Integration data quality: validation & normalization

Most integration outages start with data: wrong currency, malformed phone, invalid SKU. Use validation and normalization to prevent garbage-in/garbage-out.

Data quality checklist

• Shared dictionaries: currencies, statuses, warehouses.
• Normalize: phone in E.164, lowercase email, canonical domains.
• Validate before sending; log rejected events.
• Version your payload schema.

Recommended approach

• Fail fast at the edge: validate incoming payloads and user input before integrations run.
• Sanitize and keep raw + normalized values for auditability.
• Schema evolution: introduce new fields via versioning, keep backward compatibility for a period.