Skip to content
Calcrivo

Cloud Optimization Score Calculator

Score how well a team turns optimisation recommendations into realised savings, and value the backlog left on the table.

Inputs

items
items
items
USD/month
USD/month
days
days
%
%
%

Newer families are usually cheaper per unit of work than the ones they replace.

Optimisation Score

40.7/ 100

Grade

D — Weak

Savings Realisation Rate

43.3%

Recommendation Action Rate

40.0%

Monthly Saving Left on the Table

$34,000.00

Annual Saving Left on the Table

$408,000.00

Delivery Velocity Subscore

33.3/ 100

Architecture Coverage Subscore

43.5/ 100

Where to Focus

Realisation — recommendations are being accepted but the savings are not landing

Step by step

  1. Values used

    Recommendations still open = 180 items; Recommendations actioned = 95 items; Recommendations dismissed with a reason = 25 items; Monthly saving identified = 60,000 USD/month; Monthly saving actually realised = 26,000 USD/month; Mean time to action a recommendation = 42 days; Target time to action = 14 days; Workloads right-sized in the last quarter = 55 %; Workloads behind auto-scaling = 40 %; Workloads on a current-generation instance family = 20 %

  2. Cloud Optimization Score

    score = realisation rate × 35% + action rate × 20% + (target days ÷ actual days) × 20% + architecture coverage × 25%.

  3. Optimisation Score

    = 40.7 / 100

  4. Grade

    = D — Weak

  5. Savings Realisation Rate

    = 43.3

  6. Recommendation Action Rate

    = 40.0

  7. Monthly Saving Left on the Table

    = 34,000.00

  8. Annual Saving Left on the Table

    = 408,000.00

How it works

Identifying savings is easy and free, so the score weights realisation most heavily and treats an explicit dismissal as a valid outcome — a documented decision not to act is far better than a recommendation quietly aging in a queue. Velocity is scored as a ratio of target to actual days, so beating the target cannot inflate the score beyond 100. Most cloud cost programmes fail in the gap between the recommendation and the change, and putting an annual dollar figure on the untouched backlog is usually what unlocks engineering time. The weights are a default starting point, not an industry standard.

Formula

Cloud Optimization Score

score = realisation rate × 35% + action rate × 20% + (target days ÷ actual days) × 20% + architecture coverage × 25%.

realisation rate
Savings actually banked divided by savings identified
action rate
Recommendations either actioned or explicitly dismissed, out of all raised
velocity subscore
Target time to action divided by the actual mean time, capped at 100

Frequently Asked Questions

How is Cloud Optimization Score calculated?

score = realisation rate × 35% + action rate × 20% + (target days ÷ actual days) × 20% + architecture coverage × 25%. Identifying savings is easy and free, so the score weights realisation most heavily and treats an explicit dismissal as a valid outcome — a documented decision not to act is far better than a recommendation quietly aging in a queue. Velocity is scored as a ratio of target to actual days, so beating the target cannot inflate the score beyond 100.

Why does Cloud Optimization Score matter?

Most cloud cost programmes fail in the gap between the recommendation and the change, and putting an annual dollar figure on the untouched backlog is usually what unlocks engineering time. The weights are a default starting point, not an industry standard.

What values do I need to enter?

This calculator takes 10 inputs: Recommendations still open, Recommendations actioned, Recommendations dismissed with a reason, Monthly saving identified, Monthly saving actually realised, Mean time to action a recommendation, Target time to action, Workloads right-sized in the last quarter, Workloads behind auto-scaling, Workloads on a current-generation instance family. The pre-filled defaults are a realistic starting point — replace them with figures from your own environment for a result you can act on.

Why does dismissing a recommendation count toward the action rate?

Many recommendations are wrong for good reasons — the instance is oversized for a known seasonal peak, or a reservation is being held for a migration. Recording that decision closes the loop and stops the same suggestion recurring, which is a genuinely better outcome than leaving it open forever.

You might also need