ISO 50001 EnPI and energy baseline: choosing indicators, normalising data and proving improvement

What do ISO 50001 clauses 6.4 and 6.5 require?

ISO 50001 sets no target percentage, but it does expect improvement shown in numbers, and that argument rests on an indicator and a reference point. Clauses 6.4 and 6.5 define both, between the energy review (6.3) and the data collection plan (6.6). In plain terms they expect:

• Fit-for-purpose indicators. EnPIs must measure and monitor energy performance and allow improvement to be demonstrated.
• A documented method. How each EnPI is determined and updated is kept as documented information.
• A baseline built from the energy review. The EnB uses the review's information over a suitable period.
• Normalisation where it matters. When relevant variables significantly affect energy performance, both the EnPI values and the corresponding baselines are normalised.
• Revision rules. Set out in the revision section below.

Leadership has a part too: clause 5.1 makes top management responsible for ensuring the EnPIs appropriately represent energy performance.

How do you choose EnPIs that show real improvement?

Start from the energy review. It tells you which significant energy uses (SEUs) matter, which variables drive each one and who operates them; how that list is built is covered in The energy review and significant energy uses (clause 6.3). Indicators then follow two levels.

SEU level

Each SEU, whether a chiller plant, a furnace line or a fleet, needs an indicator that responds to how well it is run. If output or temperature drives its consumption, the EnPI must account for that driver, or operators will stop trusting a number that moves for reasons they cannot influence.

Organisation level

One or two indicators summarise performance across the scope; this is the figure management review reads.

Keep a candidate indicator only if it can be calculated from data you collect (or that The energy data collection plan and metering (clause 6.6) will add), moves when efficiency changes but not when only activity does, and makes sense to the people who affect it.

ISO 50006:2023, Energy management systems — Evaluating energy performance using energy performance indicators and energy baselines, is the ISO guidance on this task; certification remains to ISO 50001 alone.

How is an energy baseline set?

The baseline is a defined period of measured energy data, with the matching data for each relevant variable, against which later performance is compared. ISO 50001 requires a suitable period rather than a fixed length. In practice the period should cover a full operating cycle, so that seasons, shutdowns and product mix are all represented; for most sites that means at least a year of monthly or finer data.

Avoid years distorted by a long breakdown, an unusual order book or a major project, and record why the period was chosen, which meters and invoices fed it and how gaps were filled. Each EnPI needs its own baseline, with the same boundaries as the indicator it serves.

What is normalisation, and when is it required?

Normalisation adjusts energy figures so that the baseline period and the reporting period are compared under equivalent conditions. It is required where relevant variables, the factors that change routinely and move consumption, such as production volume, outdoor temperature or occupancy, have a significant effect. Static factors, such as floor area or installed plant, are not normalised routinely; a major change to them is a reason to revise the baseline instead.

The usual approach is to fit a simple model to the baseline data that predicts consumption from the relevant variables, then use it to calculate what the reporting period would have consumed under baseline performance. The difference from the measured figure is the normalised result. The example below is illustrative; the numbers are invented to show the method, not taken from a client.

The raw figure says things got worse; the ratio overstates the gain because the fixed load is spread over more tonnes; only the model separates volume from better operation. Keep the model, its data and fit statistics with the baseline, and state when it will be refitted.

When must the baseline be revised?

ISO 50001 names three triggers for revising an EnB:

• the EnPIs no longer reflect the organisation's energy performance, for example because the model no longer fits the data;
• there have been major changes to static factors, such as a new production line, a building extension or a closed site;
• a pre-determined method says so, such as a rule written into the documented method at the outset.

Write the rule in advance: what counts as a major change, who approves, and how old and new baselines are linked so achieved improvement stays on record. Management review, whose outputs include decisions on EnPIs and EnBs, is the natural approver. Disappointing results are not a trigger.

What the certification body checks

Certification to ISO 50001 is carried out by an accredited certification body working to ISO 50003:2021. Sterling does not certify. Certification bodies' published guidance says improvement must be shown at initial certification and each recertification by comparing EnPIs with baselines, with surveillance focused on the improvement actions; a fall in total consumption counts only where conditions have not changed significantly. ISO 50003 also treats audit evidence that energy performance improvement was not achieved as grounds for a major nonconformity.

For measurement and verification of savings, ISO 50015:2014 is the related ISO guidance.

Mistakes that undermine EnPIs and baselines

• Total consumption as the only indicator. It works only where no driver of energy use varies.
• Ratios with a large fixed load. kWh per unit improves by itself when output rises, as the example shows, and worsens when it falls.
• Variables chosen by assumption. Production is assumed to drive consumption, but the data show that weather or shift pattern matters more.
• No documented method. The number lives in one person's spreadsheet.
• Mismatched boundaries. The EnPI covers one building while its baseline includes the whole site, or a sub-meter is added without adjusting the baseline.
• Silent baseline changes. The reference year moves without a trigger, an approval or a link to the earlier record.

EnPIs and energy baselines: questions buyers ask