Skip to main content

Can you count on social media for data quality?

Richard Jones Archive

For companies seeking to gather large volumes of information about consumers for marketing purposes, there are a lot of places to look. They can search through old sales records, forms submitted either online or in person, or they can venture into social media.

Whenever consumers engage in social networking, such as Facebook, Twitter, Foursquare or another platform,they put a great deal of their personal information online. Their names, phone numbers and addresses are all usually available, as well as websites and email addresses for some parties. But the question persists - how reliable is this information?

According to Search Engine Land, data quality is actually very high with some information collected online, especially from Google and Yelp. On Google, you'll find names to be 88 percent accurate, addresses 85 percent and phone numbers 80 percent. Similarly, Yelp posts accuracy rates of 88, 85 and 81.

Greg Sterling, a contributing editor at Search Engine Land, advised that despite these impressive numbers, companies should still remain wary of relying too heavily on social media for data quality.

"Local business listings data sets are typically riddled with errors," Sterling explained. "They're also dynamic, with enhancements, changes and business closures happening all the time. Maintaining an accurate local business listings database of millions of records is one of the most challenging areas of internet content. Few people (even marketers) realize how difficult it truly is."

It's worth noting that with other social networking sites, accuracy begins to slip. On Facebook, people's listed phone numbers are only 59 percent reliable, and on Foursquare, that figure is around 50-50. You're almost better off just looking in the phone book.

Because of the shaky accuracy of social media information, it might be worthwhile for companies to invest in address management tools that will help them ensure the quality of their databases.