Channel Selection
“For each customer, which outreach channel will drive the highest response rate?”
Book a demo and get a free trial of the full platform: research 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 customer, which outreach channel will drive the highest response rate?
Marketing teams default to a single preferred channel or rotate blindly across email, push, SMS, and in-app. This wastes budget on channels customers ignore and fatigues them on channels they actually use — driving unsubscribes and opt-outs that permanently shrink your addressable audience.
How KumoRFM solves this
Relational intelligence for optimal actions
Kumo models the full customer-channel interaction graph: which channels each customer responds to, when, and in what context. The multi-label prediction scores every channel per customer simultaneously, enabling true per-individual channel routing. The result is fewer unsubscribes, higher response rates, and lower cost per engagement.
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
CUSTOMERS
| customer_id | name | segment | preferred_channel |
|---|---|---|---|
| C-2001 | Alice Nguyen | high-value | |
| C-2002 | Bob Patel | growth | push |
| C-2003 | Clara Diaz | new | sms |
| C-2004 | Dan Kim | enterprise | |
| C-2005 | Eva Chen | high-value | in-app |
OUTREACH
| outreach_id | customer_id | channel | campaign | timestamp |
|---|---|---|---|---|
| OUT-301 | C-2001 | Spring Promo | 2026-02-10 | |
| OUT-302 | C-2001 | push | Flash Sale | 2026-02-12 |
| OUT-303 | C-2002 | push | Spring Promo | 2026-02-10 |
| OUT-304 | C-2003 | sms | Welcome | 2026-02-15 |
| OUT-305 | C-2004 | Enterprise Offer | 2026-02-18 |
RESPONSES
| response_id | customer_id | channel | action | timestamp |
|---|---|---|---|---|
| RSP-401 | C-2001 | clicked | 2026-02-10 | |
| RSP-402 | C-2001 | push | dismissed | 2026-02-12 |
| RSP-403 | C-2002 | push | opened | 2026-02-10 |
| RSP-404 | C-2003 | sms | replied | 2026-02-15 |
| RSP-405 | C-2004 | clicked | 2026-02-18 |
Write your PQL query
Describe what to predict in 2–3 lines — Kumo handles the rest
PREDICT LIST_DISTINCT(RESPONSES.CHANNEL, 0, 7, days) FOR EACH CUSTOMERS.CUSTOMER_ID
Prediction output
Every entity gets a score, updated continuously
| CUSTOMER_ID | CLASS | SCORE | TIMESTAMP |
|---|---|---|---|
| C-2001 | 0.82 | 2026-03-12 | |
| C-2001 | push | 0.31 | 2026-03-12 |
| C-2002 | push | 0.78 | 2026-03-12 |
| C-2002 | 0.65 | 2026-03-12 | |
| C-2003 | sms | 0.88 | 2026-03-12 |
| C-2004 | 0.91 | 2026-03-12 | |
| C-2005 | in-app | 0.85 | 2026-03-12 |
Understand why
Every prediction includes feature attributions — no black boxes
Customer C-2001 (Alice Nguyen)
Predicted: email (0.82), push (0.31)
Top contributing features
Email click-through rate last 30 days (RESPONSES)
4 of 5 clicked
36% attribution
Push notifications dismissed 3x (RESPONSES)
3 dismissed
28% attribution
Segment = high-value, email-responsive peers (graph)
74% email pref
21% attribution
Preferred channel = email (CUSTOMERS)
15% 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: Route each customer through their highest-response channel. Lift response rates 30-50%, reduce unsubscribes by 40%, and save $1-2M in wasted outreach spend annually.
Related use cases
Explore more next-best-action use cases
Topics covered
