Conversion Attribution
“Which touchpoints drove this conversion?”
Book a demo and get a free trial of the full platform: data science agent, fine-tune capabilities, and forward-deployed engineer support.
By submitting, you accept the Terms and Privacy Policy.
Loved by data scientists, ML engineers & CXOs at
A real-world example
Which touchpoints drove this conversion?
Last-click attribution overvalues bottom-funnel channels and undervalues awareness campaigns, leading to systematic misallocation of ad spend. A brand spending $50M per year on digital ads misallocates 30-40% of budget when relying on last-click, representing $15-20M in wasted or suboptimal spend annually.
How KumoRFM solves this
Graph-powered intelligence for advertising
Kumo connects users, touchpoints, conversions, campaigns, and channels into a unified graph. The GNN learns how different touchpoint sequences contribute to conversion, capturing interaction effects between channels that linear models and even Shapley-based approaches miss. Each conversion gets a per-touchpoint attribution score grounded in the full user journey.
From data to predictions
See the full pipeline in action
Connect your tables, write a PQL query, and get predictions with built-in explainability — all in minutes, not months.
Your data
The relational tables Kumo learns from
USERS
| user_id | segment | acquisition_source | ltv_tier |
|---|---|---|---|
| U201 | Enterprise | Organic | High |
| U202 | SMB | Paid Search | Medium |
| U203 | Enterprise | Referral | High |
TOUCHPOINTS
| touchpoint_id | user_id | channel | campaign_id | timestamp |
|---|---|---|---|---|
| TP401 | U201 | Display | CMP10 | 2025-02-15 09:00 |
| TP402 | U201 | CMP11 | 2025-02-18 14:30 | |
| TP403 | U201 | Paid Search | CMP12 | 2025-02-20 11:00 |
CONVERSIONS
| conversion_id | user_id | value | timestamp |
|---|---|---|---|
| CVR101 | U201 | $12,500 | 2025-02-20 11:15 |
| CVR102 | U203 | $8,200 | 2025-02-22 16:00 |
CAMPAIGNS
| campaign_id | channel | spend | objective |
|---|---|---|---|
| CMP10 | Display | $120K | Awareness |
| CMP11 | $15K | Nurture | |
| CMP12 | Paid Search | $80K | Conversion |
CHANNELS
| channel | avg_cpa | avg_roas | attribution_window |
|---|---|---|---|
| Display | $45 | 3.2x | 30 days |
| $12 | 8.1x | 7 days | |
| Paid Search | $28 | 5.5x | 14 days |
Write your PQL query
Describe what to predict in 2–3 lines — Kumo handles the rest
PREDICT LIST_DISTINCT(TOUCHPOINTS.channel, -30, 0, days) FOR EACH CONVERSIONS.conversion_id
Prediction output
Every entity gets a score, updated continuously
| CONVERSION_ID | USER_ID | VALUE | DISPLAY_ATTR | EMAIL_ATTR | SEARCH_ATTR |
|---|---|---|---|---|---|
| CVR101 | U201 | $12,500 | 0.28 | 0.35 | 0.37 |
| CVR102 | U203 | $8,200 | 0.15 | 0.52 | 0.33 |
Understand why
Every prediction includes feature attributions — no black boxes
Conversion CVR101 -- User U201
Predicted: Multi-touch: Display 28%, Email 35%, Search 37%
Top contributing features
Email open-to-conversion time
2 days
35% attribution
Search keyword intent score
High
25% attribution
Display ad first-touch awareness
5 days prior
20% attribution
Cross-channel journey length
3 touchpoints
12% attribution
Similar user conversion paths
68% match
8% attribution
Feature attributions are computed automatically for every prediction. No separate tooling required. Learn more about Kumo explainability
PQL Documentation
Learn the Predictive Query Language — SQL-like syntax for defining any prediction task in 2–3 lines.
Python SDK
Integrate Kumo predictions into your pipelines. Train, evaluate, and deploy models programmatically.
Explainability Docs
Understand feature attributions, model evaluation metrics, and how to build trust with stakeholders.
Bottom line: A brand spending $50M on digital ads recovers $15-20M in misallocated budget by replacing last-click with Kumo's graph-based multi-touch attribution. Every channel gets credit proportional to its true causal contribution.
Related use cases
Explore more ad tech use cases
Topics covered
One Platform. One Model. Predict Instantly.
KumoRFM
Relational Foundation Model
Turn structured relational data into predictions in seconds. KumoRFM delivers zero-shot predictions that rival months of traditional data science. No training, feature engineering, or infrastructure required. Just connect your data and start predicting.
For critical use cases, fine-tune KumoRFM on your data using the Kumo platform and Data Science Agent for 30%+ higher accuracy than traditional models.
Book a demo and get a free trial of the full platform: data science agent, fine-tune capabilities, and forward-deployed engineer support.
