Attribution Modeling for Growth Teams
Who gets credit for the conversion? Understanding attribution models helps you allocate budget and understand what's actually working.
When a user converts, which touchpoint gets credit? The first ad they saw? The blog post they read? The retargeting ad they clicked? Attribution modeling answers this question — and the answer shapes how you allocate your growth budget.
Why Attribution Matters
If you don't know what's working, you can't double down on it. Wrong attribution leads to:
- Over-investing in low-value channels
- Under-investing in channels that actually work
- Inability to calculate true CAC by channel
- Bad budget allocation decisions
The Main Attribution Models
Last Click
100% credit to the last touchpoint before conversion.
Pros: Simple. Easy to implement. Cons: Ignores everything that came before. Over-credits bottom-funnel.
First Click
100% credit to the first touchpoint that acquired the user.
Pros: Values awareness and top-of-funnel. Cons: Ignores nurturing and conversion touchpoints.
Linear
Equal credit across all touchpoints.
Pros: Recognizes full journey. Cons: Doesn't reflect varying impact of touchpoints.
Time Decay
More credit to touchpoints closer to conversion.
Pros: Balances journey while emphasizing recency. Cons: May still undervalue awareness.
Position-Based (U-Shaped)
40% to first touch, 40% to last touch, 20% distributed among middle.
Pros: Recognizes importance of acquisition and conversion. Cons: Arbitrary weights.
Data-Driven (Algorithmic)
Machine learning determines credit based on actual conversion patterns.
Pros: Reflects true impact. Cons: Requires significant data. Black box.
Choosing the Right Model
There's no universally correct model. Choose based on your business:
| Business Type | Recommended Model |
|---|---|
| Long sales cycle | Position-based or time decay |
| Short sales cycle | Last click is often fine |
| Content-heavy | First click or position-based |
| Paid-heavy | Data-driven or linear |
| Early stage | Start simple, add complexity |
Multi-Touch Attribution in Practice
Setting It Up
- Track all touchpoints: Every ad click, page view, email open, demo request
- Stitch identity: Connect anonymous to known users
- Define conversions: What counts as a conversion?
- Choose time windows: How far back do you look?
- Select model: Start with position-based, evolve to data-driven
Tools for Attribution
- Google Analytics 4: Built-in models, limited customization
- Segment + Warehouse: Build your own in SQL
- Attribution tools: Rockerbox, Triple Whale, Northbeam
- Marketing platforms: HubSpot, Marketo have built-in attribution
The iOS 14+ Problem
Post-ATT, mobile attribution is harder:
- Less deterministic matching
- More reliance on modeled conversions
- SKAdNetwork limitations
Adapt by:
- Investing in first-party data
- Using probabilistic attribution
- Triangulating with incrementality tests
Incrementality: The Gold Standard
Attribution shows correlation. Incrementality shows causation.
Run holdout tests:
- Split audience randomly
- Show ads to one group, not the other
- Measure conversion lift
Incremental lift = (Treatment conversion rate - Control conversion rate) / Control rate
True CAC = Spend / Incremental conversions
Common Attribution Mistakes
Trusting one model blindly: Look at multiple models and understand the story.
Ignoring offline touchpoints: Podcasts, events, word-of-mouth aren't tracked but matter.
Too short lookback window: B2B journeys can be months long.
Not connecting to revenue: Conversion attribution without revenue attribution is incomplete.
Platform attribution bias: Every platform wants credit. Cross-reference with independent measurement.
Attribution will never be perfect. But having a model — understanding its assumptions, and knowing where it might be wrong — is infinitely better than no model. Use attribution to guide decisions, but validate with incrementality tests before making big bets.
Related
Filed under Acquisition. See also Designing Referral Programs That Actually Work, Landing Page Anatomy: 11 Elements That Convert, SEO for SaaS: A Growth Team's Guide.