How many vCPUs do affected VMs have? How many CPU cores do you have on your hosts?
If VMs have multiple vCPUs, check performance charts for those vms, pick Used and Ready counters (both in miliseconds) and check the values.
Typically, in absence of any context info, 2000ms or higher Ready (10% of 20 seconds) is considered to be a high value and usually indicates a scheduling problem. Decreasing number of vCPUs is usually the first recommendation. However if you are doing this, make sure not to cross HAL boundaries. I.e. if guest OS was installed on multiCPU system and is being switched to singleCPU, that normally requires changing Guest kernel. Moving from 4CPU to 2CPU (we're staying with SMP kernel and SMP architecture) usually is fairly safe, however you should consult your Guest OS manufacturers documentation.
As for resource pool creation, that is an expected behaviour. Once you create a cluster and add hosts to it, the cluster becomes the root resource pool and you can only create sub-pools under the cluster object (DRS needs to be turned on and licensed).
If your hosts are not CPU saturated, resource pools most probably will not solve the performance problem.
WBR
Imants