Testing Data? Ask These Questions to Evaluate Data Quality

Marketers no longer need to be convinced about the benefits of taking a data-driven approach to their work. A recent Salesforce survey found global marketers are relying on an average of 15 different data sources, up from 10 the previous year. In fact, 82% employ “declared interest/preferences data” and 77% use “inferred interest/preferences data.”

As a marketer, you naturally only want to use reliable data to enrich the data you already own.

After all, using poor-quality data can lead to big problems. You can fail to get an accurate 360-degree view of your customers, for example, leading you to target the wrong people with your ads or the right people but with the wrong products. Promising campaigns can look like failures, and ultimately, the money you invest in data is wasted.

Though data quality may seem like an abstract concept, it’s actually not. You can consider a number of measurable characteristics to determine whether or not the deliverables of a data provider can help you reach your goals.

What Are the Characteristics of Data Quality Measurement?

The data integrity company Precisely defines five traits that are integral to data quality:

  • Data accuracy: Is the information correct in every detail?
  • Data completeness: How comprehensive is the information?
  • Data reliability: Does the information contradict other trusted resources?
  • Data relevance: Do you really need this information?
  • Data timeliness: How up-to-date is the information? 

You can assess these characteristics in some very practical ways when you’re testing data.

  • Accuracy: Did the data provider give correct enrichment fields, like are the categories assigned correctly? If the data includes names and addresses, are they spelled correctly?  
  • Completeness: Are there a number of empty values in fields that are relevant to your use cases? It would be a red flag if you see a lot of blanks in columns for postal codes for certain geolocation, for instance.
  • Reliability: Can you cross-validate the data with your existing data assets, like transaction data? For example, if a data provider claims a group of users is sports fans, transaction data should bear that out. After all, sports fans generally purchase more sporting goods than consumers who aren’t interested in sports. 
  • Relevance: Is the data properly created? If you request only data that pertains to customers in the U.S., you shouldn’t receive data about customers in Europe.
  • Timeliness: Does the data have a lag? For some brands, data that is more than 96 hours old has already gone stale and become useless. Ensure the timeliness of the data suits your purposes. 

Providers have ways to confirm they’re producing high-quality data. They can use natural language processing (NLP), for instance, to measure if their categories are assigned correctly. They can set up data quality checkpoints for data deliverables with their partners. They can also receive anomaly detection alerts for irregular data that comes through their sources.

Asking providers about these and other measures they take to deliver quality data can help you choose the right partners.

An Additional Gauge of Data Quality

Achievements and credentials around data quality can help you feel comfortable about working with a data provider.  ShareThis is proud of the recognition they’ve received in this area

ShareThis is one of only six data providers to have received the IAB Tech Lab’s Data Transparency Label. The label offers a clear view of our audience segments, including where our original data was sourced, how we constructed the audience, and the topline description of each audience.

Our company has also been awarded the Neutronian Certification Badge following a comprehensive audit of the processes and procedures we use to build our audience segments. The certification covers five main categories of data quality with 10 to 15 attributes evaluated in each category. These categories are:

  • Consent and compliance
  • Sourcing transparency
  • Dataset characteristics
  • Methodology and processing
  • Performance

In addition, Neutronian releases a quarterly Data Privacy Scores report where ShareThis consistently ranks as a top data provider. 

Conclusion

Here’s the bottom line: Since quality data can lead to good decision-making, it’s always a good decision to ensure a third-party provider’s data is of the highest quality.

In addition to having an eagle eye on measurable characteristics while testing data, you can ask providers about their quality control and check for achievements and credentials they’ve received for producing quality data.

About ShareThis

ShareThis has unlocked the power of global digital behavior by synthesizing social share, interest, and intent data since 2007. Powered by consumer behavior on over three million global domains, ShareThis observes real-time actions from real people on real digital destinations.

Subscribe to our Newsletter

Get the latest news, tips, and updates

Subscribe

Related Content