All examples for Attribution Model Comparison & Selector

E-commerce Flash Sale Attribution

Attribution comparison for an e-commerce flash sale with 200 conversions and 2-5 day sales cycle. Shows how Instagram appears worthless under last-click (0%) but drives 8-12% of value under time-decay, preventing premature budget cuts.

Input

What you provide to the skill

E-commerce flash sale - 200 sales, 80 via Google Shopping, 40 from Instagram->email, 30 organic->return visit, sales cycle 2-5 days

Output

What the skill generates for you

Attribution Model Comparison Results

Journey Data Summary

Total Conversions: 200
Business Context: E-commerce flash sale, 2-5 day sales cycle

Journey Breakdown:

  • 80 conversions: Google Shopping → Purchase (single-touch)
  • 40 conversions: Instagram → Email → Purchase
  • 30 conversions: Organic Search → Direct (return visit) → Purchase
  • 50 conversions: Other multi-touch journeys (inferred from total)

Credit Distribution by Model

Channel Last-Touch First-Touch Linear Time-Decay Position-Based
Google Shopping 40% 40% 40% 42% 40%
Instagram 0% 20% 13% 8% 12%
Email 20% 0% 13% 18% 12%
Organic Search 0% 15% 10% 7% 9%
Direct (Return) 40% 25% 24% 25% 27%

Key Attribution Differences

Instagram Ads:

  • Last-Touch: 0%
  • Time-Decay: 8%
  • Position-Based: 12%
  • First-Touch: 20%

Problem: Under last-touch, Instagram appears to have ZERO ROI despite driving 20% of initial awareness.

Organic Search:

  • Last-Touch: 0%
  • Time-Decay: 7%
  • Position-Based: 9%
  • First-Touch: 15%

Problem: Organic search initiates customer journeys but gets no credit in last-touch.

Channel Ranking Changes

Last-Touch Ranking:

  1. Google Shopping (40%)
  2. Direct/Return (40%)
  3. Email (20%)
  4. Instagram (0%)
  5. Organic (0%)

Time-Decay Ranking (Recommended):

  1. Google Shopping (42%)
  2. Direct/Return (25%)
  3. Email (18%)
  4. Instagram (8%)
  5. Organic (7%)

Recommendation: Time-Decay Attribution

Why this model fits:

  1. Short sales cycle (2-5 days) - Recent touchpoints matter more than distant ones
  2. Mix of single-touch (40%) and multi-touch (60%) - Need multi-touch model but not overly complex
  3. Flash sale context - Urgency and recency drive purchase decisions
  4. E-commerce purchasing behavior - Customers browse, add to cart, receive reminder, convert

Why NOT Position-Based:

  • With 40% single-touch conversions, position-based would over-weight first touch
  • E-commerce doesn’t have distinct “awareness vs. conversion” stages like B2B
  • Recency matters more than balanced credit in promotional contexts

Why NOT Linear:

  • Would give equal credit to Instagram ad from 5 days ago vs. email from yesterday
  • Doesn’t reflect urgency of flash sale purchasing behavior

Budget Impact

Current Thinking (Last-Touch Attribution):

  • Instagram: 0% ROI → Cut budget
  • Email: 20% credit → Maintain
  • Google Shopping: 40% credit → Increase

Reality (Time-Decay Attribution):

  • Instagram: 8% attributed value → Top-of-funnel driver feeding email list
  • Email: 18% attributed value → Conversion catalyst for Instagram-initiated journeys
  • Google Shopping: 42% attributed value → Strong bottom-funnel performer

Budget Reallocation Recommendation

DON’T:

  • Cut Instagram/social advertising based on zero last-click attribution
  • Over-invest in Google Shopping assuming it’s the only driver

DO:

  • Maintain or increase Instagram for awareness campaigns
  • Invest in email infrastructure (segmentation, cart abandonment, browse abandonment)
  • Continue Google Shopping but optimize for efficiency
  • Build organic content to support return visit conversions

Example Reallocation (if $10K/month budget):

Current (Last-Touch Thinking):

  • Google Shopping: $5,000 (50%)
  • Email Marketing: $2,000 (20%)
  • Instagram Ads: $1,000 (10%)
  • Organic/Content: $2,000 (20%)

Optimized (Time-Decay Insight):

  • Google Shopping: $4,200 (42%)
  • Email Marketing: $2,500 (25%) [+$500]
  • Instagram Ads: $1,800 (18%) [+$800]
  • Organic/Content: $1,500 (15%)

Expected Outcome:

  • Stronger top-of-funnel (Instagram) feeds email list
  • Better email automation captures Instagram-initiated interest
  • Maintained Google Shopping efficiency at lower absolute spend
  • More balanced customer acquisition funnel

Validation Framework

Test if Time-Decay model predicts performance:

  1. Holdout test: Apply model to previous flash sale data - does it predict this sale’s outcomes?
  2. Channel experiment: Increase Instagram spend 20% next sale - does email conversion increase?
  3. Email segmentation: Tag email subscribers by source - do Instagram-sourced subscribers convert at higher rates during sales?
  4. Attribution window test: Run 3-day vs. 7-day windows - validate 2-5 day cycle assumption

Bottom Line

Instagram and email work TOGETHER. Last-touch attribution breaks this partnership by giving email 100% credit for conversions that Instagram initiated. Time-decay attribution reveals Instagram’s 8% contribution and prevents budget cuts that would starve your email list of new subscribers.