All docs pages (49)
Data Quality
Score one object's records on field fill rates, email, phone and URL format, and how long since they were last modified, with a CSV of every problem value.
Updated 2026-11-09
Data Quality gives one object a score out of 100 and lists the values that look wrong. It looks at three things: how often each field is filled in, whether email, phone and URL values have a valid shape, and which records haven't been modified in a long time. Open Data → Data Quality. The page is titled Data Quality Scanner.
How to use it
- Pick an object in the Object box (Search an object to scan…).
- Check what will be measured. You don't choose these by hand; they come from the object's fields:
- Format checks lists every email, phone and URL field, with the check each one gets.
- Completeness says how many fields are measured for fill rate.
- If the object has a Last Modified Date field, Flag records not modified in is ticked, with 365 days. Change the number of days, or untick it to skip the staleness check.
- Click Scan quality.
What you see
The scorecard shows:
- the score out of 100
- how many records were scanned
- records with issues
- a count per problem type, such as bad email, bad phone or bad url, shown only when there are any
- stale: records not modified within your number of days
The score is the share of scanned records that have no format problem and are not stale. A score of 90 or more is shown as good, 70 to 89 as a warning, and below 70 as bad. Fill rates are shown but do not change the score.
Field completeness lists each measured field with a bar and the percentage of records where it is filled in, lowest first.
If there are format problems, a table lists them with Record (a link that opens the record in Salesforce), Field, Problem and Value. The table shows the first 200. Export CSV downloads all of them to a file named after the object, such as Account-quality-issues.csv, with the columns RecordId, Field, Check, Value and Message.
The format rules:
- Email: one
@, a dot in the domain followed by at least two characters, and no spaces, commas or semicolons. A field holding two addresses fails. - Phone: 7 to 15 digits once spaces, brackets, dashes and other formatting are removed.
- URL: an
http://orhttps://address, or a bare domain such asexample.com/page, with no spaces.
A blank value is never a format problem. Blanks show up in the fill rates instead.
What it reads and writes
Data Quality only reads. It reads the object's field describe, then runs one SOQL query for the record Id, Last Modified Date (when staleness is on), the email, phone and URL fields, and the fields measured for fill rate. The scoring runs in your browser. Nothing is written to your org.
Limits
- A scan reads at most 2,000 records, with no sort order. When it hits that cap, the scanned count says "capped 2000". On a larger object the score describes those records only.
- Fill rates cover the first 50 fields in the order Salesforce's describe returns them. That list includes system fields such as Id, so it may not include the fields you care about most. You can't pick the fields.
- Checkbox fields always count as filled in, because unticked is still a value.
- Address, geolocation, encrypted and binary fields are not measured.
- Format checks run only on fields whose type is email, phone or URL. A text field that holds email addresses is not checked.
- Stale records are counted but not listed, and they are not in the CSV.
- The phone check counts digits only. It doesn't check country codes or area codes.
Related
- Duplicate Finder: find and merge records that were entered twice
- Data Loader: fix the values you exported by pasting them back as an update
- SOQL Builder: query the records behind a problem