in Business Marketing. She is the author of Maximizing Lead Generation: The Complete Guide for B2B Marketers, and Trade Show and Event Marketing. Ruth serves as a director of Edmund Optics, Inc., the HIMMS Media Group, and the Business Information Industry Association. Learn more at www.ruthstevens.com.