Digital Transformation & Maturity

The troubles with revenue assurance and managing fraud

In this article, Arun R Kapoor, Business Consultant at Tech Mahindra explores the business challenges facing telcos’ revenue assurance and fraud management departments, and the strategies that can be employed, keeping these challenges in mind. Read more in this white paper.

Business challenges

Data explosion

Mobile data traffic has grown 18-fold over the past five years. Global mobile data traffic reached 7.2 exabytes per month at the end of 2016. Newer smartphones powered by high speed broadband connectivity are being adopted at an exponential rate. It is estimated that global mobile data traffic will be 49 exabytes by 2021.

Despite various advancements in revenue assurance (RA) and fraud management systems (FMS), monitoring huge chunks of usage data records remains a big problem.

Analytics challenges

Apart from the continuous monitoring of risk alerts using RA and FMS systems, a 360° analysis of the potential threats and risks is needed. At present, the following are a must for proactive detection of leakage:

  • Usage pattern analysis
  • Demographic profiles
  • Fraud links
  • Fingerprinting analysis
  • Location/traffic analysis
  • Error spikes
  • Rating reconciliations

Generating such analysis reports manually, on top of legacy systems, is time-consuming and error prone. In fact, most analysts believe that a big chunk of their revenue is lost due to time spent formatting and analyzing the data manually, instead of focusing on the actual threat that is ruining in the network.

False alerts

Various RA and FMS tools assign a score to an alert that is generated when any leakage/fraud criteria are met. Usually the scores are set at the time of writing business rules which are rarely changed by the team.

Weekly or monthly reviews of business rule configuration is not always at the top of agenda. The reality of the situation is that the analysts deal with high number of false alarms every day.

If an operator does not have a 24/7 team which cycles through these alerts, then each morning the team is hit hard with big chunk of mundane and tedious analysis jobs.

Querying and toggling

Teams run ad hoc queries on various application interfaces to fetch relevant information. At times, the queries are scheduled in the backend, however the results need to be downloaded and reformatted to get useful information.

Furthermore, jumping between various application screens is also quite distracting. It thus not only increases the fraud/leakage run-time but also affects the overall productivity of a full-time analyst.

Manual updates

 

To optimize the business rules and prioritize alerts, the analysts need to configure industry alerts published by various forums. This requires registering with fraud alerts to receive regular messages from TM Forum/related websites and selecting the appropriate ones to be uploaded into the business rule configuration. Other examples of manual updates are

  • Updating or modifying geographical location (cell site) data
  • Updating rating information
  • Uploading bad debt customer details
  • Uploading credit check information, payment details from banks
  • Configuring custom blacklisted/whitelist account information

These are just some of the repetitive and time-consuming tasks performed by teams outside their actual role.

Additional responsibilities

The TM Forum Fraud Survey report found that RA and FMS teams also have additional responsibilities such as internal auditing, product management, legal and regulatory compliance checks, customized reporting etc. Such manual and laborious tasks cannot be delayed due to high risks, compliance and security aspects.

Strategy tenets

Determining the right strategy tenets considering the above business challenges is the first step in developing and adopting best practice automation and methodology solutions. Consider the following:

Real time reporting

Dedicated teams must be responsible for a wide range of controls and monitor full proportion of revenue streams, also highlighting where mistakes are occurring. At end of each day/week, the reports are generated – predominantly manually. Most of the time, the manual reports generated on fraud/leakage do not tally with management accounts.

First and foremost, management must identify and quantify manual error and volumetric tasks performed by the analysts.

Proactivity

Understanding the factors affecting the fraud/leakage run time is very crucial.

Fixing an issue before it amplifies is the best approach to mitigate risks. Proactive actions mean taking action before any revenue leakage occurs. Unfortunately, most of the controls and business rules that are supposed to be proactively applied are waiting for something to go wrong and then raising the alarms.

As such proactivity comes through better management of data and alarms, bringing more optimized governance of activities and applying risk management techniques.

 

Maturity assessment

Assessing the maturity of revenue and fraud detection is usually computed based on:

  • Organizational effectiveness
  • Scope and completeness of the processes
  • Key performance indicators (KPI) measurement
  • The degree to which tools usage support the day-to-day operations of a function

Each of these dimensions is further sub-divided into various topics. One such sub-topic in the TM Forum Revenue Assurance Maturity Model talks about the degree of automation in RA and FMS technology.

Replacing the transaction labor by making automation one of the primary goals helps to achieve a higher maturity in mitigating revenue loss.

Analytics

So far analytics are limited to running various backend scripts or frontend queries to generate customized dashboards and reports. The real-time value will come from applying analytics and business intelligence to auto-generated data.

Designing a mechanism to automatically extract, format and load data into dashboards would bring true end-to-end analytics into the heart of the business.

Dynamic configuration

Existing RA and FMS tools are equipped with dynamic thresholds and auto-action mechanisms on alerts to avoid revenue leakage. However, cognitive technologies, including machine learning and judgmental-based decisions are yet to be integrated with the existing ecosystem to make it possible to enhance automation to processes that require judgment.



Advertisement:
Share.

About The Author

Business Consultant - Tech Mahindra

Leave A Reply

Back to top