6 Steps to Data Quality

Working with a clean data set in your Salesforce org is now more important than ever before. Research from Gartner shows that bad data is the number one reason organizations fail to meet their goals. Experian found that American companies believe that a staggering 25% of their data is inaccurate. In this blog post, we will guide you through a step-by-step plan to improve your data. It does not matter if you are working in a small start-up or in a large enterprise — follow these steps and make sure you improve the customer experience, increase the effectiveness of your sales and marketing efforts, and save on costs.

Before starting with the plan, it is paramount to build a data quality team. This team needs backing up by the highest levels of leadership in your organization. Data quality efforts that lack support from upper management are bound to fail. Include team members from all departments that create, use, or manage data. Departments like marketing, sales, and IT first come to mind but there may be a lot more. With this team, you will follow the these six steps.

Step 1: Data purpose

Assess and document which department needs what data for what purpose. For example, marketing needs email addresses to send both automated campaigns as well as newsletters. Also take note of important KPIs that indicate the quality of the data (hard bounces and spam complaints, for example).

Step 2: Data assessment

This is a two-part step. First, compare the data you require (from step one) with the data you have. This will highlight areas to prioritize. You might find out that some data is completely missing, some data is not used at all, and some data is polluted with duplicates and non-validated values.

Secondly, document the data flow with the CRUD model: where is data Created, Read, Updated and Deleted?

Step 3: Data rules

Now that you have established the current state of your data, it is important to describe how you want it to be. You do this by establishing data rules. Data rules contain specific instructions on validation, formatting, required fields, CRUD rights, storage, and deletion. For example, you could introduce a rule that requires all phone numbers to be registered and to be formatted including country and area codes, using 00 as the country code prefix.

Step 4: Data rule implementation

Now comes the hard part: getting your organization to play by the data rules your team has established. In this step, it is important to make sure that all new data that is getting in follows the data rules. You do this by implementing validation solutions on the different points of entry and by using tools to format and deduplicate the new data that is coming in.

As a lot of data is still entered manually, you will need to brief your organization on the new data rules. Support from upper management is key here. 

Step 5: Clean

The previous step was about plugging the holes. This step is about cleaning your existing data. Don’t try to do this manually. There are plenty of good tools out there for formatting and deduplication. Most of them offer free versions or trials, so you can try before you buy.

Step 6: Monitor

It's time to reap the benefits of your hard work. By checking the KPIs you defined in step 1, you can see the impact of your efforts. Keep tracking these KPIs so you know when something goes wrong with your data.  

Rinse & Repeat

Unfortunately, data quality is not something you can solve by running a project. It needs regular attention. Our advice is to repeat these six steps every year or every other year. By making a constant effort toward better data, your organization will achieve what we like to call data happiness. Data happiness means that data will help you reach your goals, not hold you back.

Gijs Hovens is marketing manager at Duplicate Check, helping Salesforce users achieve more revenue and cost reductions by improving data quality. Check out Duplicate Check on AppExchange.