Every office has them: two or three PCs that everyone hates and nobody can quite prove are broken. The DEX Score turns that grumbling into a ranked repair list.
An IT team with 200 PCs and time for five repairs a week has a prioritisation problem, not a knowledge problem. The DEX Score gives you the order: every device carries a 0 to 100 score built from performance, reliability, responsiveness, application reliability and user sentiment. Sort ascending, and the top of your list is where the pain is concentrated. This guide shows how to work that list honestly — including when not to trust a low score at face value.
👆 Start at the top of this list and stop when the scores turn acceptable (example devices)
A low score is a headline, not a diagnosis. Two PCs can both sit at 40 for completely different reasons — one drowns in application crashes, the other takes five minutes to boot. The score's job is to rank; the five pillars' job is to explain. Open the device, see which pillar is dragging it down, and you know whether you are looking at a hardware refresh, a software cleanup or a dying disk before you touch the machine.
ManageEngine's bands map neatly onto a working week. Poor (0–49) — fix or replace now; these users are suffering daily. Needs attention (50–69) — schedule the fix this month; left alone, they slide into poor. Acceptable (70–89) — watch the trend, no heroics. Excellent (90–100) — leave alone entirely. The discipline is resisting the loudest voice in the corridor: the squeaky wheel is not always the lowest score, and the score is the fairer queue.
Honest triage includes doubting the tool. A device that was switched off or off the network for days reports thin data, and ManageEngine notes that missing metrics can leave scoring incomplete — a laptop back from a week's leave may wear a misleading score until it has reported for a day or two. Also check device age and role: a nine-year-old PC scoring 55 may simply be at the end of its life, where the correct action is the refresh list, not another repair attempt.
Work the bottom of the list into three buckets. Fix in place — full disks, bloated startups, one bad app version, a battery at the end of its life. Refresh — old hardware whose pillars fail together; no single repair will save it. Replace the role — occasionally the PC is fine and the workload outgrew it, like accounts software on a machine bought for email. Writing the bucket next to each device converts a score list into a budget conversation the owner can actually approve.
The loop closes when the score moves. After a repair, watch the device for a week: boot time down, crashes gone, pillar recovered. If the score does not move, your fix did not fix the user's problem — keep investigating instead of closing the job on effort. This is the habit that separates score-driven IT from dashboard decoration, and it is the fastest way to find out whether DEX's view of your fleet matches your users' reality.
Watch the video to see a real score list being worked — then try sorting your own fleet and compare the bottom five with the PCs your users complain about.
▶️ Watch Full VideoAs a working rule from ManageEngine's bands: 0 to 49 is poor experience — fix or replace now. 50 to 69 needs attention — schedule it. 70 to 89 is acceptable, and 90 to 100 is excellent. Always read which pillar is low before deciding the action.
Possibly both. Check data freshness first — a device that was off for days may have an incomplete score. Then look at the pillar: battery health or device age can lower a score before the user feels anything. Treat it as an early warning, verify, and do not argue with the user.
Yes. The category and metric weightages are configurable, so an office that cares most about boot time or app stability can weight those higher. Change weights deliberately and note when you changed them, or trend comparisons lose meaning.