Skip to main content
All CollectionsBest Practices
Best practices - Data QA
Best practices - Data QA

Best ways to verify if the Platform data matches your file

Edwin Ng avatar
Written by Edwin Ng
Updated over 4 years ago

Verifying if MSIGHTS Reporting is meeting your expectations will always be a priority when pulling Reporting in the Platform. However, eyeballing data is not the ideal way to check that the numbers you are seeing in the Platform are what you're expecting.  

We would like to share our recommendations for Data QA:

  1. The first and most important step is to make sure you are comparing the same data. By far the most common mistake when QA'ing data is that the Platform data is being compared to a different file. You have to compare the same exact data in order to have a match.

  2. Start at a high level. Our recommendation is to use a longer time frame at first before you dive into a more granular level. Does the data match at the yearly/monthly level? Does the campaign have the correct amount of impressions? If the answer to these questions is yes, we can start going into a more granular level.

  3. Avoid having extremely large comparisons. Get a good sample size that is easy to work with at first. If the data at the monthly level looks correct, there is no use in doing a complete year by day QA next!

  4. The export function is your friend. Do not hesitate to export your Widget and use it to compare to your Microsoft Excel file using pivot tables. Some users are more comfortable using Excel to QA their data, so exporting directly to Excel can facilitate the QA Process.

  5. Keep in mind that a small discrepancy can be found. Not all sources are exact, so some discrepancy is normal. We aim for 100% match rate, but anything higher than 96% is deemed acceptable by industry standard. Most of the time, the issue is due to some sources using more decimal places.

  6. Ask for help! The MSIGHTS Client Engagement team will gladly assist you if you need help, as well as guide you through our method for internal QA. At the end of the day, we are all a team with the same goal!

Our recommendations are based on the MSIGHTS day-to-day process and methods. This, however, is not the only way to do data QA.  

Please do not hesitate to reach out to MSIGHTS Support via the messenger chat window at the bottom right-hand corner of your screen or with the button below for any questions, support, or feedback. We'd be happy to help!

Did this answer your question?