Revenue and customers
Next-Best-Action
Each customer gets the action that does the most for their value over time, inside eligibility, frequency, and fairness constraints, with a global holdout to prove it.
The problem
Channel teams run their own campaign calendars, so customers receive contradictory messages, such as a retention discount and an upsell in the same session. Optimizing each touch for clicks can wear down the value of the relationship.
Who owns it
The marketing or customer-experience decisioning owner, who holds the action catalog and its constraints. A weekly trading meeting reviews how each action performs and retires or launches actions.
What it learns from
- Event streams from every channel
- The offer catalog and its costs
- Customer identity across channels
- Consent and preference state
How it works
Predict, decide, act, measure
The same four steps in every solution, in this order, each one visible to the team that owns it.
- 01
Predicts the value of each possible action
For each customer at each touchpoint, models estimate the incremental value of every eligible action, including doing nothing. Each decision records what could have been shown and with what probability, so performance can be evaluated without bias.
- 02
Decides through arbitration with constraints
A central decision service arbitrates by value within eligibility rules, margin floors, cross-channel contact caps, offer budgets, and a protected exploration budget. High-stakes offers, such as credit-line changes or retention concessions, go through approval.
- 03
Acts in every channel through one decision service
Web and app personalization, marketing automation, and contact-center scripts call the same decision service, so decisions and logging stay consistent across channels.
- 04
Measures against a global holdout
A global holdout keeps the existing rule-based policy. Program-level incrementality is measured on cohort value over time, never on clicks per touch.
What it moves
The business levers
Customer value over time
The objective is the value of the relationship, not the response to one message.
Offer spend
Discounts and incentives concentrated where they change the outcome.
Contradictory contact removed
One arbitration across channels ends the retention discount and the upsell arriving together.
Catalog health
Stale or near-duplicate actions are flagged and retired on a weekly rhythm.
Proof
How the return is proven
The return is the incremental cohort value over the global holdout, stated against the return model agreed up front. Complaints, margin, the fairness of offer distribution, and the exploration floor are reported beside it.
What you see: the ROI readout
Every solution produces the same readout, assembled from the system's own records rather than built by hand.
Correctness
The checks the pipeline passed on your own data: leakage, point-in-time, and entity resolution.
Incrementality
The incremental effect against the holdout or equivalent design, with its confidence interval.
Return
The measured return against the investment, set beside the return model agreed before work started.
Guardrails
Complaints, cost, margin, fairness, and contact limits, each shown against its bounds.
Audit extract
Every score, action, and outcome, with the model and policy versions behind it.
Governance
What is governed
Access, approvals, and the audit trail are part of the solution, not a review step added at the end.
Identity-scoped access
The action catalog, its constraints, and its performance are visible by role through your single sign-on.
Approval gates
New actions, constraint changes, and high-stakes offers need a named approver.
Audit trail
Every decision can answer why this customer saw this offer, with the alternatives and their probabilities.
Fair lending, promotions, and consent
Offer fairness is reviewed for who receives the better prices. Where an offer involves credit, ECOA and financial-promotion rules apply as eligibility, and consent and GDPR preferences are enforced the same way.
Why it holds up
Trained on your data, framed by the outcome
This solution trains models on your own data, checks every stage from data to decision, and is measured against the outcome it was built for. Because its predictive power is measured, it also shows the drivers behind the metric, so your team can fix causes as well as act on scores.
Why solutions hold upStart with one decision
Bring the problem and the systems your team works in. We scope the first solution with you, agree how its return will be measured, and run it in production with your people.