Topic 3: How to Measure Data Quality
Blog | January 30, 2023
This is part 3 of the “What is Data Quality?” series. Click here to read part 2: https://www.opinionroute.com/blog/topic-2-survey-fraud-is-a-factor-to-manage/
Too often, in market Research sample, purchase decisions are made based on qualitative factors- convenience, vendor brand identity, historical experiences, etc. But if you want to be serious about consistent data quality, then, to quote John Doerr’s book, you MUST “measure what matters.”
As a beginner course in these metrics, here are a few that matter to us:
- Fraud Rate
- Duplication Rate
- Data Acceptance Rate
Measuring these things on a single typical survey is extremely helpful. But the key to a comprehensive data quality strategy is to measure all of these on a systemic level.
For instance, you should track and report on these three metrics across all clients, projects and vendors. Then, analyze all of it on a daily or weekly basis for trends. This approach enables you to spot events that compromise vendors and proactively protect your studies.
For example, in Q4, we noticed a major sample vendor had seen a 110% spike in daily fraud blocks ratios. We immediately reviewed every project using that supply and paused their traffic where we saw the spike in our over trended data set. We provided the vendor with the technical data we had to help them investigate the main channel introducing the fraud. Once we had assurances they had cut that channel off, we slowly reintroduced their traffic in a project or two while we measured if the fraud rate dropped to normal levels.
This sort of approach not only protects your data, but it holds vendors more accountable for quality.
Next, we will highlight how Professional Expertise matters.
Please click here for Part 4 of 5.