Steppingblocks doesn't describe its data with a single accuracy percentage. Instead, every data point passes through a multi-step validation process, and one rule governs what you see: if we are not confident in a data point, we don't display it.
How we validate the data
We use layered quality checks:
Validation at the point of collection
Standardizing and deduplicating records
Cross-checking information across sources to find the most accurate, current value
Machine learning models
Automated quality and recency checks
Anomaly detection, with flagged records sent for human review
Human audits and ongoing testing
Where does a specific data point come from?
No single field relies on one source. Each value is built from hundreds of sources, then normalized, deduplicated, validated and enriched by machine learning. A given field may come from one input or several, depending on the profile, and some values, such as salary, are modeled from multiple sources.
If you need a methodology summary for a report or accreditation review, ask your Partner Success Team Member.