[ Solution ]
Energy Performance & ISO 50001
Consumption fell last month. Did you save energy, or did production fall? Every energy report that cannot separate the two tells management the wrong story.
Raw consumption lies
Most energy reports show a single curve: monthly kWh. When it falls, everyone celebrates — yet the most common reason it falls is not efficiency but less production that month. The reverse holds too: most months spent hunting waste because consumption rose were simply months with more orders.
That is why energy performance is measured not on raw consumption but on consumption normalised against its driver: kWh per tonne, kWh per room-night, kWh per degree-day. This is exactly what ISO 50001 calls an EnPI.
The module’s job is three steps: build a baseline from historical data, hold each month’s actual against that line, and track the difference cumulatively. Without all three, “we saved energy” is a claim without evidence.
How far off expectation is this month?
Once the baseline stands, every month asks a single question: was this consumption normal for this production volume?
Monthly Deviation Analysis
actual − expectedThe threshold is visible too: any deviation beyond ±5% is flagged as work to investigate and raises a notification. January’s −9.9% looks like good news, but it is investigated with the same seriousness — consumption below expectation can also be the trace of a modelling gap, a meter fault or an unrecorded shutdown.
The baseline: where does “expected” come from?
Expected consumption is not a guess, it is a regression: for every month of the reference period, production is paired with consumption and the relationship is written as an equation.
Next to the equation stands one more number, and we never hide it: R² — how much of the consumption the model explains. In the real example below, R² is only 0.14. That is a poor result, and the panel says so openly.
Why it matters: claiming savings against a low-R² baseline is like losing weight on broken scales. A low R² tells you a driver is missing from the model — degree-days, product mix, shift count. The right response is not to decorate the equation but to find and add the missing variable. If the panel hid this for you, it is you who would be caught out in an audit or a board meeting.
Regression: the relationship, visible
Regression Analysis
reference period 2023 · 14 observations · R² = 0.14Looking at the scatter makes the R² concrete: at the same tonnage the site consumed 72 thousand one month and 105 thousand another. That spread is the footprint of a factor the model has not been given. The module recommends adding a second variable — and every time the equation is updated, the old baseline stays frozen in the archive. Past reports never change retroactively.
CUSUM: from single-month noise to a cumulative signal
Monthly deviations are individually noisy; one month looks good, the next looks bad. The real question is: cumulatively, where are we heading?
CUSUM — the cumulative sum of deviations — stacks each month’s (actual − expected) on the last. If the curve is descending, savings are accumulating; if it flattens, the improvement has stopped working; if it turns upward, something has broken — and a single-month report will never show it.
The cumulative result, on one screen
CUSUM · Cumulative Deviation
Σ(actual − expected) · downward = savingSavings rate
3.1%
since the period began
Cumulative savings
5,890
kWh
EnPI
60.3
kWh/tonne · actual
SEC target
61.85
kWh/tonne · under target ✓
Figures are from our own pilot site (2026), normalised against production data.
What this means for ISO 50001
These are exactly the three things ISO 50001 requires you to prove to an auditor: the EnB (energy baseline), the EnPIs (performance indicators), and the measured effect of improvement.
The module produces all three in audit language:
- The baseline is frozen. Once the EnB is established and approved it does not change; if a revision is needed, a new version is opened and the old one stays in the archive with its date. This is the first thing an auditor looks at.
- EnPIs are tied to drivers. kWh/tonne, kWh/m², kWh/degree-day — defined per site, never hand-calculated.
- Deviation management is on record. Every deviation beyond the threshold sits with its explanation and closure note. The answer to “what do you do about significant deviations?” is shown on screen.
If you hold no certificate, the same mechanics still run — it is simply called evidence for management instead of an audit.
What we measure, and from where
Measurement provenance
- Consumption
- kWh
- Main meter and area analysers · hourly
- Production
- tonnes · units
- ERP/MES link or periodic entry
- Degree-days
- °C·day
- Hourly weather · by site location
- Baseline
- equation
- Regression · published with R² and observation count
- Deviation
- %
- Derived · actual − expected
- CUSUM
- kWh
- Derived · cumulative deviation
- EnPI
- kWh/driver
- Defined per site · tied to the EnB version
The link to ESOS and funded projects
If you are planning an improvement that needs investment — compressor replacement, heat recovery, drive conversion — what the business case asks for is exactly what this module produces: reference-period consumption, a normalised baseline and verified savings.
The same numbers serve ESOS reporting: the scheme wants credible consumption analysis and evidence that recommendations deliver. The application and the post-implementation verification speak the same method, against the same baseline.
Frequently asked
We are not ISO 50001 certified. What does the module tell us?
The same thing — just to management instead of an auditor: how much of the change in consumption came from production, and how much from efficiency. If you start the certification process, the EnB and EnPI infrastructure is already in place; if you never do, the monthly performance conversation still moves from guesswork to measurement.
What do you do when R² comes out low?
First, we say so — we do not hide it. Then we hunt the missing variable: degree-days at a cooling-heavy site, product mix on a multi-product line, shift count in shift operations. When a variable is added and the model strengthens, a new baseline version is opened; the old version and every report produced with it stay in the archive untouched. We do not rewrite history.
Why freeze the baseline? Isn't updating it as the model improves more accurate?
The model improves — as versions. If the baseline is reworked every month, the 'savings' figure changes retroactively every month too, and that destroys trust in front of auditors and boards alike. This is precisely ISO 50001's EnB logic: the reference is fixed; revisions are dated and justified.
We don't want to hand over production data. Does the module work without it?
In a limited way. Without driver data there is no normalisation; what remains is raw-consumption comparison — the very trap in this page's first paragraph. For sites that will not share absolute tonnes, we run indexed: a production index relative to the reference period does the same job, and no commercial data leaves the site.
Are the numbers on screen real?
Yes — the regression and deviation analysis are the real 2023-referenced data of our own pilot site; that is exactly why an unflattering R² of 0.14 is left on screen. Customer site data is never used on this website.
Let us measure what is happening on your site.
In a one-hour call we look at your existing setup and set out exactly which measurement points are needed and what you would be able to see.