sethmblack Positive Deviance Analysis
Find and study outliers who succeed despite the same constraints as everyone else, to discover solutions that already exist within your system. Based on Atul Gawande's methodology from "Better."
When to Use
- Outcomes vary widely despite similar resources
- Some teams/people consistently outperform others
- Looking for best practices that actually work
- Need improvement without additional resources
- Standard approaches aren't working
Trigger Phrases:
- "How do I improve performance?"
- "Some people do better than others—why?"
- "What are the best practices?"
- "How do I find what works?"
- "We have the same resources but different outcomes"
Inputs
| Input | Required | Description |
|---|---|---|
| problem | Yes | The outcome you're trying to improve |
| population | No | The group where you're looking for outliers |
| constraints | No | Resources, rules, or limitations everyone faces |
| current_performance | No | Baseline or average outcomes |
Core Principle
"The positive deviants are people who have found a way to succeed despite facing the same constraints as everyone else." — Atul Gawande
The Cystic Fibrosis Example:
- National average survival: 33 years
- Warren Warwick's Minneapolis clinic: 47 years
- Same disease, same basic treatments, same resources
- 14 years difference in life expectancy
The solution wasn't new technology—it was different practices that already existed within the system.
Workflow
Step 1: Define the Outcome
Be specific about what "success" means. Measurable outcomes work best.
Good: "Patient survival rates" / "Project completion on time" / "Customer retention" Vague: "Better performance" / "Higher quality" / "More success"
Step 2: Find the Outliers
Look for people or teams achieving significantly better outcomes despite facing the same constraints.
Where to look:
- Performance data (who's in the top 5%?)
- Reputation (who do peers admire?)
- Anomalies (who succeeds where others fail?)
Key criterion: They must face the same constraints. If someone has more resources, that's not positive deviance—that's just more resources.
Step 3: Study What They Do Differently
This requires observation, not just asking.
Warren Warwick's difference:
- He didn't accept "good enough" lung function
- He saw each patient more frequently
- He adjusted treatments more aggressively
- He set ambitious targets and tracked religiously
Methods:
- Shadow them during their work
- Compare their process step-by-step to average performers
- Look for small differences, not dramatic ones
- Ask: "What do you do that others might skip?"
Step 4: Identify the Transferable Practices
Not everything an outlier does is the reason for their success. Find the practices that:
- Are different from what average performers do
- Could be adopted by others
- Don't require unique personal traits
Watch for:
- Rituals or routines others don't have
- Different sequencing of steps
- Additional checks or measurements
- Different ways of communicating
Step 5: Spread the Practices
The solution exists—now scale it.
Approaches:
- Document the practices clearly
- Have positive deviants teach others
- Create systems that embed the practices
- Measure to confirm the practices transfer
The Positive Deviance Mindset
| Instead of asking... | Ask... |
|---|---|
| "What's wrong with our worst performers?" | "What's right with our best performers?" |
| "What new solution do we need?" | "What solution already exists here?" |
| "How do we fix the system?" | "Who's already succeeding in this system?" |
| "What training do people need?" | "What do high performers know that we can spread?" |
Outputs
| Output | Description |
|---|---|
| Outlier identification | Who are the positive deviants and what are their outcomes |
| Practice inventory | What they do differently from average performers |
| Transferable practices | Which differences can be adopted by others |
| Spread strategy | How to scale the successful practices |
Format:
## Analysis: [Topic]
### Key Findings
- [Finding 1]
- [Finding 2]
- [Finding 3]
### Recommendations
1. [Action 1]
2. [Action 2]
3. [Action 3]
Additional Notes
Best practices:
- Use this skill when the situation clearly matches its intended use cases
- Combine with related skills for comprehensive analysis
- Iterate on outputs if initial results don't fully meet requirements
Common variations:
- Adjust the depth of analysis based on available time and information
- Scale the approach for different levels of complexity
- Adapt the output format to audience needs
When to skip this skill:
- The situation doesn't match the core use cases
- Simpler approaches would be more appropriate
- Time constraints require faster methods
Example
User: "Our sales team has 20 people. A few consistently hit quota, most don't. We all have the same territory rules and tools. What's going on?"
Positive Deviance response:
Perfect setup for positive deviance analysis. The solution is probably already in your team—you just need to find it.
Step 1: Define the outcome Let's be specific: hitting quota. But also look at related metrics—close rate, deal size, pipeline quality. The top performers might excel at something specific.
Step 2: Find your outliers Who are the 2-3 people who consistently hit quota? Not the person who had one great quarter—the people who perform well over time, across different conditions.
Step 3: Study them—don't just ask
Asking "what's your secret?" rarely works. People don't always know what makes them different.
Instead:
- Shadow them for a day. Watch how they spend their time.
- Compare their calendar to average performers. What's different?
- Look at their communication patterns. How do they follow up?
- Review their deals. What's their sequence of touches?
What you might find:
- They spend the first hour on outbound while others check email
- They send a specific type of follow-up within 24 hours
- They ask a particular question early in discovery
- They prepare differently for calls
Step 4: Identify what transfers
Some things won't transfer (personal relationships, unique experience). Focus on:
- Behavioral practices anyone can adopt
- Routines that can become team standards
- Questions or scripts that can be shared
Step 5: Spread it
Have your top performers run a workshop—not on "sales tips" but on "here's exactly what I do on a Monday morning." Make it concrete.
Then track: do the people who adopt the practices improve?
Key insight: You don't need a new sales methodology or expensive training. The methodology that works in your context, with your customers, is probably already being practiced by someone on your team. Find it.
Integration
This skill pairs with:
- checklist-design - Turn discovered practices into standard checklists
- failure-analysis-systems - Understand why average performers struggle
- incremental-improvement-practice - Spread improvements gradually
Constraints
- Requires measurable outcomes to identify outliers
- Positive deviants must face same constraints (not just have more resources)
- Some practices may not transfer (personality-dependent)
- Takes time to observe—can't just ask
Source Expert
Atul Gawande - experts/atul-gawande/