July flash sale
Cost of Delay is a conversation about time
Cost of Delay estimates the economic impact of waiting. It can include missed revenue, avoidable operating cost, customer harm, risk exposure, market timing, or the delayed learning that protects a larger investment. It is not simply a large number attached to a preferred feature. Its value comes from comparing how outcomes change when work starts or finishes at different times.
Draw the delay curve before choosing a score
Some opportunities lose value steadily. Others have a fixed-date cliff, a short market window, an accelerating risk, or little immediate loss followed by a threshold. Sketching this time profile prevents teams from treating all urgency as linear. State the horizon and unit, even when using relative values later in WSJF.
| Time profile | Typical example | Planning implication |
|---|---|---|
| Linear | Recurring manual cost continues each week | Compare accumulated avoidable cost |
| Fixed-date | Regulation or contractual event | Plan backwards with confidence and contingency |
| Market window | Season, launch, or expiring advantage | Value may collapse after the window |
| Risk acceleration | Security or reliability exposure grows | Include probability and impact range |
| Learning delay | Later evidence blocks a larger decision | Prioritize the test when it preserves options |
Estimate as a range
Use low, expected, and high cases instead of one impressive forecast. Document volume, affected population, benefit per event, probability, duration, and confidence. Separate money that is genuinely lost from revenue merely shifted to a later date. For customer or safety effects that should not be reduced to currency, use an explicit risk class and decision policy.
Install challenge rules before estimates arrive
- The proposer cannot be the only source for value and urgency.
- Finance, operations, risk, customer research, or architecture reviews assumptions relevant to them.
- The same unit and horizon apply to compared items.
- Benefits already counted elsewhere are not counted again.
- A number cannot override a legal, safety, or capacity boundary.
Worked comparison: automation versus renewal risk
Feature A avoids an estimated 400 to 600 staff hours per month and can start producing value whenever released. Feature B addresses a renewal risk whose decision occurs in eight weeks; after that date most benefit disappears. Even if Feature A has the larger annual total, Feature B may have higher near-term Cost of Delay. The correct sequence depends on delivery duration, confidence, opportunity window, and whether a smaller slice can protect the renewal.
Re-estimate only when evidence changes
Constant rescoring invites advocacy and churn. Revisit Cost of Delay when customer behavior, a deadline, risk probability, delivery size, or strategic context materially changes. Record the old assumption and new evidence. This preserves decision history and shows whether forecast quality improves.
From Cost of Delay to WSJF
WSJF compares Cost of Delay with job size. Both sides can be gamed: advocates inflate urgency and minimize effort. Use cross-functional estimates, ranges, and sensitivity analysis to see whether a plausible change reverses the ranking. If it does, invest in the evidence that resolves the decision rather than debating decimals.
The economic prioritization practices in SAFe POPM certification training help product roles apply these choices to features. Leading SAFe certification connects them to Lean economics and enterprise decision-making.
Decision record to keep
Capture the delay mechanism, time profile, range, assumptions, confidence, job-size source, chosen sequence, displaced work, and next review trigger. A defensible estimate is one that another group can inspect and update, not one that wins the room.
