A better solution is to use an offer-optimization process that calculates the best possible plan prior to a customer interaction. A well-designed offer-optimization solution based on analytics considers all relevant data elements and many other complex factors, including constraints. It takes into account changing usage patterns, business objectives, competitive tactics, economic conditions, and the introduction of new devices and services.

The optimal offers should be continually revised as conditions change. Maintaining an analytically derived best plan for each customer is a proactive approach that significantly improves the odds of retaining profitable customers and selling new services.

An offer-optimization process should also revise the pre-calculated best-price plan during a customer interaction if the customer reveals new factors affecting his or her decision. For example, the pre-calculated best offer may have a low usage cap, but the customer may have revealed that he or she prefers to have a stable monthly bill. The offer-optimization system should consider this information and give preference to plans that have a fixed monthly fee rather than a usage cap.

A disciplined offer-optimization process that relies on analytically calculated price plans tailored for each customer will give your company the best possible strategy for aligning offers to business objectives. Plus, customers who receive the right offers the first time around are far more likely to remain loyal.

Enhancing collections-scoring models improves the bottom line

At many CSPs, collections departments have plenty of past-due accounts. But time limitations and budget constraints force them to focus on just a small percentage of customers who are the best candidates for collections activities.

Consider one network operator whose collections team was responsible for maximizing collections on past-due accounts. The team had limited resources, so it needed a prioritized list to identify the best candidates for daily collection actions.

Each night the operator posted payments and new charges to customer accounts. Only after this was completed could the team run its collections-scoring model, which had to be executed within a three-hour window, so the model was run against the 1 percent of customers who had the highest outstanding balances (meaning 99 percent of customer records weren’t even scored).

However, using in-database scoring, the model was run against all 40 million customer records in only 4 minutes, and the entire scoring process was completed in just 12 minutes. By including all customers, the model improved by 13 percent; likewise, the improvement in collections is projected to be more than $1 million per month.