We rolled out the EW AdTarget campaign manager and the EW Navigator platform to power omnichannel subscriber engagement for a mobile operator in Kyrgyzstan. The Eastwind team also developed Next Best Offer models, enabling the operator to increase sales and minimize churn.
Why Next Best Offer Models Matter for Mobile Operators
We launched the core EW AdTarget functionality and completed the migration from the legacy campaign manager by the end of 2024, enabling the operator to kick off marketing activities immediately while we continued enhancing the platform. The operator’s goal went beyond automation, focusing on boosting CVM effectiveness with big data to:
- build more predictable revenue streams,
- maximize campaign revenue,
- strengthen subscriber loyalty,
- minimize churn.
To meet these objectives, we developed Next Best Offer models. It`s algorithms that analyze subscriber behavior and deliver the most relevant offer in real time.
NBO models help the mobile operator:
- move bundle users to more relevant monthly tariff plans;
- remind customers of included services and options while recommending additional activations;
- boost cross-sell and up-sell conversion;
- offer at-risk customers personalized tariff plans from a closed portfolio.
What Work Was Completed During the Project
Accurate predictions at scale require a robust data infrastructure. The Eastwind team built it from the ground up:
- implemented a daily pipeline for ingesting, cleansing, and validating data across calls, SMS, data usage, charges, and subscriber interests;
- engineered data marts with 200+ features, including automated feature selection and ranking;
- trained and deployed 33 NBO models across B2C and B2B segments;
- enabled daily feature recalculation and prediction updates, with results seamlessly fed into EW AdTarget to power campaign execution.
We delivered an end-to-end solution for the mobile operator, selecting the optimal technology stack for data pipelines and data marts, deploying all required components within the customer’s infrastructure, and ensuring ongoing support and system stability through continuous monitoring.
Pavel Olifer,
Director of Business Development, Eastwind
The system continuously processes data, updates data marts, and identifies the most relevant offer for each subscriber. For users at risk of churn, a dedicated model delivers personalized retention tariff plans from an exclusive portfolio.
How NBO Integration with EW AdTarget Drives Value
NBO model outputs seamlessly flow into EW AdTarget, empowering the operator’s CVM team to launch fully personalized campaigns without any manual setup.
Benefits for the telecom operator:
- fully automated marketing: models continuously retrain in streaming mode as customer and product data changes, with no manual effort required;
- higher conversion driven by more accurate and relevant offers;
- stronger user satisfaction through proactive recommendations and reminders about included options.
In total, Eastwind specialists developed and trained 33 ML models, selecting 30–50 of the most informative features for each.
How Long Did the Project Take
From the first discussions to delivering the final models, the project took about 1 year and 4 months, covering hardware procurement, process setup, and data preparation. The final step was enabling the operator’s team through training.
With Eastwind solutions, the mobile operator can move beyond mass messaging to fully personalized omnichannel engagement. This approach boosts ARPU, builds stronger subscriber loyalty, and helps preserve the customer base.
Discover how combining EW AdTarget with machine learning can drive value for your business
