glossary · prioritization
WSJF prioritization definition
A prioritization formula that ranks work by dividing Cost of Delay by job duration, so the highest-returning item per unit of time is built first.
WSJF is a sequencing formula, not a scoring rubric. The logic is economic: building things in the wrong order carries a cost, and that cost compounds across every item in the queue. The formula, WSJF = Cost of Delay / Job Size, answers one question: which item returns the most value per unit of time? Do it first. Don Reinertsen coined the original formulation as CD3 (Cost of Delay Divided by Duration) in “The Principles of Product Development Flow” (2009). SAFe adopted and adapted it for Portfolio Kanban sequencing at Program Increment (PI) planning.
The formula and its components
Cost of Delay (CoD) is the value lost for every period an item sits unshipped. SAFe breaks this into three Fibonacci-scored sub-components:
- User-Business Value: how much customers and the business benefit from this item being available now.
- Time Criticality: how fast value decays if delivery slips. A compliance deadline scores 13; an evergreen improvement scores 1 or 2.
- Risk Reduction / Opportunity Enablement (RR/OE): whether building this reduces a future risk or unblocks options the team cannot take without it.
Job Size is relative duration, also Fibonacci-scored. CoD = UBV + TC + RR/OE. WSJF = CoD / Job Size.
One item in each category must anchor at 1. Without an anchor, all scores inflate and the proportional relationships that make Fibonacci useful collapse. The anchored item defines the scale; everything else is relative to it.
Worked example
| Feature | UBV | TC | RR/OE | CoD | Job Size | WSJF |
|---|---|---|---|---|---|---|
| Compliance audit log | 3 | 13 | 8 | 24 | 5 | 4.8 |
| Onboarding redesign | 8 | 3 | 2 | 13 | 3 | 4.3 |
| Bulk export to CSV | 5 | 2 | 1 | 8 | 8 | 1.0 |
The compliance log ranks first despite lower user-business value because the regulatory window is binary and the item unblocks downstream work. A team following intuition would ship the onboarding redesign first and miss the deadline.
The Reinertsen vs SAFe divergence
Reinertsen’s CD3 uses actual dollars: how many dollars per week does delay cost? Black Swan Farming’s example makes the stakes concrete: FIFO scheduling costs $69,000 in accumulated delay; CD3-sequenced delivery costs $27,000, a 61% reduction. That math works only when CoD is real money.
SAFe replaced dollars with relative composite scores. Jason Yip documented the consequence: a composite CoD of 9 tells you nothing about whether the feature deserves budget. You can rank items but lose the economic grounding that makes the ranking defensible outside the room. At SAFe shops, relative scoring is the expected answer. At lean or AI-native companies, dollar-denominated CoD reads as more rigorous.
When to use WSJF vs alternatives
WSJF is the right lens when delay cost varies sharply across items: regulatory deadlines, competitor windows, platform changes that become irreversible once someone else ships. Use RICE when your backlog is heterogeneous and confidence is a meaningful variable (WSJF has no uncertainty dimension). Use ICE for a fast initial cull. Use MoSCoW for fixed-scope releases with a hard boundary, not for continuous sequencing.
WSJF in 2026: the numerator dominates
AI tooling has collapsed job duration estimates. A feature that required a 13-point epic now spikes as a 3 when a vibe-coded prototype settles the core design question in a day. When Job Size shrinks across the whole backlog, the Cost of Delay numerator dominates rankings. Productboard now auto-scores CoD components by pulling usage data, revenue impact estimates, and segment weighting, shifting the input from gut-feel to structured data. The PM’s job shifts from filling out the rubric to auditing whether the inputs reflect what is actually true about the market.
The conceptual shift matters more than the tooling. In 2026, the question “can we build this?” rarely blocks a decision. The question WSJF forces is the one that does: what does it cost us every week we do not ship this? WSJF is less a scoring ritual and more a forcing function for that question.
What clears the bar in the interview
weak
"WSJF equals Cost of Delay divided by job size, where Cost of Delay is user-business value plus time criticality plus risk reduction, each scored on a Fibonacci scale." Reciting the formula and stopping signals no judgment: you have not explained how you score CoD in practice, not acknowledged that SAFe's relative scale loses economic meaning, and not defended a trade-off under pushback. At non-SAFe companies this can read as needing a process scaffold to think economically rather than thinking economically first.
strong
"WSJF is rooted in Reinertsen's insight that building things in the wrong order has a real cost: accumulated delay across everything waiting in the queue. The formula, Cost of Delay divided by duration, tells you which item returns the most per week of team time. For a recent PI I scored three features: a compliance audit log, an onboarding redesign, and a bulk export. The audit log ranked first despite lower user-business value because the regulatory window was binary and it unblocked downstream work. The team's intuition was to ship the redesign first; walking through the time-criticality scoring changed that. One flag on SAFe's version: relative composite scores let you rank, but they mask whether the absolute CoD justifies the investment. Where I can, I push for dollar-per-week estimates so the ranking holds in a budget conversation, not just in grooming. I also reassess scores mid-increment if a competitor ships or a deadline moves, because WSJF has no uncertainty dimension and CoD estimates can go stale within a sprint."
For the full backlog prioritization prompt and follow-ups, see how do you prioritize your backlog?. For sibling frameworks, see RICE and MoSCoW.