Real-time and Machine Learning for Marketing Campaigns

Prior Situation / Scenario:

  • Batch campaign operation
  • Too many messages with little or no context being sent to users
  • No offers being sent based on end-user actions (Purchase, top-up balance, etc.)

Client Challenges:

  • Subscribers complain of receiving too many messages
  • Incentives sent almost to every suspected subscriber
  • How to automate the Contextual Real-Time Campaigns

Strata Solution/ Key Enablers:

  • Real time campaigns for timely, appropriated and relevant offers
  • Developed an algorithm to predict the probability to recharge or consume balance.
  • Apply a contact policy that integrates batch and real time campaigns
  • Run control groups and A/B testing to understand the impact of offers and campaigns
  • Analytics dashboard to monitor performance


  • Real time contextual offers are aligned with subscriber’s context and activity
  • Consumers perceive the uniqueness of his/her offers
  • Increase in campaign effectiveness


Within the first six months of the CBM & Marketing Automation program, subscribers’ incentives were reduced by 42% and take-up rate increased by 29%.