You may have seen in Summary\Cloud Readiness the "Headroom" fields which appear in the Cloud Sizing views:
These fields are captured by inventory and then refined by ARC to build an accurate picture of a machine's requirements. Movere then uses a mixture of CPU, RAM, IOPS, and throughput for a given machine to decide the Cloud profile required in Azure or AWS.
These calculations are really helpful for manually fine-tuning the actual profile to purchase. The headroom fields show the extent to which this metric is driving Movere's sizing recommendation. A higher the value e.g. 75% shows that there is lots of "headroom" or room to spare in the memory usage. However, a lower number might indicate that a machine is bumping up against using all the memory on a regular basis.
For example, if a machine has RAM at 3% headroom, CPU at 50% headroom and IOPS at 60% headroom, then the critical factor on that machine is RAM. Memory is setting the bar for which cloud profile to select.
In terms of manual refinements - it makes sense to look across the set and see where there are low numbers and potentially "trade-up" to a profile with a higher capability.