Fde Review Value

Check the Value and Decide What Happens Next

Use plain language. Explain any necessary technical word immediately.

Facts needed

Get the baseline and live-test dates, number and mix of items handled, completed work, cases sent to a person, staff minutes, quality, mistakes, sales or successful results, customer problems, staff corrections, extra profit or prevented loss, software and connection costs, review and support costs, and the one-time project cost.

Calculate

Compare like periods or exposure units. Convert every recurring benefit and cost to the same 30-day basis before combining them. Normalize shorter periods only when their volume and work mix are comparable; otherwise return wait and name the additional evidence required.

Monthly value = monthly extra profit + monthly losses prevented + monthly useful work created from freed time − monthly software cost − monthly connection cost − monthly staff review cost − monthly ongoing support cost

Payback days = one-time project cost ÷ real monthly value × 30

Count freed time only when it reduces spending, avoids hiring, completes more valuable work, or creates another measured result. Do not count the same benefit twice.

If real monthly value is zero or negative, report no payback instead of calculating payback days. Do not expand; choose stop when the result is conclusive or wait when the test is too small or too short.

Decide

• Expand: Quality holds, real results show payback within 30 days, value safely beats ongoing cost, failures stay limited, and the owner accepts responsibility.
• Fix once: One small, understood repair can likely make it pass without making the project bigger.
• Stop: Payback is too slow, value or volume is too low, trusted information is missing, staff will not use it, review uses up the benefit, or risk cannot be contained.
• Wait: The test was too small or too short; name the exact extra group or time needed.

Read [value-scorecard.md](references/value-scorecard.md) for the results sheet and ownership checklist.

Return

Return old versus new results, measurement periods and exposure, normalization method, assumptions, monthly value, payback days or no payback, quality and safety status, decision, next test size, owner, and next review date. Keep estimates separate from real results.

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: inbusiness23
• Source: inbusiness23/fde-skill
• License: MIT

Install and usage instructions live in the source repository linked above.