Built for: CX and insight teams.
What you can do
A score for everyone
An estimate of how each customer would answer your survey question, with an honest range, and a clear dash where there is too little to go on.
Predicted beside surveyed
Predicted and surveyed NPS side by side, who you are hearing from, and who to survey next.
The people who need attention
Every scored customer, and the ones whose predicted score needs attention now, with alerts when risk spikes.
Why a score is what it is
Open any customer to see how far each signal moved their predicted answer. The drivers view ranks the same signals across everyone, and lets you follow a driver to the customers it moved.
Consumers or accounts
One score per person, or one per account built from its contacts weighted by role.
Export what you see
Export the customers and figures on screen for your own analysis.
Where the score comes from
- Your own survey answers, through an opt-in connection to Customer Experience that an admin switches on, or answers imported from another survey provider.
- Signals that exist for every customer, such as deliveries, service tickets, usage and billing.
- Anonymous surveys, restricted responses and people who opted out are never used to train it.
- Scores refresh nightly, when new answers arrive, and on demand.
Whether to believe it
A share of survey answers is held back from the model, and a model has to beat a simple baseline on those answers before anything is published. The trust view shows the result and whether the model is drifting. The best model is chosen automatically unless an admin pins one, and every change, and every time someone opens a named customer’s score, is recorded in the activity log.
Works with
- Churn Modeling: which customers are likely to leave, and why.
- Customer Experience: a complete feedback programme across locations, accounts and journeys.
- Closed-Loop: turn unhappy responses into owned, tracked cases.