The Future of Customer Experience
Is Data-Driven

history, and transaction history. Product features include constituent attributes, price points, eligibility rules, and market features such as retention, margin, sales volume, and revenue characteristics. The AI models then estimate the market performance of newly created products by simulating “counterfactuals”—for example, hypothetical markets where the newly designed product is available for purchase. 

The outcome of the model is an estimate of the market characteristics of the product, which can then be compared with the product manager’s own estimates to design market-optimal products.The user journey while designing a new product offer could be characterized as follows:

The implementation of the key solution components such as product catalog, ordering and inventory management, and more, complies with the TM Forum’s Open Digital Architecture (ODA) and open application programing interfaces (APIs). An open, standards-based information model ensures a common data model and a common data dictionary to be adopted, measured, and benchmarked.

Figure 1: Offer design journey
click to enlarge

Let’s look at a specific example. A new mobile offer is being designed in preparation for launch in a major market. The offer bundles together data, SMS, and a collection of digital services, including chat, music, and movies with a promotional free period of three months and certain exclusions such as international roaming. Based on analysis by the business analyst, the pricing recommended was Rs 399/month.

The bundle is targeted at both existing and new subscribers, the goal being to increase revenue with bundled services. The specific KPIs expected to be met were:

1. Increase revenue by 3 percent.
2. Grow the number of subscribers by 7 percent.  
3. Reduce churn by 3.5 percent.

The offer specifics and KPIs were fed to the optimization and modeling engine (AI product engine in Figure 1 above). 

Figure 2: Model Accuracy
click to enlarge

The model recommended increasing the package price by ~9 percent would still deliver the three business KPIs. Had the offer design been done without the pricing optimization engine, quite a bit of money would have been left on the table. The results