The problem with the conventional savings study
The format is familiar: a plant walkthrough, a list of measures, an estimated savings percentage for each and a payback period. It looks complete and is usually well intentioned.
The problem shows up when someone asks where the percentage came from. It almost always comes from a typical factor in the literature applied to annual consumption — not from a measurement at that plant. And when the measure is implemented and the savings do not appear, there is no way to know whether the measure failed, production changed, or the number was never realistic.
An investment committee that has lived through that once will not approve on the basis of that document again. And it is right.
Requirement one: adjusted baseline
Savings are a subtraction against something. That something is the baseline, and if it is not adjusted, the subtraction means nothing.
Adjusting means correcting for the variables that changed and have nothing to do with the project: production, temperature, number of shifts, working days. If your plant produced 18 % less in the project year, your consumption fell for reasons you owe to no one.
Without an adjusted baseline, any result is debatable, and it will be debated — usually by finance, usually the following year.
Requirement two: measurement and verification plan
The plan is written before implementation, not after. It defines what will be measured, with which instrument, at which point, how often, and how savings will be calculated from that.
The reference is IPMVP, the International Performance Measurement and Verification Protocol. It is not a decorative seal: it is the method that lets two parties with different interests —you and the provider, or you and whoever finances— arrive at the same number.
A study without an M&V plan can promise whatever it likes, because it does not define how it would be verified.
Requirement three: quantified uncertainty
A number without a range is an opinion with decimals. Every measurement has uncertainty, and stating it does not weaken the study: it makes it credible.
The difference between "will save 1.2 million a year" and "will save between 1.0 and 1.35 million a year, under these assumptions" is the difference between a brochure and a document you can make a decision with.
Whoever does not state their uncertainty either did not calculate it or would rather you not see it.
Requirement four: return measure by measure
A package with an overall return hides that inside it are measures that pay back in months and measures that never pay back, averaged into a single attractive number.
Separated, you can decide: implement the first three this year and leave the rest for when there is budget. Together, you have to take it or leave it.
That separation is also what lets you start small. Most projects that work started with the shortest-payback measure, not the biggest one.
What all this does in practice
It makes the study financeable. A document with an adjusted baseline, an M&V plan, stated uncertainty and return per measure is the one a committee approves and a financial institution can read.
And it changes the conversation with the provider. When the verification method is written in advance, the result stops being a matter of opinion at the end of the project.
It is more work and costs more than a checklist walkthrough. That is why we first ask what you need it for: if what you urgently need is a report to submit, that is a shorter, cheaper scope, and we tell you so.
Ask three things: how they will adjust the baseline, what the M&V plan is and what the uncertainty of the number is. If all three answers exist, you have an investment-grade study. If not, you have an estimate.
