DCEP study materials
DCEP study resources

Choosing a Benchmark When Metrics Disagree

DCEP · Put your knowledge to work in context.

All case studies

Choosing a Benchmark When Metrics Disagree

A 2.5 MW facility reports PUE 1.38, but its DCiE equivalent is about 72%. The operator also computes a partial PUE excluding UPS losses, which gives 1.30. A peer comparison shows similar facilities at PUE 1.25-1.45. The CFO wants a single benchmark figure for an annual report and asks whether the facility is 'efficient.' The facility uses evaporative cooling in a dry climate, runs at 55% average IT utilization, and has no waste-heat reuse. Reporting boundaries differ across the peer set.

1. How should the analyst reconcile PUE, DCiE, and partial PUE, and what should be reported?

Read the worked answer

PUE and DCiE are reciprocals (DCiE = 1/PUE), so 1.38 and ~72% are the same result expressed differently; presenting both is redundant and can confuse. Partial PUE of 1.30 is lower because it excludes UPS losses, so it is not comparable to full-boundary PUE and must not be placed alongside peer PUEs without labeling. Report the full-boundary PUE as the primary metric, state the boundary and measurement method, and disclose the partial figure only as a diagnostic with its exclusion noted. Reconcile the numbers explicitly so the CFO sees one consistent story rather than three competing figures.

2. Is the facility 'efficient' relative to peers, and what context must accompany the comparison?

Read the worked answer

PUE 1.38 sits within the peer range of 1.25-1.45, so it is broadly typical, not standout. But PUE is climate- and design-dependent: evaporative cooling in a dry climate lowers cooling energy, which helps PUE, while 55% utilization means fixed overhead is spread over less IT work, which hurts it. Peer boundaries differ, so a like-for-like comparison requires matching boundary, climate normalization, and utilization. The honest answer is that the facility is within the normal band, with a plausible explanation for both favorable (evaporative, dry climate) and unfavorable (low utilization, no heat reuse) factors. Recommend reporting PUE with these context factors rather than a bare 'efficient' verdict.