Your team needs data to make critical decisions and run your business. When you’re in that critical meeting where someone is making a case for a bold, new initiative, nothing is worse than when the credibility of the data starts to get questioned. The meeting falls flat. The decision gets tabled. And like many other companies with data integrity issues, your company goes into data-paralysis.
Improve your Enterprise Data Integrity
Here are a few tips that will allow your team to trust your data and make more efficient decisions to boost your organizational performance.
1. Single Source of Data
If you’re like most businesses, you have data all over the company. An easy question such as “How many new clients did we get last year?” can yield several different answers. Should you look at your CRM? Your Project Management Software? You Accounting System? Your Support Management software?
What if you asked a more sophisticated question like: “Which new Strategic Clients that are buying less than $50k worth of our products/services have the highest support cost?”
Many companies use excel spreadsheets from various different systems to try and cobble together an answer. Often this results in multiple answers as different people try merging the data in different ways. Sometimes, the same person might not be able to get the same answer twice. To address this issue, enterprise companies invest in a single-source Data Warehouse. This data warehouse pulls data from all the critical systems in your business and normalizes them so that you can quickly and efficiently make queries across multiple systems. A consolidated Data Warehouse is key in improving your data integrity issues.
2. Single Client with Written and Agreed Upon Definitions and Calculations
To improve Data Integrity, precise data definitions are key. A seemingly simple question such as “How much revenue did our company make in April?” depends on a host of other questions. Fiscal Month or Gregorian Month? Closed Revenue or Recognized Revenue? What you don’t want is a flurry of Excel spreadsheets being passed around through email with incremental adjustments being made by various people. Everyone needs to be able to look at a set of metrics in a single client system and be able to know the definition and calculation of that metric. There should be a single person or team that is a clearinghouse for critical metrics so that changes to the metric can be vetted and communicated.
Your organization should adopt an executive-sponsored, standard look and feel for your reports so that everyone can look at a report and quickly understand the implications. Use standard color and icon indicators. Some data consumers may want to tweak the reports because they like to look at their business in a certain way. You can provide them self-service reporting tools such as Microsoft SQL SSRS, Tableau, or Excel PowerBI. However, reports that are used across businesses or reviewed by execs should be standardized so that the review can be about the data and not about how to read the report.
3. Data Ownership
Each metric should have an owner. This owner should be the approver of calculations and definitions. This owner should also be responsible for the target values if there are any. You don’t want to be in a meeting when someone asks, “who changed the definition of this metric?” and everyone is looking around shrugging. Your reporting system should have the ability to have metric owners and they should be the only one with the authority to approve changes to the metric. The metric owners should also understand the metric and be able to validate if there are ever issues with the numbers. The alternative to this is a logging mechanism to track any changes that were made to each metric. Data integrity requires data accountability.
4. Automate where Possible
Manual metrics are almost always complained about and rarely trusted. Manual metrics are one of the key reasons for poor data quality. These metrics depend on people entering in data on a daily, weekly, or monthly basis. Because it’s people entering in data, sometimes there are absences, mistakes, or just plain neglect. If possible, you should always look at automation options. Most of your data are stored in systems around your company and with some code most of that data can be pulled into your reporting system. Sometimes, it’s better to have no data than wrong data, as you don’t want to make critical decisions based on a completely wrong data point. You will find that people stop looking at data that is unreliable over time. If you want to improve data integrity, automate as much as you can.
5. Update Frequency
A question that is often heard in meetings where data is being presented is, “When was this last updated?” Your reporting system should make clear when each metric and/or report was last updated. Ideally, you want your data as real-time as possible. Increased update frequency increases data quality and data integrity. Sometimes, this is very difficult or not possible. The next best thing is to let the data consumers know when the last update was done so that everyone in the room knows the time range of the data. A primary tenet of data integrity is that the users’ understand the data.
Affirma has been helping companies for more than a decade in making sense of their data and grooming it so that it is usable by the organization. We can help you develop or improve your data reporting environment so that you can make the critical decisions to help your business thrive. Contact us for a free consultation and let us know what data integrity challenges you are having. Our experts are happy to suggest some recommendations to help you get the information you need or help you develop and implement a data reporting environment.
Comments are closed.