The only AI-powered attribution API with scientific guard rails. Our 8 verified causal inference methods prove which marketing channels actually cause conversions—not just correlate with them.
Trusted by marketers who need statistical proof, not guesswork.
U.S. companies waste $400 billion annually on ineffective marketing. Why? Because traditional attribution confuses correlation with causation.
"80% of buyers saw your Facebook ad"
Correlation detected. But did the ad cause the purchase? Or did people who were already going to buy just happen to see it? You're making $100K decisions based on guesswork.
Result: Wasted ad spend on channels that don't actually drive sales.
"Facebook ads caused a 15% lift in sales"
Causation proven. Using methods from economics and medicine (Propensity Score Matching, Instrumental Variables, Doubly Robust Estimation), we isolate the actual causal effect of each channel.
Result: Reallocate budget to channels that provably drive revenue.
Real client outcome: 40% reduction in wasted ad spend after switching from correlation-based to causal attribution. See how →
We use the same rigorous methods that economists and medical researchers use to prove causation—now available via simple API.
What it does: Combines two models (propensity and outcome) to estimate causal effects. Even if one model is wrong, the estimate remains valid.
Why it matters: More accurate than any single method. Reduces bias by 60-80%.
What it does: Matches customers who saw an ad with similar customers who didn't. Compares outcomes to find true causal effect.
Why it matters: Mimics randomized controlled trials. Used in economics and medicine for decades.
What it does: Finds "natural experiments" in your data (e.g., email sent due to time zone, not user interest) to isolate causal effects.
Why it matters: Handles unmeasured confounding. Gets causation when nothing else can.
What it does: Game-theoretic method that fairly distributes credit across channels based on marginal contribution.
Why it matters: Provably fair attribution. No channel gets over or under-credited.
What it does: Automatically discovers causal structure in your data. Learns which channels influence which outcomes.
Why it matters: Uncovers hidden relationships. No prior assumptions needed.
What you get: Every result includes p-values (significance), confidence intervals (uncertainty), and standard errors.
Why it matters: Know exactly how confident you can be. Make decisions with statistical certainty.
We combine AI/ML models with rigorous statistical validation to deliver the most accurate causal estimates possible. Every AI-generated result passes through scientific guard rails before you see it.
How it works:
All 8 methods verified and operational. See technical documentation →
Average response time: 4-32ms. Fast enough for real-time decisioning.
One API call with customer touchpoints (email, ads, social, etc.). Works with any marketing stack.
Our algorithms apply 8 proven causal methods to determine which channels actually caused conversions (not just correlated).
See exactly which channels drive revenue, with p-values, confidence intervals, and dollar amounts attributed to each channel.
curl https://api.causalmma.com/api/v1/attribution \
-H "X-API-Key: YOUR_KEY" \
-H "Content-Type: application/json" \
-d '{
"touchpoints": [
{"channel": "email", "timestamp": "2025-01-01T10:00:00", ...},
{"channel": "facebook", "timestamp": "2025-01-02T14:00:00", ...}
],
"attribution_model": "data_driven"
}'{
"attribution_weights": {
"email": 0.45, // Email caused 45% of conversions
"facebook": 0.55 // Facebook caused 55%
},
"confidence_intervals": {
"email": {"lower": 0.38, "upper": 0.52}, // 95% confidence
"facebook": {"lower": 0.48, "upper": 0.62}
},
"p_values": {
"email": 0.001, // p < 0.05 = statistically significant
"facebook": 0.0001
},
"method_used": "doubly_robust" // Which causal method was applied
}We're not another correlation tracker. We're the only attribution API built on peer-reviewed causal inference from economics and medicine.
Them: Track last-click, first-click, or time-decay attribution (all correlation-based)
Us: 8 verified causal inference methods (Doubly Robust, Propensity Score Matching, Instrumental Variables, Shapley, PC Algorithm). We prove what causes conversions.
Them: Show you numbers without confidence intervals or significance testing
Us: Every result includes p-values, 95% confidence intervals, and standard errors. Know exactly how reliable your insights are.
Them: Batch processing with 24-hour delays. Not suitable for real-time optimization
Us: 4-32ms average response time. Real-time API for instant insights and automated optimization.
Them: Proprietary "black box" algorithms with no scientific validation
Us: Methods from peer-reviewed economics and medical research. Transparent, reproducible, and scientifically rigorous.
Them: Enterprise platforms charge $20K-$100K+ per year (Windsor.ai, Northbeam, Rockerbox)
Us: Start at $149/month with 14-day free trial. Same science, fraction of the cost. See pricing →
Them: Complex integrations requiring weeks of engineering work
Us: Simple REST API with code examples in 6+ languages. Integrate in under 10 minutes. View examples →
If you're making $100K+ decisions based on which marketing channels to fund, you need statistical proof of causation, not correlation. We're the only platform that provides it.
Start with our cost-effective cloud API, or deploy our SDK locally if you have your own AI infrastructure.
Self-hosted infrastructure:
Our Cloud API: $149-$799/month
Save 95% on infrastructure costs!
Already have GPUs and AI infrastructure? Deploy our SDK locally for maximum privacy and control.
1. Install locally: pip install causalmma-client
2. Algorithms run on your machine: Data stays 100% local
3. Control plane validates license: Metadata only (no data sent)
4. Results computed locally: 30x faster for large datasets
Choose Cloud API if:
Choose Local SDK if:
Start free. Scale as you grow. No setup fees, cancel anytime.
$149/month
$399/month
$799/month
Custom
All plans include: Full API access • Complete documentation • Code examples • Statistical confidence intervals • Multiple attribution models
Any business making 6-7 figure decisions on marketing budget needs causal proof, not correlation guesswork.
Problem: Spending $100K/month on Facebook ads, but is it actually driving sales or just reaching people who would buy anyway?
Solution: Propensity Score Matching proves Facebook's causal impact. Reallocate budget from channels that don't drive sales.
Result: 30-40% reduction in wasted spend. More revenue per ad dollar.
Problem: Prospects interact with 10+ touchpoints over 6 months. Which channels actually cause them to convert?
Solution: Shapley Values fairly attribute credit across the entire customer journey. Instrumental Variables handle unmeasured factors.
Result: Optimize multi-touch campaigns with statistical confidence.
Problem: Clients demand ROI proof. "80% saw the ad" isn't proof the ad worked.
Solution: Show clients p-values, confidence intervals, and causal lift percentages. "Your LinkedIn campaign caused a statistically significant 18% increase in conversions (p < 0.001)."
Result: Win bigger clients. Justify higher fees with scientific rigor.
Problem: Your platform tracks attribution, but users complain results don't match reality.
Solution: Integrate our API for true causal attribution. Differentiate from competitors still using correlation-based methods.
Result: "Powered by causal inference" becomes your competitive advantage. API docs →
Common thread: If you're spending $50K+/year on marketing and need to prove ROI with statistical certainty, causal inference is the answer. Get started →
Get statistical proof of which channels drive conversions. All plans include a 14-day free trial—no credit card required.
📧 Email: durai@infinidatum.net
⏱️ Response Time: Usually within 24 hours
🎁 Free Trial: Test Starter, Professional, or Business plan for 14 days
📚 Learn More: Quick Start Guide | API Documentation | Code Examples