WSJF and cost of delay, in plain English

RICE and Value/Effort mostly ignore when work happens. For a lot of decisions that's fine. But some work has a clock on it — a launch window, a compliance deadline, a competitor about to ship — and for those, you want a framework that prices time. That's WSJF: Weighted Shortest Job First.

The formula

WSJF = Cost of Delay ÷ Job Size. You do the work with the highest cost of delay relative to how long it takes — the most time-valuable thing you can finish soonest. It comes from SAFe, but you don't need the rest of SAFe to use it.

Cost of delay, decomposed

Cost of delay is what it costs you for every unit of time this isn't done. It's usually estimated as the sum of three things:

• User/business value — how much value this delivers once shipped.
• Time criticality — does the value decay if you're late? A tax-season feature shipped in May is nearly worthless.
• Risk reduction / opportunity enablement — does this unblock other work or kill a major risk?

You score each on a relative scale (often a modified Fibonacci: 1, 2, 3, 5, 8, 13), add them, and divide by a similarly-scored job size.

Why the division matters

Dividing by job size is what makes WSJF clever. A high-value item that takes six months can be worse to start than three medium-value items you can each finish in a month — because finishing the quick ones sooner captures their value sooner and frees you up. WSJF formalizes the intuition that, all else equal, shorter jobs that deliver value now beat long jobs that deliver later.

The honest caveat

WSJF scores are relative, not absolute. The numbers only mean something compared to each other within the same exercise — don't treat a WSJF of 8 as a real dollar figure. Use it to rank a set of time-sensitive initiatives against one another, and lean on RICE when timing isn't the dominant factor.