Want to see demo?    
Guide

Data Profiling in Salesforce

Essential Definitions & Best Practices

Salesforce Data Quality Guide

Depending on its age, your Salesforce org may have dozens or even hundreds of objects. To start data profiling, aim to build a strong foundation for a few key objects before you scale your efforts. Focus on a subset of 3-5 objects and prioritize improving their data quality. Implement continuous monitoring to maintain the data quality of these initial objects. Over time, incorporate additional objects into your data profiling strategy.

Which Salesforce Objects Should You Profile?

Start by profiling your most important Salesforce objects. These are the objects your organization relies on to do business. Because users interact with these key objects daily, they contain the most data, presenting the greatest opportunity for optimization. In addition to the suggestions below, look at your org’s Storage Usage to identify other large objects.

Key Sales Cloud Objects to Profile

Start by profiling Opportunity, Account, Contact, and Lead. If your business uses them, also profile Campaign, Contract, Product, Price Book, and Quote.

Key Service Cloud Objects to Profile

Profile Case and Contact. Next, consider profiling Service Appointment, Service Contract, and Asset.

Run a Full Scan Profile for Each Object

Your profiling definitions should start broad and then narrow to more and more specific scenarios. During your analysis, you will create multiple profiling definitions for the same object. Be explicit with your naming and category classification.

Start with a full scan of your selected objects. Create a profiling definition that includes all fields and records for the object. This will give you a quick understanding of record volumes and age, overall field fill rates, record types, and data management rules.

In Cuneiform for CRM name the profiling definition {{Object Name}} – All Fields – All Records and give it a Category of “Full Scan”. Note: Remember to add new fields to your profiling definitions as you create them.

Example data profiling definition
Example data profiling definition

Take note of which objects use record types and any other segmentation attributes.

Create Targeted Profiles for Each Object

The full scan provides a good baseline of the object and will help you target data quality improvements. However, these broad profiles can hide critical nuances in your data.

For example, an org may have several thousand customers and 50 partner Account records. Partner Accounts may represent only two percent of the data, but 40% of revenue. Assessing partner-specific fields alongside the entire Account dataset may lead to inaccurate results.

To ensure thorough assessment and avoid overlooking key insights, develop targeted profiling definitions to evaluate data in context. Avoid making decisions solely based on low record volumes. Instead, analyze data patterns across various scenarios to gain a comprehensive understanding.

Segmentation-based Profiling Definitions

Some fields may be used widely across the business, while others are only used within specific record types. This can cause fields to appear underutilized when they are highly populated for the relevant record type(s). Create a profiling definition for each record type and any other important segmentation attributes.

  • Record Types
  • Business Units
  • Demographic segments

Time-based Profiling Definitions

Time plays an important factor in data analysis too. As businesses grow and evolve, so do their processes and data collection in Salesforce. Data captured ten years ago may no longer be relevant today, and vice versa. Prevent older records from skewing results by looking at recent records first.

Transactional objects, like Opportunity and Case, accumulate large volumes of records rapidly. Streamline your data analysis by profiling records from the current year plus a specified number of previous years. Determine the appropriate timeframe based on what is relevant to your business. If uncertain, begin with three years and assess the results to determine how to refine the profile definition.

For master data objects in Salesforce, such as Account and Contact, consider segmenting on recency and activity. For example, a seven-year-old Contact with two Cases in the last year is valuable. Conversely, a two-year-old Contact without any business interaction is likely not.

Profiling recent records will inform your go-forward data quality strategy. Profile historical records to formulate your data archival strategy. These results will provide insight into the relevance, importance, and impact of historical data in current and future operations. Use profiling insights to determine which records to retain, archive, or delete.

Armed with rigorous data profiling results, you are ready to identify specific actions for cleanup and enhancement. Implement targeted strategies to ensure the integrity, reliability, and relevance of your Salesforce data, driving greater business success and efficiency. Read on to clean up unused Salesforce fields.


Want to see our platform in action?


By leveraging the Cuneiform Platform, you can obtain and use more accurate, data-driven insights through effective data quality monitoring. Learn more about how we can help you with your important tasks.