A to Z on Cleaning Your Customer Data, and Keeping It Clean

Data is everywhere! And new technologies frequently become available to help agencies capitalize on that information. There are solutions that identify cross-sell and upsell opportunities to those that can conduct sentiment analysis. Most are designed to boost customer retention to helping agents provide more personalized customer service.

But while agents have access to significant amounts of data, the information will only yield benefits if the sources for that data are accurate and complete.

To help keep your data clean, up to date, and protected, follow these seven practices:

1. Create a data role within your agency: Assign one person who can implement, incentivize, and enforce good data maintenance procedures. The individual can create guidelines on how data should be input into your agency management system. They can train others to follow the guidelines. Your data czar can take a big picture view of where the data is coming from, how it is being stored, and possible sources that could potentially corrupt the information. They can regularly review reports coming out of the systems and identify issues that need to be fixed.

   2. Put a priority on clean data: It takes teamwork to keep data clean. For example, if the data czar identifies something to be corrected, it’s important to inform the person who owns that customer record to correct the issue. This will help the team understand problem areas and make them more alert to bad data.

   3. Correct any errors at the source: Some solutions will pull information from an agency’s system and use this data to populate a document. Often the agent has a chance to review the document and make corrections before forwarding it. Stress to your staff the importance of making fixes in the agency management system. This will save team members from having to correct the same information over and over again and eliminate the chance that something might be missed.

   4. Know your vendor contracts: Ultimately, the data belongs to the insured. And agencies need to ensure that the customers’ information is being used responsibly. Some vendors have contracts that give them control of the data enabling them to monetize it for their own use. They can also make it harder and expensive for agencies to get their data back. Thoroughly review your vendor contracts to know how they plan to use the information. If you are having trouble deciphering the legalese, consider putting it in ChatGPT and asking it to explain the terms to you in simple language. 

   5. Develop data procedures to boost consistency: Making sure everyone enters data the same way is critical. If different employees input information in different ways, the solution might not be pulling all the necessary information. Things to consider include:

   - are the names of carriers and products uniform?

   - is the client and prospect information entered the same way?

   -are agents using the same terms to describe policies and client requests?

   Create a set of guidelines to ensure people enter information the same way every time. Stress to your employees that everyone should follow the guidelines when entering information.

   6. Be clear on what information is required: Systems can have many different fields to input information into. Indicate which sections must be completed, which sections you would prefer be filled in and which areas can be skipped. This will help reduce the number of incomplete records.

   7. Don’t skip over regular data cleaning practices: Agencies should audit their data at least once a quarter. Review your book of business, and ask these questions:

   - Are there any duplicate records that need to be combined?

   -Are any records missing information?

   -Are there any old records that you no longer need?

   If you are hesitant to completely delete an old record for fear you might end up needing it, export that record to an Excel sheet. If after 6 – 12 months you have not touched it, you no longer need it, and it can be permanently deleted.