Production information about the equipment that is the source of your IoT data is also valuable to throw into the mix. Failures are often linked to the manufacturing date, plant, and line where the equipment was assembled. Subcomponents and third-party parts are tied to the equipment identifier.

Production data can also be used to estimate time in the field if that is not directly available. This can be done by using manufacturing date as a starting place and estimating time to sale and in field operation. Different equipment configurations may also end up being key attributes that you can use to make assumptions for other data that you may not have.

For example, if a company builds vehicles, and red flashing lights are only included on vehicles built as fire trucks, you can probably classify the customer of these vehicles as fire departments even if you do not have this information directly in the customer data.