Pave's level predictions for our team were off by a few levels. How do you verify that the companies in the dataset level their employees correctly?

We use a mixture of machine learning and human validation to review and massage Benchmarking data before it enters the production environment. When companies adjust and confirm employee levels in the Level Your Employees tool, the user input flows into Pave's backend and is used to help train future predictions.

If our compensation experts disagree with a predicted value, the data is adjusted and leveled accordingly before it is released into the Benchmarking product.