Step Effectiveness Tracking: Personalised Intervention Ranking

Not all negotiation steps work equally well for everyone. Whistl tracks the effectiveness of each step in your 8-Step Negotiation, learning which techniques work best for YOU. This data then reorders your intervention sequence—putting your most effective steps first.

Why Step Effectiveness Matters

The 8-Step Negotiation Engine isn't one-size-fits-all:

Individual Differences

• Some respond to breathing: Physiological regulation works
• Some respond to reflection: Cognitive awareness helps
• Some respond to visualization: Future-focused motivation resonates
• Some respond to partner support: Social accountability is key

The Problem with Fixed Order

• Wasted time: Ineffective steps delay what would work
• Frustration: Users disengage from unhelpful interventions
• Missed opportunities: Best steps may come too late
• One-size-fits-none: Average effectiveness ≠ individual effectiveness

How Step Effectiveness Is Measured

Whistl tracks multiple metrics for each step deployment:

Success Criteria

# Step effectiveness calculation
def calculate_step_effectiveness(step_id, user_id):
    # Get all deployments of this step for this user
    deployments = get_step_deployments(step_id, user_id)
    
    successful = 0
    total = len(deployments)
    
    for deployment in deployments:
        # Measure success across multiple dimensions
        urge_did_not_return = check_urge_status(deployment, hours=2)
        user_completed_step = deployment.completed
        user_rated_helpful = deployment.helpfulness_rating >= 4
        
        # Count as successful if urge didn't return
        if urge_did_not_return:
            successful += 1
    
    effectiveness = successful / total if total > 0 else 0
    
    return {
        "effectiveness": effectiveness,
        "sample_size": total,
        "confidence": calculate_confidence(total)
    }

Tracking Metrics

Effectiveness Tracking Dashboard

Users can view their personal step effectiveness data:

Personal Step Rankings

┌─────────────────────────────────────────┐
│  Your Step Effectiveness               │
│                                         │
│  1. Acknowledge     ████████████ 89%   │
│     (156 deployments, 139 successful)  │
│                                         │
│  2. Partner         █████████░░░ 83%   │
│     (42 deployments, 35 successful)    │
│                                         │
│  3. Delay           ████████░░░░ 71%   │
│     (89 deployments, 63 successful)    │
│                                         │
│  4. Reflect         ██████░░░░░░ 67%   │
│     (134 deployments, 90 successful)   │
│                                         │
│  5. Visualize       ██████░░░░░░ 61%   │
│     (98 deployments, 60 successful)    │
│                                         │
│  6. Breathe         █████░░░░░░░ 54%   │
│     (112 deployments, 61 successful)   │
│                                         │
│  7. Alternative     ████░░░░░░░░ 48%   │
│     (87 deployments, 42 successful)    │
│                                         │
│  8. Commit          ███░░░░░░░░░ 42%   │
│     (56 deployments, 24 successful)    │
│                                         │
└─────────────────────────────────────────┘

Adaptive Step Ordering

Based on effectiveness data, Whistl reorders steps for each user:

Reordering Rules

• Effectiveness >80%: Promote to positions 1-3
• Effectiveness 60-80%: Keep in middle positions 4-6
• Effectiveness <60%: Move to later positions 7-8
• Effectiveness <10%: Consider removing from sequence

Example: Sarah's Personalised Order

Sarah responds best to visualization and alternatives:

Example: Marcus's Personalised Order

Marcus responds best to tough reflection and partner support:

Confidence Thresholds

Whistl requires sufficient data before reordering:

Minimum Sample Sizes

• 1-9 deployments: No reordering (insufficient data)
• 10-29 deployments: Minor adjustments only
• 30+ deployments: Full reordering allowed

Confidence Calculation

# Confidence in effectiveness estimate
def calculate_confidence(sample_size, effectiveness):
    # Wilson score interval for binomial proportion
    z = 1.96  # 95% confidence
    
    n = sample_size
    p = effectiveness
    
    denominator = 1 + z**2 / n
    centre = (p + z**2 / (2*n)) / denominator
    margin = z * ((p * (1-p) + z**2 / (4*n)) / n)**0.5 / denominator
    
    confidence = 1 - (margin * 2)  # Width of confidence interval
    
    return max(0, min(1, confidence))

# 10 samples at 70% = 45% confidence (low)
# 50 samples at 70% = 78% confidence (medium)
# 100 samples at 70% = 89% confidence (high)

Step Cooldown and Frequency Rules

Based on effectiveness, steps have different cooldown periods:

Cooldown Rules

Real-World Impact

Personalised step ordering improves outcomes:

Before vs. After Personalisation

User Testimonials

"I noticed Whistl started showing me the breathing thing less and the partner thing more. Makes sense—calling my mate works better for me." — Marcus, 28

"The app knows I hate the commitment step. It barely shows it now. Instead it shows me my Bali photos, which actually motivates me." — Sarah, 34

"It's like Whistl learned me. The interventions feel custom now, not generic." — Jake, 31

Conclusion

Step Effectiveness Tracking transforms the 8-Step Negotiation from a fixed sequence into a living, adaptive system. By learning which steps work best for you personally, Whistl delivers the right intervention at the right time—maximising your chances of success.

Your negotiation sequence is uniquely yours—optimised by data, refined by experience, and focused on what works for YOU.

Related: 8-Step Negotiation Engine | Intervention Type Predictor | A/B Testing Engine