Wastes 5 & 6: Task Switching and Defects
The hidden costs of multitasking and the expensive rework of bugs.
1The Context Switching Tax
Every time you switch tasks, you pay a tax.
Research consistently shows that context switching costs 15-25 minutes to return to full productivity. Not to start working—to reach the same depth of focus you had before the interruption.
If you switch contexts 4 times in an afternoon, you might spend more time re-orienting than actually working.
In manufacturing, this waste is called "transportation" and "motion"—unnecessary movement of materials or people. In software, it's the unnecessary movement of attention.
Sources of task switching:
- Too much WIP: Juggling multiple projects means constant switching
- Interruptions: Slack messages, tapped shoulders, "quick questions"
- Meetings: Every meeting is a context switch away and back
- Email/notifications: Checking constantly fragments attention
- Unclear priorities: Not knowing what to work on leads to thrashing
The cost isn't just time—it's quality. Shallow work produces shallow results. Deep work requires sustained focus. Task switching prevents deep work.
The Multitasking Myth
Humans don't multitask—we task-switch. We're bad at it. What feels like productivity is actually cognitive overhead. The person 'getting lots done' by juggling tasks is often less productive than someone focused on one thing.
2The Defect Multiplier
Defects are waste for obvious reasons: they require rework. But the real cost is worse than it appears.
Defects found late cost exponentially more than defects found early. A bug caught in development might take 10 minutes to fix. The same bug found in testing might take an hour (reproduction, diagnosis, fix, re-test). Found in production, it might take days (investigation, customer communication, emergency fix, post-mortem).
Defects beget defects. Fixing a bug under pressure often introduces new bugs. The codebase becomes a minefield. Developers slow down. Quality spirals.
Defects destroy trust. After enough production incidents, every change requires extensive approval, extensive testing, extensive sign-off. The process bloats to protect against the defects the process creates.
The hidden waste: inspection and testing. If you have defects, you need extensive QA. QA is necessary waste—it doesn't add value, it just catches the waste created by defects. Build quality in, and you need less inspection.
A developer handles 'quick questions' throughout the day: Slack messages, tapped shoulders, email. They feel helpful. But their own work never reaches depth. Complex tasks take weeks. Bugs slip through because focus is fractured.
A team establishes 'focus time' from 10am-2pm: no meetings, no Slack, no interruptions. Complex work gets done in the morning. Communication happens in batched windows. Quality and velocity improve.
3Reducing Switching and Defects
For task switching:
WIP limits: The most direct intervention. If you can only work on one thing at a time, you can't context-switch. Finish before starting.
Batched communication: Check email/Slack at defined intervals, not constantly. Let people know when you're available and when you're not.
Meeting-free time: Designate blocks for focused work. Protect them fiercely.
Clearer priorities: If people don't know what to work on, they thrash. Make priorities explicit and stable.
For defects:
Test-driven development: Write tests first. Build quality in from the start, don't inspect it in at the end.
Pair programming: Real-time review catches errors before they become bugs.
Continuous integration: Find integration bugs immediately, when context is fresh and fixes are cheap.
Stop and fix: When you find a bug, fix it now. Don't add it to a backlog to grow stale. Toyota's andon cord principle.
Blameless postmortems: When bugs escape, understand why without blame. Fix the system that allowed the bug, not just the bug itself.
The Detection Fallacy
Adding more testing doesn't reduce defects—it detects them. And detection is expensive. The goal is to prevent defects from being created in the first place. Build quality in; don't inspect it in.
- Context switching costs 15-25 minutes per switch
- Multitasking is a myth—humans task-switch poorly
- Defects found late cost exponentially more than early ones
- Testing catches defects but doesn't prevent them
- WIP limits and TDD address root causes
- •Adding more QA instead of building quality in
- •Allowing interruptions because 'collaboration is important'
- •Treating defect backlogs as normal instead of alarming
- •Believing multitasking makes you more productive