Salesforce orgs pick up data quality problems over time, as the org grows more complex and more people type into it. The platform is fine; the data is what rotted, and it drags reporting, forecasting and any trust in the system of record down with it. You see it as duplicates, half-filled fields and information that is simply wrong, which turns into missed opportunities and worn-out customer relationships.
Defining high-quality Salesforce data
Work out what "good" data means for your organization before you start fixing any of it. Aim for records that are complete, with the key fields consistently populated; accurate, reflecting real and up-to-date information; consistent, using standardized formats and values; and trustworthy enough to report and decide on.
The 4-stage data quality framework
A structured way to get Salesforce data quality up and keep it there:
1. Audit your data
Start with what you actually have. Built-in Salesforce reporting will surface the empty fields and the placeholder records. For larger datasets or deeper analysis, reach for:
- Salesforce reports, which show records in a spreadsheet-like format.
- Data Export or Data Loader, to pull data out for external analysis.
- OrgCheck, for visibility into the data model, role hierarchy and metadata quality.
- AppExchange tools, when the analysis gets complicated.
2. Deduplicate existing records
Duplicates are the usual pain point, and Salesforce ships tools for them. Matching rules define what counts as a potential duplicate, by exact name match or fuzzy name match. Duplicate rules decide what happens when one turns up: alert the user, or block the record from being created.
Those rules handle the common scenarios. Complex edge cases need real merging, and for large-scale deduplication and merging you are into third-party applications.
3. Define what to control and how
Work out which fields are critical and write down the data entry standards.
- Mandatory or optional. Be careful about which fields require data. Make too many of them mandatory and users get frustrated and start typing placeholders.
- Data entry stages. Some details only need to be there later, so "Projected Revenue" might be optional at prospecting and mandatory at the quote stage.
- Data formats. Say which one you want: open text, picklist, numerical.
4. Enforce policies with Salesforce functionality
Native features to hold those standards up:
- Schema Builder, to see the field structure and requirements you already have.
- Picklists and dependent picklists, which standardize entry with predefined options and filter the choices as the user goes.
- Dynamic Forms, to display or require fields based on user actions or earlier input.
- Validation rules, for granular accuracy and completeness logic: require a justification for a discount, prompt for follow-up information when specific criteria are met.
- Salesforce Flow, for the more involved validation and routing, such as deal size checks or approvals that have to reach a specific team.
Leave a Comment