Data and Analytics

Stanford University: Is Your Data Good Quality?

Stanford University: Is Your Data Good Quality?

February 16, 2021 | Tricia Gabberty

Defining Quality Data 

As marketers, we rely on data. But whether we use it to make key business decisions or simply cite the results of a third-party survey, can we really trust the data we’re using? 


According to online sources, “data quality refers to the state of quantitative and qualitative information. Data is generally considered high quality if it is fit for its intended uses in operations, decision-making and planning.”

In the latest episode of our Data Dialogues podcast, I interviewed data expert, Alice Siu. She is Associate Director at Stanford University’s Center for Deliberative Democracy, which helps government policy makers consult the public on policy issues, while ensuring the public is providing informed opinions.  When I asked Siu how she defines quality data, she emphasized that data must be used for its intended purpose. 

“With data quality, at least in public opinion polling, we have to make sure that we’re not only gathering the population accurately, but we’re using it to generalize to the right population,” Siu explained. “So, if we’re intending to represent the entire country, then we need to get a survey that actually represents the entire country. And that means good, random representative sampling of the population.” Listen to our podcast for the full interview.


Buyer Beware

There are several factors to consider when using data. Some of them are accuracy, completeness, reliability, relevance and timeliness. Siu said we can’t assume data has these attributes when we acquire it; therefore, it’s critical that we verify it. “If the seller is not able to give you a straight answer one way or another, then maybe you shouldn't be buying from them. It’s also important when thinking about first-party data and understanding where the streams of data come from,” said Siu.

“You need to ask the right questions. How often is this data refreshed? How often is a certain percentage of the data verified? I think there are basic, common sense questions that people should be asking before they purchase data. And they shouldn’t assume that the seller will do all the work for you.”

The Consumers’ Responsibility

Of course, all marketers are consumers too. So how much responsibility falls on us to make sure the data we consume in news articles or from a brand survey is reliable? Siu maintains that consumers have a shared responsibility before accepting the information as fact. “If a survey is talking about those who are 18-24 or Gen Z or those over 65, was that survey actually done for that group of people? Sometimes survey outlets may do a survey of a larger population and then subset it and say, ‘Hey, this population from there, maybe 20 people said this.’ And all of these details are never asked explicitly when reporting the survey data. And it's really unfortunate,” said Siu. 

“A very reputable survey organization will report everything, but nowadays anyone can go out and do a survey. So, I think as consumers of market research and data and polls... we should do a little bit more research before we actually process and store it away in our head.”  Learn more about this podcast series and listen to other episodes of Data Dialogues now.