How much should we standardize our hardware?

I’d standardize the supported options, not insist on one exact machine for every person. A small lineup, such as a standard laptop, a lighter option, and a higher-performance model, keeps imaging, docks, and troubleshooting manageable without making every purchase a one-off.

Keep the configurations consistent within each option and test new models before making them generally available. That gives you some choice without multiplying the number of driver and firmware combinations you need to own.

There’s a real downside to putting the whole fleet on one model. A bad battery design, firmware release, or driver can turn into a company-wide problem. I’d keep some diversification, but in controlled groups rather than a grab bag of whatever happened to be available.

For example, use a primary model for most staff and a second qualified model for a defined part of the fleet. Track which devices have which firmware so a problem can be contained and updates can be staged.

Scale and support capacity matter a lot here. If your team has limited time, the support cost of many models can outweigh the small savings from chasing the cheapest configuration. A single vendor with a short list of business-class models is often a reasonable middle ground.

Also check how long a model stays available. A stable product line makes it easier to buy replacements and keep a consistent baseline, even if the fleet naturally includes several generations as devices are refreshed.

I’d separate the purchasing rule from the emergency rule. Pick a standard set of approved configurations, but define what happens when supply is tight or a model develops a recurring fault. Otherwise “buy the cheapest in stock” can quietly create a support zoo.

Keep a record of model, configuration, dock compatibility, and approved firmware. When a defect shows up, you can tell whether it is isolated or shared, and decide whether to pause that model instead of guessing across the whole fleet.

For regulated or tightly managed environments, fewer baselines can save a lot of validation and documentation work. I’d start with the minimum number of device types that actually cover people’s needs, then add another only when there’s a clear reason.

And don’t confuse standardization with never testing alternatives. A small pilot can catch compatibility problems before a purchase becomes a fleet decision.