CASE STUDY

Delivering 1:1 Personalisation Through Algorithmic Experimentation

What if your organization could instantly and simultaneously process, understand & crunch all your terabytes of data that you have collected over the years, and provide you with the deepest level of personalization possible within marketing today?

What if you had at your disposal a truly AI powered marketing engine that could experiment with millions of combinations of offers & microsegments everyday, and provide you with the best offer for each customer?

Find out how qbo Decisions helps one of the largest telecom companies in Asia achieve an unimaginable scale of personalization, using its proprietary, AI-powered marketing robots.

Goals

Improve key KPIs such as Mobile Internet penetration, Customer Retention & ARPU stretch on the telecom customer base

Adopt marketing experimentation at scale powered by Artificial Intelligence to better understand customer behavior and preferences.

THE CHALLENGE

Marketing Experimentation on Big Data

Running an optimal marketing experiment on a large customer base with millions of customers with hundreds of available features is next to impossible using traditional A/B testing & marketing experimentation tools. Traditional marketing experimentation is hamstrung by a number of key disadvantages…

Selection bias in segmentation: Traditional experiments tend to reuse marketing segments from previous experience

Non-optimal usage of data due to cognitive overload: Traditional experiments are manually configured by users, and will inherently ignore large portions of insights that are lying hidden in the big data that a telco possesses.

Incorrect hypothesis: Due to the above a marketing experiment starts out with an incorrect hypothesis, resulting in sub-optimal conclusions on the optimal offer to be sent to each subscriber

THE SOLUTION

Higher take-up rates vis-a-vis BAU performance with qbo decisions

qbo decisions’ marketing robots runs marketing experiments on hundreds of thousands of customers iteratively. The Robots are able to identify key features among hundreds of available customer attributes, and experiment with thousands of combinations of features and offers over the course of multiple iterations.

qbo decisions runs these marketing experiments on a subset of a defined customer base while other, traditional campaigns are still in flight. The system constantly benchmarks itself with learnings from each iteration and compares the results of experiments with customers that are part of control groups as well as the business as usual campaigns

THE RESULTS

Tangible results in weeks not months with AI-driven marketing execution

Discovered thousands of segments & their optimal combination with the best offers, resulting in increased take up rates.

Improved key performance indicators around mobile internet up-sell, retention & ARPU stretch.

Uncovered hundreds of high potential micro-segments that were previously under marketed to by the organization.

Definitive insights for the business on which offers work, and which ones don’t.

“The machine has no feelings, it feels no fear and no hope ... it operates according to the pure logic of probability. For this reason I assert that the robot perceives more accurately than man.”

— MAX FRISCH, Homo Faber: A Report.