Prioritisation Techniques Explained

Weighted Shortest Job First

WSJF: optimising for cost of delay

WSJF (Weighted Shortest Job First) is a prioritisation model from the Scaled Agile Framework (SAFe). It scores each job by dividing its Cost of Delay (CoD) by its Job Size — because shorter jobs with high CoD should almost always be done first. The magic is in how CoD is calculated: it is the sum of Business Value, Time Criticality, and Risk Reduction / Opportunity Enablement.

Business Value measures the relative benefit to users and the organisation. Time Criticality captures urgency — does delay hurt disproportionately? Risk Reduction / Opportunity Enablement scores what you gain by acting (or lose by waiting). Each is scored relative to the others in your backlog using a Fibonacci-style scale.

WSJF is ideal for Lean and SAFe teams with a continuous-flow backlog. It excels in environments where a queue of similar-sized jobs needs to be ordered by economic value. It requires discipline to score each cost-of-delay dimension consistently.

Maximum Difference Scaling

MaxDiff: bias-free priorities from real stakeholder choices

MaxDiff (Maximum Difference Scaling) is a research-grade prioritisation method where stakeholders do not rate items — they choose. In each round, they see a small subset of backlog items and identify the one that matters most and the one that matters least. Because it is always a comparative choice, not a rating, it eliminates the anchor bias and leniency bias that plague traditional scoring.

IbisFlow generates a balanced survey design: every item appears alongside different peers across multiple rounds. Stakeholders never feel overwhelmed because they only see a small set at once. After all rounds are complete, IbisFlow aggregates choices into a statistically sound priority ranking — not a simple average.

Rating scales ask people to judge items in isolation, which invites social desirability bias ("everything is important"), anchoring (first items inflate later ones), and scale differences between respondents (one person's 7 is another's 9). MaxDiff sidesteps all of this because you can only choose — you cannot hedge.

MaxDiff is the right choice when you have a large backlog (15–100+ items), multiple stakeholder groups whose priorities may differ, and when you want the result to be defensible and statistically robust. It works especially well for customer advisory boards, product reviews, or when internal gut-feel has lost credibility.