One of my clients maintains a data quality dashboard of key metrics that is shared with their board of directors quarterly. All the metrics shared are related to the association’s overall strategic plan and track progress on goals. One of those data points is their overall data quality score.
The data score shared originated from the findings of a data quality assessment conducted by a vendor partner. It looked at all the data points used as contact information, as well as a handful of the most used demographics identified by the marketing department as critical to their ability to promote products and services to the members and customers.
Calculating Your Data Quality Score
A data quality score measures 3 criteria:
- Completeness – do you have the information collected? Only records with data are counted as complete. Nulls and blank records are not. It also excludes those marked as a known bad address or known bad email.
- Validity – is the data contained in the field usable and appropriate for that data point? For example, a zip code that was accidentally keyed into the state/province field is not counted as valid.
- Timeliness – how recently was the data collected. Does it reflect current preferences or is the information potentially out of date?
After gathering these measurements for each data point, the scores are combined. You can play with the weighting of the formula, but you must be consistent with it to track your progress over time. In many cases, I weigh completeness as 50% of the score since collecting the data is literally half the battle, then 25% each for validity and timeliness.This calculation represents the data quality score for each data point. You can take an overall average of all fields measured to calculate your overall data quality score. This is the value my client reports out to the board. It is a great metric for overall data health.
Metrics to Include in Your Data Quality Dashboard
Say I want to build a data quality dashboard of key metrics for the executive team to keep an eye on overall data integrity and data quality problems while looking one level deeper.
Here are the top data quality metrics I'd include in my dashboard:
- The percentage of complete and validated mailing addresses for members. Keeping an eye on the percentage of quality mailing address on file is a good idea if you still send a physical magazine to members monthly.
- Count of Email Reach – the count of valid email addresses for all “usable” data. It is critical to first define your usable data set of records. It should include all current members, recently lapsed members for the past x years, current customers over the past x years, and prospects added to the database in the past x years. This count should exclude all those who have opt-ed out of your commercial email messages.
- Top Demographics – I would pick the top 5 or 10 key demographics and list the data score for those demographic fields.
- Overall Scores by Audience – I would segment the data reviewed by audience type and report the overall data quality score for each segment. For example, I would segment volunteer leaders as the most important segment, then members, existing customers, and prospects as additional segments.
What other key metrics would you add to your association’s data quality dashboard?