Cross-Device Matching
“For each anonymous device session, which known user does it belong to?”
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
For each anonymous device session, which known user does it belong to?
With third-party cookie deprecation, 40-60% of user sessions are anonymous. Cross-device identity relies on probabilistic matching using behavioral and relational signals. Kumo learns from the graph of sessions, devices, IPs, and browsing patterns to resolve anonymous sessions to known users — without cookies. Every unresolved session is a lost personalization opportunity, and ad-tech platforms estimate $10-30B in annual industry-wide waste from fragmented user identities.
How KumoRFM solves this
Relational intelligence for identity resolution
Kumo builds a relational graph connecting sessions to devices, IP hashes, browsing patterns, and known user profiles. Instead of relying on deterministic cookie matching, Kumo learns that Anonymous Session S-901 browses the same product categories, visits from an IP range associated with User U-442, and exhibits the same time-of-day patterns as U-442's known sessions on other devices. The link prediction model resolves anonymous sessions to known users through these structural signals — maintaining identity resolution in a cookieless world.
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
ANONYMOUS_SESSIONS
| session_id | device_fingerprint | ip_hash | pages_viewed | timestamp |
|---|---|---|---|---|
| S-901 | FP-X7K2 | IP-44A1 | 12 | 2025-09-14 20:15 |
| S-902 | FP-M3P9 | IP-88B2 | 8 | 2025-09-15 08:30 |
| S-903 | FP-X7K2 | IP-44A1 | 5 | 2025-09-15 21:00 |
SESSIONS
| session_id | user_id | device_fingerprint | ip_hash | timestamp |
|---|---|---|---|---|
| S-501 | U-442 | FP-R1N8 | IP-44A1 | 2025-09-13 09:00 |
| S-502 | U-442 | FP-R1N8 | IP-44A1 | 2025-09-14 12:30 |
| S-503 | U-781 | FP-T5W3 | IP-88B2 | 2025-09-14 10:15 |
USERS
| user_id | signup_date | primary_device | |
|---|---|---|---|
| U-442 | sarah@email.com | 2024-03-12 | FP-R1N8 |
| U-781 | david@corp.com | 2024-07-20 | FP-T5W3 |
| U-195 | lisa@mail.com | 2025-01-05 | FP-K8J4 |
Write your PQL query
Describe what to predict in 2–3 lines — Kumo handles the rest
PREDICT LIST_DISTINCT(SESSIONS.USER_ID, 0, 7, days) FOR EACH ANONYMOUS_SESSIONS.SESSION_ID
Prediction output
Every entity gets a score, updated continuously
| SESSION_ID | MATCHED_USER_ID | SCORE | TIMESTAMP |
|---|---|---|---|
| S-901 | U-442 | 0.91 | 2025-09-21 |
| S-902 | U-781 | 0.86 | 2025-09-21 |
| S-903 | U-442 | 0.88 | 2025-09-21 |
Understand why
Every prediction includes feature attributions — no black boxes
Anonymous Session S-901 (FP-X7K2)
Predicted: 91% match with User U-442 (sarah@email.com)
Top contributing features
IP hash overlap with known sessions
IP-44A1 match
30% attribution
Browsing category similarity
0.89
25% attribution
Time-of-day pattern match
Evening user
20% attribution
Page-view depth similarity
12 vs 14 avg
15% attribution
Session duration pattern
18 min avg
10% 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: Resolve 40-60% of anonymous sessions to known users without third-party cookies — unlocking personalization and recovering millions in ad attribution value.
Related use cases
Explore more entity resolution 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.
