Configuring Thresholds

Thresholds determine when LENS triggers an alert. Setting them correctly balances catching real issues against avoiding alert fatigue.

Understanding Thresholds

A threshold defines the boundary between "normal" and "anomalous":

                    Normal Range
         ◄─────────────────────────────►
    ─────┬───────────────┬───────────────┬─────
         │               │               │
   -30%  │     -20%      │      +20%     │  +30%
         │               │               │
         └── Alert ──────┴───── Alert ───┘
           (drop)                 (spike)

Threshold Types

Percentage Thresholds

Most common type. Alert when metric changes by a percentage:

threshold:
  type: percentage
  value: 25
  direction: decrease

Best for: Conversion rates, traffic, revenue

Absolute Thresholds

Alert when metric crosses a fixed value:

threshold:
  type: absolute
  value: 100
  direction: below  # Alert when metric drops below 100

Best for: Minimum session counts, revenue targets

Standard Deviation Thresholds

Alert when metric deviates from the mean by N standard deviations:

threshold:
  type: standard_deviation
  value: 2  # 2 standard deviations

Best for: Highly variable metrics, statistical analysis

Adjusting Rule Thresholds

Method 1: Quick Adjust

1. Go to LENS → Rules
2. Find the rule
3. Click the threshold value
4. Enter new value
5. Click Save

Method 2: Full Edit

1. Go to LENS → Rules
2. Click the rule name
3. Navigate to Threshold section
4. Modify all threshold settings
5. Click Save Changes

Threshold Recommendations by Metric Type

Traffic Metrics

Metric	Recommended Threshold	Notes
Total sessions	±30%	Higher variance expected
Organic traffic	±25%	Algorithm changes cause swings
Paid traffic	±20%	Should be more stable
Direct traffic	±35%	Highly variable

Conversion Metrics

Metric	Recommended Threshold	Notes
Conversion rate	±20%	Critical metric, tighter threshold
Transaction count	±25%	Volume-dependent
Add-to-cart rate	±25%	Earlier funnel, more variance
Checkout rate	±15%	Late funnel, tighter monitoring

Revenue Metrics

Metric	Recommended Threshold	Notes
Total revenue	±25%	Primary business metric
Average order value	±20%	Stable metric
Revenue per session	±25%	Combined metric

Engagement Metrics

Metric	Recommended Threshold	Notes
Bounce rate	±25%	Varies by source
Session duration	±30%	High natural variance
Pages per session	±25%	Content-dependent

Sensitivity Presets

Use presets for quick configuration:

Low Sensitivity

# Fewer alerts, only major changes
traffic_threshold: 40%
conversion_threshold: 30%
revenue_threshold: 35%
technical_threshold: 20%

Use when:

• New to LENS
• High-variance business
• Limited resources to investigate

Medium Sensitivity (Default)

# Balanced approach
traffic_threshold: 30%
conversion_threshold: 20%
revenue_threshold: 25%
technical_threshold: 10%

Use when:

• Established tracking
• Regular monitoring capacity
• Moderate business variance

High Sensitivity

# More alerts, catches subtle changes
traffic_threshold: 20%
conversion_threshold: 15%
revenue_threshold: 20%
technical_threshold: 5%

Use when:

• Stable, predictable business
• Dedicated analytics team
• Critical metrics requiring tight monitoring

Advanced Threshold Settings

Consecutive Breaches

Require multiple consecutive breaches before alerting:

threshold:
  percentage: 20
  consecutive_periods: 3  # Must breach 3 periods in a row

Benefits:

• Reduces false positives from temporary blips
• Ensures sustained change before alerting

Minimum Data Requirements

Set minimum data volume before rule evaluates:

minimum_requirements:
  sessions: 500
  conversions: 50
  revenue: 1000

Benefits:

• Prevents alerts on statistically insignificant data
• Avoids noise in low-traffic periods

Day-of-Week Adjustments

Account for weekly patterns:

day_of_week_adjustment: true
comparison_day: same_weekday  # Compare Monday to Monday

Benefits:

• Reduces false positives from weekend drops
• More accurate comparison base

Time-of-Day Adjustments

Account for daily patterns:

time_of_day_adjustment: true
comparison_hours: same_hours  # Compare 2-3 PM to 2-3 PM

Benefits:

• Accounts for business hours patterns
• Better for real-time monitoring

Tuning Thresholds Over Time

Review Alert History

1. Go to LENS → Analytics
2. View Alert Distribution
3. Identify:
   • Rules with too many false positives
   • Rules that never fire (thresholds too high)
   • Rules that fire constantly (thresholds too low)

Recommended Tuning Process

1. Week 1-2: Use default thresholds
2. Week 3: Review alert history
3. Week 4: Adjust thresholds based on review
4. Monthly: Ongoing threshold refinement

Threshold Tuning Checklist

Signal	Action
>5 false positives/week	Increase threshold by 5-10%
Alert for known issue was missed	Decrease threshold by 5-10%
Rule never fires	Verify data, then decrease threshold
Every alert is actionable	Thresholds are well-tuned

Per-Segment Thresholds

Apply different thresholds to different segments:

rule: conversion_rate_drop
segments:
  - name: "US Traffic"
    filter: country == "US"
    threshold: 15%  # Tighter for main market

  - name: "Other Markets"
    filter: country != "US"
    threshold: 30%  # Looser for smaller markets

Available on: All plans

Threshold Templates

Save threshold configurations for reuse:

1. Configure thresholds for a rule
2. Click Save as Template
3. Name the template (e.g., "E-commerce Strict")
4. Apply to other rules via Apply Template

Pre-built Templates

Template	Best For	Characteristics
E-commerce Standard	Online retail	Balanced, revenue-focused
E-commerce Strict	High-value retail	Tight revenue thresholds
Lead Gen Standard	B2B, lead gen	Form and conversion focused
Media/Content	Publishers	Traffic and engagement focused
SaaS	Software	Signup and activation focused