A number you can defend to your own auditor.
Four stages, each one measured against data you already have — your own meters and published day-ahead prices — so a savings figure can always be traced back to its calculation, not taken on trust.
Baseline — locked before your pilot starts
We build a documented estimate of what your electricity cost would have been without Hetkin's recommendations, from your historical meter data and production pattern. Once locked, it doesn't move — every later comparison is against this same fixed reference, the same way a fixed budget doesn't get quietly revised after the fact.
Recommendations — same-day, delivery-risk labeled
We look for a cheaper window on the same calendar day as each flexible production block, with the same duration — never suggesting a shift that risks a delivery date. Every recommendation shows its projected savings and a delivery-risk rating (low / medium / high) side by side, so you're weighing the trade-off, not taking it on faith.
Verified savings — measured, not projected
Each month, we compare your actual metered cost against your locked baseline, scaled to that period. The difference is verified savings — a backward-looking, measured number, not a forecast. If a month goes the wrong way (a bad price spell, a schedule change outside your control), verified savings can come out lower than expected, or zero — never negative into a charge.
Shadow review — two internal checks before it's billed
A newly computed verification record doesn't go straight to your invoice. It's reviewed and confirmed internally, then approved a second time, before it's ever used to bill you. New pilots run in shadow mode — verification runs and is reviewed, but nothing is billed on it — for the first full billing cycle or two, so we're both confident in the number before it has financial weight.
We call our own methodology “provisional,” on purpose.
Every baseline, recommendation, and verification we generate today is labeled internally as a provisional methodology, with an explicit version number. That's not a hedge — it's the same rigor we'd want from a supplier of our own: as our models improve (more price-data history, more nuanced production constraints, cross-checks against a second market data source), the version changes and the change is visible, rather than a silent revision to how “savings” is defined underneath you.
Every savings assessment and invoice states which methodology version produced its numbers, so a figure from month one and a figure from month twelve are always comparable — or clearly flagged as not, if the methodology changed in between.
What this methodology deliberately doesn't do — yet
It doesn't model multi-day or cross-week shifts, only same-day windows — the lower-risk starting point while we're proving the model out with early design partners. It doesn't yet cross-check ENTSO-E price data against a second source. Both are real roadmap items, prioritized by what our first pilots actually need, not built ahead of that evidence.
See what your schedule is worth.
Applications are reviewed individually. Tell us about your process and we'll follow up with a qualification call.