Jul 19, 2018
Is bad data negatively impacting your business? What if you could identify data quality issues before they wreak havoc on your organization? You can with a proactive data monitoring solution. Here are five tips for developing a strategy for round-the-clock monitoring.
- Know your starting point: Benchmarking your data with a full volume data analysis is the first step in setting up data quality monitoring. Unfortunately, there isn’t a single, all-encompassing list of data quality metrics to keep track of—it’s very business and function specific, which is why it’s important to analyze all of your data. This will show you exactly how significant your data challenges are.
- Identify the data quality issues that have the most significant impact on the business: Now that you’ve identified the most critical data quality issues, you can prioritize which need to be addressed in a timely manner and which can be addressed later. Prioritization will vary based on what is most important to your business—you may choose to tackle data issues that have the greatest revenue impact, the heaviest burden on your business stakeholders, or the greatest impact to your customers.
- Build acceptable threshold definitions for each of your metrics: Most monitoring dashboards have different threshold criteria. These definitions should be agreed upon with business stakeholders. You should set up relevant alerts for users based on your accuracy definitions. Some examples of alerts that you might want to have are if data quality falls below a certain threshold, or if it improves and meets an accuracy goal.
- Ensure good visualizations are available to the business and data stewards: It is important to show business users the state of data quality in a meaningful way. Oftentimes this is done using dashboards with red, amber and green thresholds. Making your monitoring dashboards easy to read and consume will keep business users and leadership engaged without overwhelming them with complex data.
- Review your metrics on an on-going basis: It’s important to fully review your metrics any time a significant new data set is added, as well as periodically (many organizations do this annually, but decide on a cadence that works for your business.) Reviewing your metrics will ensure that you’re still measuring and addressing the most relevant data definitions. If the business changes their focus or way that they’re using data, the accuracy requirements will also change, so frequent audits and reviewing of metrics is critical. One important note about metrics: if you change metrics you make monitoring over time inconsistent, so try to add metrics rather than replace them. If you have a strong reason for changing metrics, be sure that you have some way of compensating for the change in metric definition.
Data monitoring can sound complex, but it doesn’t have to be. Not only does it empower business users to proactively monitor data quality without relying on IT, but it can also help mitigate the negative effects of poor data quality on your business.