current and projected

The "and projected" is the future proofing I'm advocating for. A lot of companies plan to replace devices every 4 years, but they don't buy devices that will last that long. If you're buying PCs that have 2-year old CPUs in them out of the gate, then you're potentially losing 2 years off the end of their lifespan.

There's a lot of companies that don't factor their lead time into their planning either, so they only start the process after the devices hit 4 years. Particularly with big device batches each of the steps take longer; financing, stock backorders, imaging & testing, user cutover.

based on your monitoring data

Device replacement should be proactive, not reactive. If you're waiting for your monitoring data to show issues then it's already too late. It costs more to have someone on the payroll doing nothing for an hour each week (even in little increments) because their laptop is unuseable than it does to buy a new machine.

Literally, a person on $80K salary only needs to lose ~10 minutes of their average day to make up the ~$2K cost of a new laptop in one year. $80K = ~$55ph @ 40hr week, $2000 ÷ $55ph ÷ 48 working weeks a year ÷ 5 workdays a week x60 (convert hours to minutes).

That 5-minute logon storm caused by every sync/indexer/EDR/RMM agent kicking off a scan, the extra few minutes having to go hunt for a charger & wall socket because the battery can't hold a charge, the extra 30 seconds typing TOTP MFA codes into each of the 10 apps they use because the machine doesn't have a compatible TPM to support WHfB. They all add up.