There’s an old adage that asserts, “The minute a printed directory hits the streets, it’s out of date.” It’s true, but the same can be said for any database, anywhere. Many users (and often management) are under the mistaken impression that if enough effort is expended, and the proper processes are applied, then the data will be 100% accurate. Unfortunately this can never be true.
According to the US Postal Service, more than 44-million Americans change their addresses every year. That fact alone suggests it is impossible to keep your database 100% accurate. And consider this: According to e-maillabs.com, for first quarter 2005, e-mail deliverability was just less than 85%. How can you have a “perfect” database when 15% (or more) of the e-mails in your system won’t deliver properly? The simple answer is, you can’t. As consultant Alan Weiss frequently exhorts, “Life is about success, not perfection.” The same can be said for your databases.
So what should you consider “successful”? That will depend on your customer base and other needs. For example, if you have businesses or other institutions as your members (trade assn), then you should be able to keep track of a very high percentage of those business addresses (more than 95%, in my opinion), because businesses move very infrequently. On the other hand, if your primary membership is college students, you can expect a much lower rate of accuracy, as most college students change addresses at least once per year, if not more often.
Having said that, here are some steps you can and should take to ensure the highest possible accuracy of your organization’s data:
- Centralize your data. In order to maintain the highest data integrity, there needs to be one source for data within your organization. I frequently tell the story of a former client of mine who managed to anger a board member whose address they could never get right, even though she had made several requests to correct it. The problem wasn’t bad data; it was bad data management. This particular assn was managing mailing lists in five different databases. You’re guaranteed to have problems when you have decentralized data.
- Build data collection processes into everything you do. For example, every time you “talk” to your members, be it through e-mail, snail-mail, on the phone or via your Web site, ask them to check their contact information and update it where needed.
- Create and regularly execute data integrity reports. For example, create a report that pulls all records that do not have an “@” symbol in the e-mail address field (it can’t be a valid e-mail address without the “@” symbol) and run the report regularly. Do the same for other fields or combination of fields (e.g., ZIP codes for US addresses).
- Finally, set realistic benchmarks for data accuracy. What constitutes realistic benchmarks depends on a number of factors: who your customers are, costs for acquiring and maintaining the data, etc. For example, you should be able to achieve nearly 100% accuracy for your board of directors, as this list is relatively small compared to your entire database. By the same token, you should be able to have higher accuracy with members vs. nonmembers, and customer vs. noncustomers.
One hundred percent data accuracy is impossible. Accept that, and focus on where you can be successful, not perfect, in your data management.
This article originally appeared in the September 2, 2005, issue of Association TRENDS. Copyright 2005 Association TRENDS. Reprinted with permission.