Most teams describe the same pain using the same language: "We keep getting interrupted." "Too much unplanned work." "We cannot protect focus time."
That language is understandable, but it hides the real problem. In Kanban, this is not just an interruption problem. It is a demand problem.
When teams relabel "unplanned work" as demand entering the system, they stop treating it as random noise and start treating it as measurable system input. That shift matters because what can be measured can be reduced.
The Reframe: Failure Demand vs. Value Demand
In simple terms:
- Value demand is work that delivers new capability or planned customer value.
- Failure demand is work that arrives because something failed upstream: defects, rework, emergency fixes, preventable escalations, workflow breakages.
"Unplanned" sounds like weather. "Failure demand" sounds like engineering and operational debt that can be observed, categorized, and systematically reduced.
This is one of the highest-leverage language shifts I use with teams and leaders. The moment a leadership team sees failure demand as a category, instead of "just the way software works," priorities change.
Why This Reframe Improves Decision Quality
The old framing creates coping strategies: work later, work harder, add people, run more meetings.
The demand framing creates improvement strategies:
- Which failure-demand categories are growing?
- Which upstream process is generating the most avoidable rework?
- How much capacity is being consumed by preventable demand?
- Which policy change reduces that demand permanently?
Now your conversation is no longer "How do we survive interruptions?" It becomes "How do we reduce the system conditions creating reactive demand?"
How to Measure Demand vs. Capability
You do not need a complex model to start.
Step 1: Tag work items at intake
When work enters the board, classify each item as one of:
- Planned/value demand
- Unplanned/expedited demand
- Optional: failure-demand subtype (defect, environment break, rollback, support escalation)
The critical move is consistent tagging. Imperfect but consistent data beats perfect but missing data.
Step 2: Track the ratio weekly
Calculate:
- Planned ratio = planned items / total items completed
- Unplanned ratio = unplanned items / total items completed
Trend this weekly or monthly. You are looking for movement, not a single "good" week.
Step 3: Compare demand ratio with cycle time and throughput
When unplanned ratio spikes, cycle time and throughput usually degrade. Plotting these together exposes the cost of reactive demand in business terms, not just team frustration.
Step 4: Estimate capability consumed by failure demand
Show leaders a simple capacity view: if 49% of completed work is failure demand, roughly half of available engineering capacity is being consumed by system failure recovery instead of forward value.
This is where executive attention shifts fast.
What Healthy Ratios Usually Look Like
There is no universal benchmark, but I use practical working ranges:
- 20-30% unplanned: manageable in many complex environments
- 30-40%: warning zone, predictability begins to degrade
- 50-60%: reactive mode, planned delivery will keep slipping
The point is not to hit a perfect number. The point is to move the ratio in the right direction and stabilize it.
Real Pattern I See Repeatedly
In one engagement, failure demand started at 49% of total work. Almost half of available delivery capacity was being spent on reactive work.
Over several months, we made the ratio visible, reviewed it in monthly Service Delivery Reviews, and addressed root causes category by category. The ratio dropped to 41%, then eventually to 12%.
No new headcount was added.
That is the key lesson. Capacity improved because demand quality improved, not because the team expanded. When failure demand shrinks, planned work flows faster with the same team.
Why Leadership Should Care Immediately
When an executive sees that nearly half of delivery capacity is consumed by failure demand, they finally have a concrete lever:
- Invest in upstream quality controls
- Fix broken handoffs across teams
- Tighten release and rollback policies
- Fund automation where recurring failures are obvious
Without this ratio, leadership decisions are made on anecdote. With this ratio, decisions become traceable to measurable system constraints.
Practical Root-Cause Loop for Failure Demand
Once you can see the ratio, run a lightweight monthly loop.
- Identify top 3 failure-demand categories by volume.
- Pick one category and trace 5-10 recent examples.
- Find the repeatable failure mode, not the one-off event.
- Define one policy or engineering intervention.
- Measure category movement next review.
Examples of useful interventions:
- Add automated regression checks for a failure-prone area.
- Introduce explicit entry criteria before handoff to QA.
- Add a release readiness checklist for high-risk changes.
- Limit expedited lane WIP to prevent runaway interruption behavior.
The goal is not to eliminate all unplanned work. The goal is to stop treating preventable failure demand as inevitable.
Common Mistakes to Avoid
Mistake 1: Treating all unplanned work as equal
Not all unplanned work is failure demand. Some urgent demand is legitimate and valuable. Separate categories so you do not optimize away critical responsiveness.
Mistake 2: Tagging at close instead of intake
Retroactive tagging is noisy and biased. Tag when the item enters the system.
Mistake 3: Reviewing ratio without action
A chart alone changes nothing. Tie demand ratio reviews to explicit decisions and owners.
Mistake 4: Blaming teams for high failure demand
High failure demand is usually a system design signal, not an individual effort signal. Keep the conversation at system level.
How This Connects to Predictability
If your team keeps missing dates, start here. Predictability is mostly a function of:
- Stable demand profile
- Controlled WIP
- Reliable handoffs
- Consistent completion behavior
Demand ratio is one of the fastest ways to see whether those conditions exist.
This is why I pair demand ratio tracking with cycle-time and throughput views in the same monthly review. One chart tells you reactive load. The others show delivery consequences.
Where to Start This Week
If you want to implement this quickly:
- Add planned/unplanned tags to your board this week.
- Publish a weekly ratio trend for the next 6 weeks.
- Review the trend in a 30-minute monthly system review.
- Pick one failure-demand category and run one root-cause intervention.
That sequence is enough to start changing system behavior.
If you want the technical tactics that reduce time-in-system once demand is controlled, read How to Reduce Cycle Time: 5 Measurable Strategies. If you want to run the reporting cadence that keeps this visible for leadership, read The Service Delivery Review. If you need this set up as part of a broader diagnostic, the Process Diagnostic & Design service is the right entry point.