KNOWN
Directly supported by supplied or source data. Public records remain source-reported information, not independently verified facts about current operations.
Built Matters methodology
Evidence
before conclusions.
A useful finding explains what was observed, how it was calculated, what it might mean and what remains to be established. These principles define the intended Built Matters Energy analysis.
Data sources
The conclusion starts with the source.
Potential inputs include customer interval energy data, tariffs and bills, operating schedules, production or output data, and optional submeter, BMS or equipment information.
Relevant public or open data, published ESOS information and weather data can provide context where appropriate. Not every source will be available for every operation.
Source, period, meter boundary, units, missing intervals and operational coverage should be recorded. Historical public information and current customer data answer different questions and must not be treated as interchangeable.
The evidence model
Four labels.
Different kinds of conclusion.
CALCULATED
Derived transparently from the available data. The method, period and assumptions should explain how the figure was reached.
INFERRED
Built Matters’ interpretation of the evidence. An explanation to test, distinct from a directly observed fact.
INVESTIGATE
A plausible issue requiring additional evidence or specialist assessment before action.
Confidence
Precision is not the same as certainty.
Confidence should reflect completeness, consistency, coverage and the strength of the explanation. A directly measured pattern can have strong evidence while its physical cause remains uncertain.
Findings should show the relevant assumptions and missing evidence. Cost calculations depend on the tariff, period and charge basis used. Estimates of avoidable expenditure need to distinguish necessary consumption from a plausible opportunity.
AI can help interpret data and organise findings. It does not remove the need to show the calculation, qualify inference or seek professional judgement where appropriate.
Limitations
A pattern does not prove a physical cause.
- Interval data may reveal when a pattern occurs without identifying the precise machine or system.
- Correlation with weather or production does not establish causation.
- Incomplete or inconsistent data can reduce confidence or prevent a useful estimate.
- Energy-price assumptions affect financial calculations.
- Weather, production mix and operational changes can confound before-and-after comparisons.
- Specialist engineering investigation may be needed before operational or capital changes.
PROVE methodology
Baseline. Intervention. Subsequent consumption.
Define an appropriate baseline and record what changed, when it changed and which meters or operating boundaries are relevant. Compare subsequent consumption over a meaningful period.
Where data permits, consider weather, production volumes, schedules and other relevant influences. Where it does not, make the limitation explicit rather than attributing every difference to the intervention.
Separate measured consumption changes from their estimated financial effect. A tariff change can alter cost without altering consumption. Evidence of savings should state the comparison, assumptions and remaining uncertainty, not promise scientific certainty.
Independence
No incentive to sell a particular intervention.
Built Matters does not sell equipment, energy contracts or FM services, and receives no supplier commission. It has no commercial incentive to recommend a particular physical intervention.
When a finding needs specialist work, the intended Investigation Brief explains the question and evidence. The customer chooses who investigates and decides what to implement.
Practical questions
Data, evidence and confidence.
How does Built Matters analyse energy consumption?
The approach checks data coverage and units, examines consumption over time and relates patterns to operating context. Relevant patterns are translated into financial implications and prioritised, with assumptions and evidence distinguished from interpretation.
What is the difference between KNOWN, CALCULATED, INFERRED and INVESTIGATE?
KNOWN is directly supported by supplied or source data. CALCULATED is derived using an explicit method and assumptions. INFERRED is an interpretation to test. INVESTIGATE marks an issue needing additional evidence or specialist assessment. Source-reported public information is not independently verified current fact.
How does Built Matters calculate financial impact?
Relevant consumption is combined with applicable tariff information over a stated period. Estimates must identify the rate, baseline and assumptions. Consumption cost is not automatically avoidable cost, and an opportunity estimate is not a guaranteed saving.
How does Built Matters measure whether a saving happened?
Compare consumption after an intervention with a suitable baseline, considering weather, production, operating hours and other changes where data permits. State the comparison method and uncertainty, and separate energy changes from the financial effect of tariff changes.
Can half-hourly data identify which machine caused high consumption?
Not on its own. A whole-site meter records combined demand. Timing can suggest questions to investigate, but assigning consumption to a machine requires supporting evidence such as submetering, equipment logs or a specialist inspection.
How does weather affect energy analysis?
Heating and cooling demand can change with weather even when efficiency has not changed. Where relevant and supported by data, weather should inform the baseline comparison; otherwise it remains a limitation on attribution.
How does production volume affect energy analysis?
Higher output can raise total energy use without worsening efficiency. Production volume, product mix, shifts and downtime help explain changes. Comparing matched periods or energy per unit can add context, but only where output measures are comparable.
What does a confidence level mean?
Confidence describes how strongly the available evidence supports a finding. Coverage, consistency, assumptions and alternative explanations all matter. It is not a guarantee or a calibrated probability unless a specific statistical method establishes one.
What public data can Built Matters use?
Published ESOS information can provide context about reported consumption, opportunities, plans and progress where available. Each source needs a reporting period and clear attribution. Current supplied data is needed to understand what is happening now.
How accurate are Built Matters findings?
Accuracy depends on the supplied data and the question being asked. Transparent calculations can be checked; interpretations remain subject to uncertainty. Missing data or changing conditions limit conclusions, and the product does not promise a universal accuracy percentage.
When is specialist engineering investigation required?
Use a qualified specialist where a finding requires physical diagnosis, safety assessment, equipment design or implementation advice. A consumption pattern can justify investigation without establishing its cause or the correct engineering intervention.
From evidence to action
Find the opportunity.
Decide what matters.
Prove what changed.
Independent energy intelligence for the decisions that affect your operating costs.