Energy baseline (ISO 50001)

The energy baseline is your reference point: this is where we stand now. You compare all improvements that you measure with your EnPIs against this baseline. Without a baseline, you cannot demonstrate that your energy performance has actually improved, and that is exactly what ISO 50001:2018 asks of you (clause 8.2). The NEN version of the standard refers to it as energiereferentielijn (EnB).

The energy baseline is not the same as your energy target. The baseline is your starting point (where are you now?), and the target is your end point (where do you want to go?). Your EnPIs measure the progress in between.

What is an energy baseline under ISO 50001?

An energy baseline (English: Energy Baseline, abbreviated EnB) is a quantitative reference point against which you compare your energy performance. Usually you choose a recent full year as the baseline, for example 2024, and in the years that follow you measure whether you perform better or worse.

It sounds simple: “We used 520 MWh in 2024, now 480 MWh, so we are saving 8%.” In practice, it is more complex. What if you produce 20% more in 2025? Or if it was a harsh winter? Then an absolute figure is misleading. That is why you must normalise.

Why normalise?

Normalising means correcting your use for factors you cannot influence. This way, you compare like with like.

Example without normalisation:

Year	Use	Production	Conclusion
2024 (baseline)	520 MWh	50,000 units	Reference year
2025	572 MWh	60,000 units	+10% use?

Without normalisation, it looks as though you are getting worse. But you produce 20% more.

Example with normalisation:

Year	Use	Production	kWh per unit	Conclusion
2024 (baseline)	520 MWh	50,000 units	10.4 kWh/unit	Reference year
2025	572 MWh	60,000 units	9.5 kWh/unit	-8.5% per unit

Now you see that you have become more efficient per product. That is what the auditor wants to see.

Step by step: setting the energy baseline

Step 1: Choose the baseline period

Choose a period that is representative of your normal operations. Usually a full calendar year.

Choice	When it suits
Previous calendar year (for example, 2024)	Standard choice, full seasonal cycle
Last 12 months (rolling)	If you start in the middle of the year
Average of 2 to 3 years	If you have large annual fluctuations

Avoid using a year with exceptional circumstances (COVID lockdown, major renovation, long production stop) as your only baseline.

Step 2: Identify relevant variables

Which factors affect your energy use but that you cannot control?

Common variables:

Variable	Relevant for	Example
Production volume	Production companies	More orders = more use
Outdoor temperature (degree days)	Buildings, HVAC	Cold winter = more gas
Occupancy rate	Offices, hotels	More guests = more use
Operating hours	All sectors	More running hours = more use
Floor area	Real estate	More m² = more heating

You do not have to include all variables. Choose the 1 to 3 that have the greatest impact on your use.

Step 3: Determine the relationship between use and variables

You want to know: if my production volume rises by 10%, how much does my use rise?

Simple method (ratio): Divide use by the variable. Example: 520 MWh ÷ 50,000 units = 10.4 kWh per unit. This works well if the relationship is linear.

Advanced method (regression analysis): Use historical monthly data to build a statistical model. Excel can do this (scatter plot with a trend line). This works better if you have several variables or if the relationship is not linear.

Regression example: You plot monthly electricity use (Y-axis) against production volume (X-axis) for 24 months. The trend line gives you: Use = 15 MWh + 0.0095 × production volume. Now you can calculate the “expected” use for any production volume.

Step 4: Document the baseline

Record:

Baseline period: Which year or period
Total use: Per energy type (electricity, gas, etc.)
Variables: Which variables you use for normalisation
Normalisation method: Ratio or regression, with formulae
Reference values: The values of your variables in the baseline period

This document is part of your EnMS documentation and is checked at the audit.

Step 5: Calculate the normalised baseline

For each measurement point (month, quarter), you calculate the “expected use” based on your baseline and the current value of your variables. The difference between expected and actual use is your improvement (or deterioration).

Example:

Month	Production	Expected use (baseline)	Actual use	Difference
Jan 2025	4,500 units	46.8 MWh	43.2 MWh	-7.7%
Feb 2025	5,200 units	54.1 MWh	50.8 MWh	-6.1%
Mar 2025	4,800 units	49.9 MWh	46.5 MWh	-6.8%

You see that, corrected for production, you consistently use 6 to 8% less than your baseline. That is the improvement you can claim.

When do you adjust the energy baseline?

The baseline is not set in stone. ISO 50001 requires you to adjust it if:

Situation	Action
Major change in processes (new production line, closure of a department)	Revise the baseline
EnPIs no longer reflect reality	Revise the baseline
Method error discovered (wrong data, incorrect formula)	Correct the baseline
Regular improvement through measures	Do not adjust the baseline

Important: Do not adjust the baseline just because you are missing your targets. The baseline is your honest reference point. Adjusting it when results disappoint is like lowering the bar after missing a jump.

The auditor checks whether your baseline was set honestly. If you adjust it regularly without good reason, that is a red flag. Always document why you make an adjustment.

Examples per sector

Production company

Element	Value
Baseline period	2024
Electricity	520 MWh
Gas	80,000 m³
Normalisation variable	Production volume (50,000 units)
EnPI	10.4 kWh/unit (electricity), 1.6 m³/unit (gas)
Target 2027	8.8 kWh/unit (-15%)

Data centre

Element	Value
Baseline period	2024
Total use	8,500 MWh
IT load	5,300 MWh
Normalisation variable	IT load (as this determines cooling demand)
EnPI	PUE = 1.60 (total ÷ IT load)
Target 2027	PUE = 1.35 (-16%)

Office building

Element	Value
Baseline period	2024
Total use	180 MWh electricity, 40,000 m³ gas
Normalisation variables	Degree days (weather), occupancy rate
EnPI	18 kWh/m²/year (electricity), 4.0 m³/m²/year (gas)
Target 2027	15 kWh/m²/year (-17%)

Common mistakes with the ISO 50001 energy baseline

Using only absolute use as the baseline

“We used 520 MWh in 2024” without normalisation. If your company grows or shrinks, this figure says nothing about efficiency. Always normalise for at least one relevant variable.

Choosing the wrong baseline period

Choosing a year with a renovation lasting three months as your baseline gives a distorted picture. Choose a representative year or correct for the exception.

Using too many variables

Three normalisation variables is often the maximum. More makes the model complex and hard to explain. The auditor must be able to follow your method.

Never revising the baseline

After 5 years with the same baseline, the ratios may no longer be correct (different products, different machines, a different building). Revise it after major changes.

Relationship with other ISO 50001 elements

The energy baseline does not stand alone. Everything is connected:

Element	Relationship with the baseline
Energy review	Provides the data for the baseline
SEUs	A baseline per SEU enables targeted control
EnPIs	Compare current performance with the baseline
Objectives	Formulated as a % improvement compared with the baseline
Audit	Auditor checks whether the baseline is correct and honest