Altitude: Team Diagnostic Frame
The 20 Team Gaps
The same friction comes up in every retrospective: requirements that bounce back, work stuck behind another team, priorities that change mid-iteration. Each fix you have tried has faded. The friction has specific causes, and this page names 20 of them.
What a gap is
A gap is a condition that prevents your team from delivering value, and that you can observe in how work arrives, how it moves, and what comes back once it ships.
Explanations like “process,” “discipline” and “we need to do the framework properly” name no condition the team can see and change. So the next retrospective reaches for a new board, a work in progress (WIP) limit or a new retro format, and the old friction returns.
As you read, picture the team at this morning’s standup, not the team described in last quarter’s health check. Mark every gap you have watched happen on that team; one incident can earn more than one mark.
Two questions for every gap
The first question is which barrier a gap sits in:
- Alignment Drift breaks direction. The team cannot connect its work to an outcome.
- Choked Flow breaks movement. Work starts faster than the system lets it finish, because it waits on approvals, dependencies and handoffs.
- Broken Feedback breaks learning. The loop between work and learning fails to close, in a retrospective, a review or after release.
The second question is whose move it is. Every gap has a part the team can change with the authority it actually holds, and a part that needs whoever controls the constraint. For example, a ceremony the team runs, burning an hour a week for nothing, is the team’s to drop if nobody outside the team requires it. A routine decision that needs sign-off from three levels up is not the team’s to make, and the work behind it waits.
When you escalate, bring the specific constraint, its cost in the team’s own evidence, and the decision you need. Each escalation line ends with a linked name for the condition above the team. The name tells people who use the framework which larger mechanism you mean.
The 20 gaps carry no severity order, because which one bites hardest depends on the team.
The 20 gaps
Alignment Drift
“We don’t know why we’re building this, but it’s due Friday.” Leadership has a vision; the team receives a list of features.
1. Feature Factory
You ship features, and nobody tracks whether anyone uses them. The team works hard, the maintenance load grows, and customers notice nothing.
Within your control: Before work starts, write down the outcome each item should produce. After release, check whether the outcome happened, even informally.
Requires escalation: Lasting change needs outcome measures at the portfolio level. Show leaders the features that shipped and still have no evidence they mattered. The enterprise names are Delivery-as-Success and Motion-as-Progress.
2. Requirements Roulette
You build it. They reject it. You rebuild it. Rework eats capacity, estimates become guesses, and the team gets called unreliable for a problem that started upstream.
Within your control: Ask “How will we know this is done?” before accepting an item, and write down the criteria and your assumptions.
Requires escalation: Propose a standing rule: nothing enters the iteration without testable acceptance criteria. If nobody can make the criteria testable, the enterprise name is Strategy Vapor.
3. Whiplash Priorities
Monday’s priority is Tuesday’s afterthought. Everything is Priority #1. Each pivot pays a switching cost that shows up only as “the team seems slow.”
Within your control: Ask “What should we drop?” when work arrives mid-iteration, record the interruptions, and limit what the team itself starts.
Requires escalation: Ask the leaders above the team to own the “no”: to set up one channel for requests, and to require a trade-off before any mid-iteration addition. The enterprise name is Priority Collision.
4. Proxy Trap
The team builds for customers it has never met, from requirements written by people who haven’t met them either. Until after delivery, the team cannot tell a real need from an assumed one.
Within your control: Use the product yourself, and ask to observe support calls and to read unfiltered customer feedback.
Requires escalation: Propose standing access: regular observation sessions, the support ticket stream and a seat in customer interviews. The enterprise name is Customer Distance.
5. Decision Owner Absence
The person who can say yes isn’t in the room, the building or even the same time zone. So the work waits, or the team guesses and redoes it when the decision-maker disagrees.
Within your control: At planning, flag the decisions you will need. Write down every assumption you proceed on.
Requires escalation: Ask the decision-maker for delegation boundaries by scope, risk and reversibility, and show how long work sat waiting. Above the team, this is Decision-Rights Fog when nobody is sure who can say yes, and Governance Drag when the approval costs more than the risk it manages.
6. Framework Theater
The ceremonies happen. The hard decisions don’t. Improvement stalls, and when results disappoint, the framework gets the blame.
Within your control: Each ceremony that the team runs should end in a decision. Start with planning: end it with an iteration goal the team can act on.
Requires escalation: Propose that ceremonies be judged by the decisions they produce, and bring the hours they cost. The enterprise name is Motion-as-Progress.
Choked Flow
“It takes longer to move the ticket than to do the work.” The work itself takes hours. Getting it through the system takes weeks.
7. WIP Explosion
Everyone carries six or seven items. All are started, and none is close to done. Each extra item adds waiting and rework, so output does not rise in proportion.
Within your control: Limit how many items the team itself starts, and when a request arrives at the limit, ask what should stop to make room.
Requires escalation: Ask leaders for one enforceable intake rule: at the limit, a named decision-maker defers the new request or stops an item in flight. Above the team, this is Capacity Illusion when leaders authorize work as if the team had spare capacity. It is Change Saturation when the organization runs more initiatives than it can absorb.
8. Dependency Hellscape
Your team can’t finish until Team B finishes. Team B is waiting on Team C. The finish date stops being yours, because you can estimate your own effort but not when a dependency will clear.
Within your control: Map every external dependency at planning. Request each prerequisite early, and track it on the board.
Requires escalation: Escalate the specific prerequisite. Name its owner, the decision you need and the cost of waiting. A recurring dependency points to Silo Survival when functional ownership outranks end-to-end flow, and to Missing Runway when teams lack a shared platform capability.
9. Handoff Collision
Code waits for review. The review waits for approval. A transfer can add a wait and lose context, so work that finally clears may need rework.
Within your control: Make each piece review-ready, with its context and edge cases written down, and shrink the batch.
Requires escalation: Ask that the receiving role help redesign the transition. The options are to embed review, to combine stages, or to agree acceptance criteria before work begins. Above the team, this is Governance Drag when the approval costs more than the risk it manages, and Decision-Rights Fog when nobody knows who may accept the work.
10. Hero Bottleneck
Only one person knows the system, and that person is on vacation or drowning. Throughput caps at one person’s capacity, however much time everyone else has.
Within your control: Using time the team already controls, have teammates take turns pairing with the hero, and write down the reasons behind the hero’s decisions.
Requires escalation: Leadership has to trade current delivery for protected time to pair and to hand on the hero’s knowledge. The enterprise name is Silo Survival.
11. Technical Bankruptcy
You’re building on a swamp, and the plan still assumes solid ground. Routine changes take longer than they used to, with the same team.
Within your control: Name new debt when you take it on, and track where existing debt slows changes.
Requires escalation: Ask leaders to choose openly between feature demand and the smallest credible repair, instead of covering both with overtime nobody sees. Above the team, this is Hidden Debt when the debt is invisible in investment decisions. It is Missing Runway when the groundwork that future delivery needs keeps being refused.
12. Stage Mismatch
What works in dev fails in staging. What passes staging breaks in production. So every release needs more checking.
Within your control: List the known differences between environments, and test against production-like conditions as early as you can.
Requires escalation: Ask for a named owner for matching the environments, and for a decision on which differences are acceptable. When the mismatch keeps recurring, the enterprise name is Missing Runway.
13. Unplanned Work Dominance
The plan survives until Tuesday. Firefighting consumes the rest. Unlike Whiplash Priorities, the priorities hold; the capacity to work on them is what goes.
Within your control: Where you control the team’s planning, reserve capacity for interruptions, sizing the reserve from the team’s own history. Sort the interruptions by source.
Requires escalation: Escalate by source. Incidents may need stability investment, and customer escalations may need a change in service design. The enterprise name depends on the source too: it may be Hidden Debt or Change Saturation.
14. Meeting Sprawl
Back-to-back meetings fill the day. No one has time to do the work they’re meeting about. Decisions that need a meeting wait for a free slot.
Within your control: Count the team’s meeting hours, then stop or shorten the meetings the team runs that don’t earn their time.
Requires escalation: Escalate the whole set of recurring meetings, not one invitation, and ask the accountable leaders which to remove, combine or delegate. The enterprise name is Decision-Rights Fog when routine decisions lack a clear owner, and Governance Drag when meetings work as approval gates that cost more than the risk they manage.
Broken Feedback
“We keep talking, but nothing changes.” Work ships and the ticket closes. Nobody finds out whether the work mattered.
15. Zombie Retrospectives
Every two weeks, the team gathers to complain about the same things. Unlike Framework Theater, the retro does surface problems, but nothing changes. So workarounds become permanent.
Within your control: Make each retro produce one change the team owns, finish it before the next retro, and report the result.
Requires escalation: Items beyond the team’s authority need a path upward, and someone above the team who accepts ownership of the response. Without that owner, the enterprise name is Leadership Absence.
16. Vanity Velocity
The velocity chart climbs. The customer sees nothing new. Once velocity is used to judge the team, estimates become a negotiation, and planning gets worse.
Within your control: Use velocity only for the team’s own planning, and report it beside outcome evidence, such as adoption.
Requires escalation: Go to whoever decides how delivery performance is reported. Ask them to add outcome evidence to the report, and bring one case where velocity climbed and the outcome did not. The enterprise name is Data Illiteracy.
17. Inspection Bottleneck
The work took three days to build. It has sat “ready for review” for three weeks. Handoff Collision costs time and context at any transfer. Here, what is lost is the feedback itself. It comes back too late, and in batches too large to use. One review queue can earn both marks.
Within your control: Make each review cover less work, and answer the reviewer’s likely questions in advance.
Requires escalation: Ask the review-policy owner, not the reviewers, to match review effort to each change’s risk. Above the team, this is Governance Drag.
18. Release Trauma
War rooms. Pizza. The 3 AM call. Releasing hurts, so the team does it as rarely as it can. The release cadence caps how fast the team learns what only production use can show.
Within your control: Break changes into smaller pieces, and record what each release costs.
Requires escalation: Ask whoever allocates engineering investment to fund a first step toward automated testing, decoupled deployment or reliable rollback. When that investment keeps being deferred, the enterprise name is Hidden Debt.
19. Experiment Prohibition
“We should test that assumption.” “We don’t have time.” Even a small test has to clear approvals built for permanent changes, so the assumption ships untested. If it is wrong, you learn late, when fixing costs most. The enterprise gap of the same name is different: teams skip security review or drown in it.
Within your control: Describe each proposed test as what it is: a way to reduce investment risk.
Requires escalation: Ask the owner of the change-control policy to decide which changes need the full approval. Bring one worked case as evidence: the assumption that shipped untested, what it cost, and the test that would have checked it. The enterprise name is Evidence Stalemate when proof is demanded before the test. It is Governance Drag when approvals built for permanent changes apply in full to small, limited tests.
20. Local Optimization
Your team hits every target. The system gets worse. Unlike Vanity Velocity, the gain is real; the problem is its scope. Gains at one stage move cost to the next.
Within your control: Learn what the teams before and after yours need, and follow your work past your boundary.
Requires escalation: Ask whoever owns how delivery is measured across teams for one end-to-end measure, and bring a case where every team hit its target and the customer still waited. Above the team, the name to carry up is Altitude Error: the improvement was real, and the constraint sits where the team can’t reach.
What your marks tell you
Where they cluster. If most of your marks fall in one barrier, start the diagnosis there. If they spread across all three, the barriers may be feeding each other, so a fix in one can fade. The Vicious Cycle shows how.
One stalled item. Take one item that stalled in the last iteration. Follow it through three stops: where it came from, where it waited, and what feedback or evidence came back after it shipped. If the item has not shipped, nothing has come back yet. At each stop, ask which marked gap was at work.
Where each constraint sits. For each marked gap the item ran into, find where its constraint sits. A fix that stalls does not prove the constraint is above you. Check instead what decision or resource the fix needs, who controls it, and whether the team holds that authority. Misplace the constraint, and the team grinds against one it cannot move, or waits for permission it never needed. That mistake is Altitude Error.
What comes first. Pick the marked gap that costs the team most, judged by its own evidence, such as time lost to waiting or work that had to be redone. Then do the team’s part, and escalate the rest as its entry describes.
The 20 gaps are our framework for organizing what you observe, not a scored assessment, so your marks show where to look without grading the team.
Where to go next
To locate the constraint that binds your team’s flow, read The Constraint Horizon. Each gap will get its own page on this site. Applied End-to-End Flow: Team covers all 20 in full. It also covers the Team-Level Value Acceleration Process and the four team amplifiers (Trust Erosion, Accountability Avoidance, Micromanagement Creep and Burnout Accumulation), which make the gaps harder to fix.
Sources and lineage
The 20 team gaps and the Three Barriers form part of the Applied End-to-End Flow framework developed by Curtis Hibbs and Joshua Barnes. The complete treatment appears in Applied End-to-End Flow: Team. These are authored framework models, not externally validated diagnostic scores.