Agile Team Anti-Patterns with Real Examples: What Broken Processes Look Like

Agile Team Anti-Patterns with Real Examples

Anti-patterns sound abstract. Let's see them in the wild—what broken agile actually looks like on real teams.

---

Example 1: The Silent Standups (Theater, No Substance)

The Scenario:

Team of 8 developers, one manager (Dave), daily standups at 10:00 AM.

Monday 10:00 AM:
Dave: "Alright team, status updates. Go around."

Alice (Backend): "I worked on the API endpoint. I'm working on 
error handling. I'll work on tests tomorrow."

Bob (Frontend): "I fixed the login bug. Working on the dialog 
component. Will keep working on it."

Sarah (DevOps): "Infrastructure updates. Same as yesterday, really."

[Dave is typing in Slack, not looking up]

6 minutes pass. Everyone is bored.

Charlie (Junior Dev): "Um... I'm blocked on the database migration? 
The table schema doesn't match what Alice said yesterday..."

Dave: "We don't have time. Bring it up after standup."

Charlie: [Stays blocked for 3 more days]

10:06 AM standup ends. Nothing is solved.

What's broken:

1. No time for problem solving — standup is a status report
2. Authority not present — Dave is typing, not facilitating
3. Blocker not handled — Charlie's issue buried
4. Theater not therapy — people avoid raising real problems
5. Psychological safety broken — junior dev raised hand, got shut down

What happened the next sprint:

• Charlie's blocked task delayed a story by 4 days
• Bob finished login and realized he needed Alice's API work (not done)
• Sarah discovered infrastructure issue mid-sprint
• Velocity dropped 30% (everyone waiting on each other)

---

Example 2: The Ghost Retrospective (Process Theater)

The Scenario:

Friday 4:45 PM, sprint ends. Retro scheduled for 5:00 PM.

5:00 PM Retro starts (15 people)

Scrum Master (Tom): "What went well this sprint?"

[Long silence. Dave, the manager, is also attending.]

Alice: "Um... we shipped the feature?"

Tom: "Good. What could be better?"

Bob: "Communication could be better."

[Dave leans forward, makes a note—Bob looks worried]

Charlie: [Says nothing. Looks at his shoes]

Sarah: "I think our process is fine."

[More silence]

Tom: "Okay, great. Everything's good then?"

Alice, Bob, Sarah: "Yeah, everything's fine."

Tom: "See you all next sprint!"

5:12 PM — Retro ends. Nothing discussed. No action items.

What happened after:

Monday 10:00 AM standups still fail. No one said anything because:

1. Manager in the room — team doesn't trust they can be honest
2. No psychological safety — Bob's "communication" comment made him nervous
3. No action items — so nothing changes
4. "Fine" is not feedback — it means people are hiding

The real problems (nobody voiced):

• Standups are pointless (wasted 5 hrs last sprint)
• Alice is overwhelmed (too many meetings)
• Bob's code reviews take 2+ days (blocking others)
• Charlie is struggling on his first sprint (needs more support)
• Sarah has a tech debt idea nobody wants to hear

All these stayed buried. Sprint 2 repeated the same problems.

---

Example 3: The Scope Creep Sprint

The Scenario:

Sprint starts Monday with 55 points committed.

MONDAY 10:00 AM Sprint Kickoff:
Scrum Master: "We're committed to these 5 stories, 55 points total.
sprint goal is 'Improve checkout flow.' Any questions?"

Team: "Looks good."

WEDNESDAY 2:00 PM:
Dave (Manager): "Hey team, I just got an email from a VIP customer.
They want feature X. Can we squeeze it in?"

Alice: "We're at capacity... but I guess we could—"

Dave: "Great. I'll add it to Jira."

[Dave adds 13-point story without asking team to remove anything]

THURSDAY 10:00 AM Standup:
Tom: "Uh, Dave... we committed to 55 points. We're now at 68 points.
That's physically impossible."

Dave: "Just work faster. Crunch time."

THURSDAY 5:00 PM:
- Team in panic mode
- Code review skipped (no time)
- Testing skipped (no time)
- Alice and Bob working overtime
- Charlie goes home stressed

FRIDAY 3:00 PM Sprint ends:
- 54 of 68 points done (23 points added mid-sprint)
- 4 bugs discovered in checkout flow (code review was skipped)
- Two stories are incomplete
- Velocity is meaningless (55 committed, 68 added, 54 done)

What's broken:

1. No sprint boundary — mid-sprint additions without removal
2. Impossible commitments — math doesn't work
3. Quality sacrificed for quantity — skipped code review = bugs
4. Team burnout — overtime and stress
5. Velocity is now useless — can't predict anything

Long term consequences:

• Burnout cycle continues (team exhausted, makes mistakes, more bugs, more pressure)
• Next sprint: senior dev quits (fed up with crunch)
• Onboarding new person takes time they don't have
• Velocity never stabilizes

---

Example 4: The Blocked Sprint (No Cross-Training)

The Scenario:

Team of 5 has sprint story: "Set up payment system with Stripe"

MONDAY:
- Alice (Expert Payment Dev): "I'll build the Stripe integration"
- Assignment: Only Alice can work on this (she knows Stripe)

WEDNESDAY:
- Alice is sick
- Bob is assigned to "continue payment story"
- Bob has never used Stripe API
- Bob spends 6+ hours trying to understand Alice's code
- Bob asks questions—no one knows the answers

THURSDAY:
- Alice still sick
- Bob is completely blocked
- Story cannot progress
- Rest of team cannot start dependent stories (waiting on payments)
- Burndown chart freezes

FRIDAY:
- Alice returns
- Realizes Bob's work was 80% wrong (different approach)
- Alice reworks the code herself
- Payment feature finishes at 4:30 PM Friday (crunch)

What's broken:

1. Knowledge silos — only one person knows payment system
2. No pairing — Alice never shared knowledge
3. No documentation — payment setup not written down
4. Blocking dependencies — entire sprint depends on one person
5. Bus factor = 1 — if Alice leaves, payment feature is gone

Long term consequences:

• When Alice gets promoted, entire payment system is a black box
• New dev hired—takes 3+ weeks to learn payments (should be 1 week with docs)
• Vacation time expensive (team is blocked whenever Alice is gone)
• Alice burned out (everyone depends on her)

---

Example 5: The Micromanaged Sprint (No Autonomy)

The Scenario:

Team of 4, manager (Dave), PO (Sarah). Sprint 1.

SPRINT PLANNING (90 mins):
Dave: "Here are the stories I want you to build."
[Dave has already assigned which dev builds which story]

Dev 1 (Alice): "Actually, I think we should—"
Dave: "I've thought about this. Do it my way."

[Alice gets a story: "Build profile page"]

WEDNESDAY STANDUP:
Alice: "I'm building the profile page, but I think we should use 
Apollo GraphQL instead of REST to reduce API calls."

Dave: "We're using REST. I decided this."

Alice: "But GraphQL would be 40% faster and—"

Dave: "I don't want to hear it. Stick with REST."

FRIDAY:
[Profile page is slow, needs REST calls to fetch user + preferences + avatar]

Alice: "Yeah, this is exactly why GraphQL would be better."

Dave: "Then make REST calls in parallel."

[Alice adds complexity to avoid GraphQL]

SPRINT REVIEW:
Dave: "See, my way worked fine."

[But the page is slow, code is messy, and Alice is frustrated]

What's broken:

1. No autonomy — Dave makes all technical decisions
2. No trust — doesn't listen to team's expertise
3. Poor decisions — Dave is a manager, not a principal engineer
4. Low morale — Alice is shut down
5. Worst ideas win — because they're the manager's ideas

Long term consequences:

• Sprint 2: Alice is quiet in meetings (learned not to speak up)
• Technical debt accumulates (best practices ignored for Dave's ideas)
• Team is 40% less productive (smart people not trusted)
• High performers leave (burnout, lack of autonomy)
• Code quality drops (no one cares anymore)

---

Example 6: The Unfinished Sprint (No Definition of Done)

The Scenario:

Sprint goal: "Improve user authentication." 8 stories, 55 points.

MONDAY Sprint Starts:
- Alice: "Building the login redesign (8 pts)"
- Bob: "Building password reset flow (5 pts)"
- Charlie: "Building two-factor auth (13 pts)"

Team definition of "done": "Code written and merged"
[No mention of tests, docs, or deployment]

WEDNESDAY:
Alice: "Login redesign is done" [Merges code to main]
[No tests written. Works on localhost only. Not in staging.]

THURSDAY:
Bob: "Password reset is done" [Merges code]
[Forgot to handle edge case: what if user doesn't have an email?]
[No documentation. QA has no idea how to test it.]

FRIDAY 4:00 PM Sprint Review:
Stakeholder: "Wait, the login redesign isn't live?"

Scrum Master: "It's... done, but not deployed to staging yet."

Stakeholder: "But it's Friday. How is it not live?"

Alice: "We merged it Monday. It should be live by now, but... 
no one deployed it?"

[Discovery: the code has a bug—crashes on Safari, and tests 
would have caught it]

What's broken:

1. "Done" is undefined — code merged ≠ code in production
2. No tests — bugs discovered later
3. No deployment — code never reaches users
4. No docs — QA can't test, support can't explain it
5. Wasted work — 26 points of "done" stories are now incomplete

Long term consequences:

• Each sprint, same problem: code never deployed
• Backlog of "done" but not deployed code stacks up
• QA has no idea what to test
• Stakeholders think team is slow (they delivered 50 points this sprint, but only 20 are actually live)
• Trust erodes

---

Example 7: The Velocity Illusion (Climbing Velocity = Burnout Incoming)

The Scenario:

SPRINT 1: 30 points delivered (team new, finding rhythm)
SPRINT 2: 40 points delivered
SPRINT 3: 50 points delivered
SPRINT 4: 55 points delivered
SPRINT 5: 60 points delivered

Manager sees: "Great! Velocity is growing 10 points per sprint!"
Manager plans: "Q3 will be 80 points per sprint!"
Manager tells C-suite: "We're accelerating!"

[But here's what's actually happening]

SPRINT 3-5 (The crunch):
- Team is working overtime
- Code review is 1 day (should be 2) = bugs slip through
- Testing is reduced
- Alice is sleeping 5 hours/night
- Bob has skipped 3 gym sessions
- Charlie is arguing with his wife because of work stress

SPRINT 6 (The crash):
- Alice gets sick, takes a week off
- Bob is burned out, stops caring, goes through motions
- Charlie quits

VELOCITY DROPS:
- Sprint 6: 20 points delivered (half the team down)
- Sprint 7: 25 points (team is demoralized, new hire onboarding)
- Sprint 8: 30 points (new hire learning)

Manager: "What happened?? Velocity collapsed!"
Team (internally): "We were never sustainable. You pushed us too hard."

What's broken:

1. Mistaking velocity growth for improvement — unsustainable velocity collapses
2. No team feedback — ignored signs of burnout
3. Short-term thinking — cramming instead of sustaining
4. Burnout cycle — team burns out, quits, velocity crashes, new people need training
5. No learning — same cycle repeats

---

Summary: Common Team Dysfunction Patterns

| Dysfunction | Symptoms | Outcome | |------------|----------|---------| | Silent standups | Blockers not raised, status-report theater | Blockers snowball, no problem-solving | | Ghost retros | No action items, "everything fine" | Same problems repeat every sprint | | Scope creep | Mid-sprint adds, impossible commitments | Burnout, low velocity, quality drops | | Knowledge silos | One person blocks whole team | Burnout, high turnover, bottlenecks | | Micromanagement | Managers make technical decisions | Poor code, low morale, top talent leaves | | Undefined "done" | Deployed code is surprise, not plan | Rework, bugs, wasted effort | | Velocity climbing | Team accelerates to burnout | Crash and burnout cycle |

---

How to Diagnose Your Team

Ask these questions honestly:

• Do people raise blockers in standup, or do they hide problems?
• After retro, do people remember the action items?
• Do stories added mid-sprint get balanced (something leaves)?
• Can someone other than the expert complete a story?
• Do developers make technical decisions, or does management?
• Does "done" mean in production, or just "coded"?
• Is velocity stable, or does it yo-yo?

If you answer "no" to most, you have anti-patterns. Pick one, fix it, measure it. Then tackle the next.