Good gender pay data is defined as that which reports the difference in earnings between equivalent job roles across an entire organization, ideally split by quartiles of pay, so comparing gender differentials in pay across similar roles within the top 25%, then the next 25% up to 50%, etc. Good metrics also report the number of women in senior positions.
The UK is an example of where this is done. Frequent updates are essential however and, although the UK has broken ground in doing this, low publishing rates caused by the relaxation of rules on publishing data means meaningful analysis over time is difficult at present.
With this data, comparisons can be made between organizations in the form of ranking lists. This particularly true of government departments and strategic suppliers, which are often more limited in number
One of the challenges of this approach lies in different organizations within a holding company’s umbrella reporting different results. In this case, using the top company’s figure or averaging the amount across organizations reported.
Categories such as UNSPSC or CPV that are published or assigned through machine learning can be used for sector analysis. This means aggregating the contract value calculated by start date and end date for each sector, such as Information Technology, and then comparing this to gender pay metrics.
With good, consistent publishing, analysis can be carried out over time to identify trends, specifically the delta of gender metric between one year and another.