What is the best way to measure INDEX impact on production?

We are in process of creating nonclustered indexes on tables on QA environment on SQL Server 2016. Tables are the same in terms of structure (columns, clustered indexes) but tables on production have more rows and more partitions. Business would like to have some stats/data thanks to which will be able to decide If we can create index on production environment. So we prepare: - Logical reads before and after index for select queries and insert batches - CPU time (and elapsed time to check if we need parallelism) before and after index for select queries and insert batches - Duration of select queries and insert batches before and after index - Storage size of new indexes Do you have any advice ? Is there any list or process for that ?