6 Salesforce data-quality fixes that make reporting more reliable

A report can be technically correct and still lead people in the wrong direction. If key fields are empty, values mean different things to different teams or the same customer appears three times, the chart simply presents those problems more neatly.

The useful response is not always a large data programme. A team can often improve confidence by choosing a few important fields, fixing how information is captured and monitoring the exceptions. Salesforce identifies missing, duplicate, inconsistent, incomplete and stale records as common data-quality problems. The platform behaviours below come from current Help and Trailhead guidance; the sequence and operating recommendations are the Ostrelis view.

1. Measure the gaps that affect a real decision

For admins and managers: start with the report or business decision that is causing concern. Identify the small number of fields it depends on, then measure how often those fields are blank or unusable. A sales forecast may depend on close date, stage, amount and next step. A service view may depend on status, priority, owner and the date of the last customer response.

Salesforce Trailhead recommends defining data-quality measures that can be monitored, such as reviewing leads without industry information each month. This is a better starting point than trying to clean every field on every object at once.

Quick action: build one completeness report for five decision-critical fields and record today’s baseline.

2. Standardise values before adding more reports

For admins and users: look for fields where the same idea appears in several forms. Free-text entries such as “United Kingdom”, “UK” and “U.K.” create separate report groupings even when users mean the same place. Similar problems appear in reason codes, customer types and referral sources.

Agree the definition first. Then use an appropriate controlled input, such as a picklist, where the business process genuinely has a stable set of choices. Avoid forcing nuanced information into a long, confusing list merely to make a chart tidy. Existing integrations and automation also need reviewing before values are renamed or retired.

Quick action: run a report grouped by one important categorical field, identify variations and agree the preferred values with the team that owns the process.

3. Tune duplicate controls around the way records arrive

For admins: Salesforce uses matching rules to identify possible matches and duplicate rules to decide how those matches are handled. Depending on configuration, users can be alerted or blocked, and duplicate record sets and reports can support wider cleanup.

Do not assume an active rule catches every route into the org. Salesforce documents important considerations for edits, imports, APIs, simultaneous saves and user access to matching fields. Review where records come from—manual entry, web forms, integrations and imports—then test representative examples as the users and processes that create them.

For users: when Salesforce shows a potential duplicate, check the existing record rather than changing the spelling to bypass the warning. Give users a clear escalation route when they cannot decide whether records should be merged.

Quick action: test one common duplicate scenario through each major record-entry route and confirm who owns the resulting review queue.

4. Ask for information when it becomes necessary

For admins: Salesforce validation rules check entered data against conditions before a record can be saved and display an error when the condition is met. Required fields, page layouts, validation rules and Flow can all support data standards, but they do not serve exactly the same purpose.

Make a field mandatory at the stage when the business genuinely needs it. Requiring too much at initial creation encourages placeholders and frustrates users who do not yet know the answer. A useful validation message should explain what needs changing in plain language and, where possible, point to the relevant field.

Test new controls against integrations, mobile use and existing records as well as the standard desktop form. This testing sequence is an Ostrelis recommendation; the exact impact depends on the organisation’s configuration.

Quick action: review one frequently triggered validation rule with users and rewrite any message that does not explain the correction.

5. Make stale records visible and assign ownership

For users and managers: a complete record is not necessarily a current record. Old close dates, departed contacts and queues owned by inactive users can quietly undermine otherwise polished reporting.

Define “stale” for the process rather than applying one arbitrary age everywhere. An untouched open opportunity might need review after 30 days, while a long-term account can remain valid for much longer. Use last activity, last modified date and relevant business dates carefully: an automated update can change a system timestamp without proving that a human checked the record.

Quick action: create a focused view of active records whose meaningful business date is overdue, then ask owners to confirm, update or close them.

6. Turn exceptions into a small monthly routine

For admins and managers: cleanup lasts only when somebody can see whether quality is improving. Keep a small set of exception reports for missing critical values, non-standard entries, potential duplicates, stale active records and unowned work. Give each measure a named owner and a realistic target.

Track the trend, not just the total. A falling backlog may hide a new source that continues to create poor records; a rising total may simply reflect business growth. Compare the exception rate with the relevant population and investigate sudden changes after releases, imports or process updates.

Quick action: schedule a 20-minute monthly review with one owner for each measure and agree one corrective action, not a long wish list.

Start with the data behind one important report

Reliable reporting grows from clear definitions, sensible controls and regular ownership. Choose one report people use to make a decision, identify the data risks beneath it and apply the smallest fix that prevents the problem recurring. That creates visible progress without turning data quality into an endless cleanup exercise.

Ostrelis can help teams improve Salesforce data, administration and reporting while keeping changes practical and supportable. Explore our Salesforce data and integration services, day-to-day Salesforce administration and reporting and dashboard support.