Four task dimensions that locate friction

'Customers find it difficult' is a dead-end diagnosis; you cannot fix 'difficult'. The Task Dimensions model splits difficulty into four nameable kinds of friction. Manually intensive steps demand too many actions or too much effort. High cognitive load steps force people to think, remember, or interpret when they shouldn't have to. Error-prone steps produce the same predictable mistakes from customer after customer. Knowledge-dependent steps assume people know account numbers, policies, or rules they simply don't. Run any struggling step through these four lenses and the vague complaint becomes a specific defect with an obvious direction for the fix.

Why it matters to the business

Each dimension is a form of customer effort, and effort is the strongest known predictor of disloyalty. CEB research behind HBR's 'Stop Trying to Delight Your Customers' found the Customer Effort Score predicts loyalty better than CSAT or NPS, and follow-up work reported that 96% of customers who have high-effort experiences become more disloyal, against 9% after low-effort ones. The most dangerous dimension is knowledge dependence, because it is invisible to the people who built the system. They already have the knowledge, so the step feels easy in every internal demo and fails in the field.

How to use it

• Score each step of your top three journeys against the four dimensions; a simple red-amber-green grid works.
• Pull your most repeated support contacts and error states, and tag each with the dimension that caused it.
• For error-prone steps, redesign upstream so the mistake cannot happen; an error message is a symptom, not a fix.
• For knowledge-dependent steps, make the system recall or explain whatever it currently expects customers to know.
• Re-score after each release to prove friction is actually falling.

Where teams get it wrong

Teams respond to predictable errors with better error messages, and to knowledge gaps with training and tooltips. Both treat the customer as the defect. If confident users make errors, the system is the problem, and polishing the failure notice does not reduce the failures.

Ask your team

• What are our five most common customer errors, and what upstream design change would make each one impossible?
• Which steps in our core journey assume knowledge, such as codes, policies, or terminology, that a first-time customer won't have?
• When we last tested a flow, did we watch outsiders, or insiders who already knew the answers?

If confident users make errors, the system is the problem.