Load safely, keep it clean
Choose the right tool, plan migrations in order with testing, prevent duplicates with matching and duplicate rules, and protect data with backups and validation.
- ToolsImport Wizard, Data Loader and upsert.
- QualityValidation, matching and duplicate rules.
- SafetyBackups, storage and migration planning.
Bad data ruins adoption. Admin interviews ask how you load, update, clean and protect data — and how you choose between tools. Accuracy here shows you can be trusted with production.
1. Import Wizard or Data Loader — how do you choose?
Choosing a data tool
How many records, and which objects?
- Under 50,000, common objects→ Data Import Wizard (browser, duplicate checks, simple)
- Up to millions, any object→ Data Loader (insert, update, upsert, delete, export, scheduled via CLI)
- Recurring integration→ APIs, ETL tools or MuleSoft
2. What is upsert and why use an external ID?
Upsert updates a record if a match is found and inserts it otherwise. Matching on an external ID (a unique id from the source system) lets you load the same file repeatedly without duplicates and link parent records without knowing Salesforce Ids.
3. How do you plan a data migration?
Migration steps
- MapSource fields to Salesforce fields and picklists
- CleanDe-duplicate, standardise formats, fill required fields
- OrderParents before children (Accounts → Contacts → Opportunities)
- TestLoad a sample into a sandbox, validate with users
- LoadTurn off non-essential automation, load, then re-enable
- VerifyReconcile counts and spot-check records
4. How do you prevent duplicate records?
Use Matching Rules (how records are compared — exact or fuzzy on name, email, phone) with Duplicate Rules (what happens — block or allow with an alert, and report on duplicates). Clean existing duplicates with merge and duplicate jobs.
5. Write a validation rule that requires a reason when an Opportunity is Closed Lost.
Validation rules block saves when the formula evaluates to true, showing your message. They apply to the UI, API and data loads.
AND(
ISPICKVAL(StageName, "Closed Lost"),
ISBLANK(TEXT(Loss_Reason__c))
)
/* Error message: Please select a loss reason before closing the opportunity as lost. */6. How do you back up and restore Salesforce data?
Use Salesforce Backup (or a third-party backup tool) for regular automated backups and point-in-time restores; Data Export Service for weekly or monthly CSV exports; and keep metadata in source control. The Recycle Bin keeps deleted records for 15 days.
7. What are data and file storage, and how do you manage them?
Data storage holds records; file storage holds attachments and files. Monitor usage in Setup → Storage Usage; archive or delete old records (for example, closed cases older than five years), reduce attachments and use Big Objects or external storage for historical data.
8. How do you keep data quality high over time?
Required fields and validation rules for key data, picklists instead of free text, duplicate rules, data quality reports and dashboards, ownership of data by teams, regular clean-up jobs and training users on why the data matters.
Remember
Key points
- Parents before children
When loading related data.
- Upsert on external ids
No duplicates on re-runs.
- Validation fires on loads
Plan bypasses carefully.
- Test in a sandbox first
Then load production.
Interview drill
Follow-up questions
What the interviewer asks after your first answer. Turn on practice mode, answer each one aloud, then reveal.
-
1 How do you mass delete records? Basic
Mass Delete Records in Setup for some standard objects, or Data Loader delete/hard delete using exported Ids.
-
2 What is the Bulk API? Intermediate
An API that processes large data sets asynchronously in batches; Data Loader can use it for faster loads.
-
3 How do you load records with a lookup to another object? Intermediate
Provide the parent's Salesforce Id, or map the lookup to the parent's external ID field during upsert.
-
4 What is hard delete? Intermediate
Deleting records without sending them to the Recycle Bin (Data Loader with Bulk API and the right permission) — irreversible.
-
5 Can validation rules be bypassed for integrations? Intermediate
Yes — add a check for a custom permission or hierarchy custom setting in the rule formula so the integration user skips it.
-
6 How do you schedule a Data Loader job? Advanced
Use the Data Loader command-line interface with process configuration files, scheduled by the operating system.