Measuring realized savings
Track recommendations through implementation and measure what they actually saved from cost data, not from the original estimate.
Most cloud tools report estimated savings — the sum of what every open recommendation claims it would save. That number is a pipeline, not an outcome, and it does not survive scrutiny in a budget review.
Flux measures what actually happened.
The lifecycle #
Every recommendation carries a status an administrator can set:
| Status | Meaning |
|---|---|
| Open | Identified, no decision yet |
| Accepted | Agreed, not yet implemented |
| Implemented | The change has been made |
| Dismissed | Deliberately not doing this |
Dismissing is a real outcome. It takes findings out of the pipeline so the remaining number means something.
How realization is measured #
When a recommendation is marked implemented, Flux captures a baseline: the resource’s actual monthly run-rate at that moment.
From then on, realized saving is measured as:
realized = baseline − current actual cost (floored at zero)
Two properties matter here.
It comes from cost data, not the estimate. If a right-sizing was projected to save $400 and actually saved $260, Flux reports $260. If the change was never really made, it reports nothing saved.
It cannot go negative. If cost rose after the change, realized saving is zero rather than a negative number implying the recommendation cost you money. Cost may have risen for unrelated reasons; Flux does not claim to know.
The savings report #
The report shows the funnel and its outcome:
- Accepted count and estimated monthly value
- Implemented count and estimated monthly value
- Realized monthly — measured, and how many resources it was measured on
- Dismissed count
Plus a per-recommendation table of baseline, current cost, and realized saving.
The gap between estimated and realized is the interesting column. A large persistent gap means the estimation method needs revisiting — which is worth knowing.