The importance of data quantity, quality, coverage
Today’s businesses and organizations are awash in a sea of data. According to a recent study released by Gartner, more than 75% of them are using data to achieve important business goals. Among them, better marketing and reduced costs.
Can your marketing efforts benefit from modeling data?
Let’s take a look at some examples. Say you’re a large member organization looking to market co-branded credit cards. To be more effective and help reduce costs, you need to match prospects against lifestyle, econometric, purchase history and aggregated, non-regulated credit information. Your goal is to pinpoint individuals who can afford and want it, and are likely to qualify for it.
Or, let’s say a vacation property seeks to sell timeshares. They may want to target individuals who possess certain financial, demographic, lifestyle and real estate-related data points that are predictive of purchase. And, what about non-profits seeking to expand donor bases? They might look to financial, real estate, and demographic data to pinpoint prospects that have the necessary funds and are most likely to donate.
In each example, the most predictive data attributes are discovered empirically through testing, collecting responses, and analyzing results.
Better data produces superior results
For modeling purposes, quality, cleanliness and reliability of data is paramount. Source and scope are also integral to successful predictive outcomes. Necessarily, you need good data to build decent models.
If you are purchasing data externally, be sure potential vendors answer these questions:
- What kinds of data does each vendor have?
- In general, where did they get it?
- How do they cleanse data before adding it to their file?
- What do they do to ensure data is as accurate as possible when put to file?
- How do they update single attributes when multiple pieces of information are not the same? How do they update key predictive attributes?
- How do they compare to key competitors in regard to coverage, accuracy and performance?
- What percent can actually be used for modeling?
- What percent can be licensed?
- How often is their data updated?
Finally, be sure to ask what the database match rate is. While you can obtain reams of data, if you’re not getting high match rates what is it doing for you? Companies like AnalyticsIQ apply sophisticated, multi-faceted matching algorithms to search databases to optimize match rates. By using the best, most accurate data (such as household and Zip+4 level), we create more robust models and produce match rates in the range of 95-plus%.
While different data is predictive of different things, accurate data produces superior modeling results. Use it to develop a broader and more precise understanding of your prospect and customer values, interests, behaviors, priorities, and characteristics.
That’s the proven way to increase response rates.