Artificial Intelligence

Opti-Num partnered with an international telecommunications provider to design and deploy a real-time, regionally-personalised, dynamic pricing system. 

CLIENT
International Telecom Provider
sector
Telecommunications
Read time
5 Min
Overview
The Challenge
The Solution
The Results
The numbers

The Challenge

An international telecommunications provider was struggling to stay competitive in a fast-paced prepaid market driven by price-sensitive customers with a strong preference for voice calling. Customer churn was high, and the provider lacked the tools to quickly respond to competitor offers or dynamically manage network capacity.

Demand across the network was uneven and unpredictable, spiking at certain times and locations due to subscriber circumstances, pay cycles, new offer launches, seasonal events (e.g. Christmas, Easter), and large gatherings like sporting events or the morning rush hour. Static pricing simply couldn’t keep up with the speed of change in customer behaviour based on location, timing, or marketing cycles.

They needed a solution that could adapt to these conditions in real time, improve customer retention, and provide subscribers with maximum value while optimising revenue return.

The Solution

Opti-Num partnered with the telco to design and deploy a real-time, regionally personalised dynamic pricing system. At the core was a robust data science and software engineering solution that combined behavioural modelling, geographic segmentation, and time-based triggers to deliver tailored offers at scale.

We built predictive models to analyse how prepaid users engaged with various price offers. These models learned over time, using live feedback to fine-tune predictions and pricing strategies. For example, a prepaid user might receive a targeted offer at a specific time and location—say, during the morning commute in Cape Town—based on past buying behaviour.

Technical Features

  • Granular Discounting Engine: Adjusted pricing every 15 minutes at the cell-tower level across South Africa.
  • Behavioural Models: Built using supervised learning with feedback loops for continuous improvement. These models incorporated usage patterns, geographic data, and temporal trends.
  • Elasticity Forecasting: Simulated pricing strategies and forecasted impact on uptake, revenue, and network load.
  • Event Response Logic: Triggered real-time responses to demand spikes driven by pay cycles, seasonal trends, or high-traffic events.
  • Model Operations: CI/CD pipelines enabled seamless deployment, retraining, and versioning of models.
  • System Integration: Fully embedded in the client’s operational environment—no manual updates required.

Example: A user commuting through a high-traffic zone at 7:45 AM might receive a targeted SMS discount based on recharge behaviour and current network load on that tower.

The Outcome

The result was a dynamic pricing platform that responded quickly to market shifts and user behaviour. In 2021 alone, up to 30% of the client’s billion Rand prepaid market was supported by our solution.

It quickly became the go-to system for responding to market events and recovering revenue when KPIs were under pressure. Senior leaders came to rely on it not just for operational flexibility, but for hitting business-critical targets.

Beyond revenue growth, the telco experienced:

  • Improved customer retention
  • Better demand balancing across the network
  • Increased competitiveness through pricing agility

What started as a pricing tool evolved into a strategic platform for network-wide optimisation. Opti-Num Solutions maintained and enhanced the system for over a decade, continuously expanding its reach with advanced algorithms and broader product coverage.

Figure 1: Dynamic Pricing Methodology

Figure 2: Visualising Market Response percustomer segment, over an average workday
Products used
Hardware Used

Latest Posts