Why Revenue Assurance Should Embrace Data Science

One good thing about running a website like Commsrisk is that you see objective data which shows what matters to other members of the community. I recently wrote an article entitled ‘Revenue Assurance Has Lost Its Way’ and it was clear from the visitor statistics that many feel the same. The piece ended with a plea for leaders to offer a new direction to those revenue assurance professionals who find themselves stuck doing the same old leakage monitoring for revenue streams that are now in permanent decline. I asked for new leadership although I am not confident that enough of the right people will step forward to fill the gap. How else might a revenue assurance professional enhance their career prospects if their bosses refuse to outline a new roadmap?

Before we get too pessimistic, let me emphasize that it is still possible to re-establish revenue assurance on firmer foundations, leading to fresh investment in the discipline and renewed growth. The key would be to have leaders who are not just calculating how to move on from revenue assurance in order to secure a better job for themselves, but who will instead show how more can be delivered by revenue assurance. They would need to describe the route forward to their peers in other telcos and the vendors seeking to sell to them, so that the establishment of a new consensus leads to a focusing of resources. That kind of direction will only be authoritative and credible if it comes from individuals who manage actual RA teams. It cannot come from vendors, consultants or other commentators like me. The problem is that telco managers may see no advantage in encouraging innovation; promotions are more easily secured by meeting a boss’ expectations than by changing a boss’ expectations. Innovation was central to revenue assurance when there were no telcos that had revenue assurance functions, but the desire to innovate was depleted as revenue assurance steadily became the norm. This norm has evolved into a prison for many revenue assurance professionals who are expected to keep following yesterday’s norms instead of repurposing revenue assurance for the challenges of tomorrow.

The majority of conference speakers who work in revenue assurance will assert the discipline has a role that extends beyond the analysis of data. Their point is that there is a difference between observing patterns in numbers and transforming those patterns into meaningful stories about how to change the business and why it should change. Many revenue assurance teams will have sometimes wasted the efforts of their staff by tasking them to identify anomalies without anyone knowing what to do about those anomalies. Fixing flaws and improving the business also requires alignment with the motivations of other managers and an appreciation of all the business’ goals. Revenue assurance is not just data science. On the other hand, we should not glorify revenue assurance by pretending the average RA team always has a good understanding of the work done by other parts of the business. If the goals of RA clash with other functions then that might reveal a limitation in the thinking of RA practitioners as well as problems with the targets and incentives given to other staff. Good revenue assurance can be much more than seeing errors in a stream of data, but not all revenue assurance is that good. Whether revenue assurance is good or bad, it should still use data to identify costly failings that nobody else has identified. In that sense, there should always be a fundamental overlap between the skills needed for revenue assurance and those developed by a data scientist.

The suggestion that RA professionals should construct a career path by becoming more adept data scientists often encounters resistance. There is no obvious downside to suggesting a revenue assurance practitioner should be taught the blend of advanced methods in statistics and programming that is the prerequisite for an effective all-round data scientist, and it is obvious that many telco departments make no effort to provide such training to their teams. Those few that do spend worthwhile amounts on training are as likely to teach their staff about information security as data science, which is like opting to teach a financial auditor how to fight a literal fire instead of training them how to extract and visualize data from information systems. More fire fighters is an advantage when there is a fire, but little is gained by giving an individual an understanding of how to manage risks if they will not be expected to manage those risks in real life. The majority of RA professionals should be conscious of how their work experience is aligned to other roles involving data analysis. RA teams soon learn why organizations have bad data, and this can be a difficult lesson to teach to the uninitiated. They may only be focused on specific kinds of data, but their experience is relevant to the generation and analysis of data more generally. The transferability of skills between revenue assurance and data science should at least be acknowledged, even if some individuals will prefer to extend their skills by mastering allied regions within the domain of enterprise risk management.

A recent article by Frank Scarduzio, Vice President of Revenue Assurance and Operations at analytics software company Qlik, illustrates how revenue assurance skills could extract value from data for other kinds of companies. The main requisites are that the analyst has access to the relevant data and the freedom to make full use of it. Scarduzio prefers to talk about revenue operations, or RevOps for short, but it is obvious that this is just another name for a discipline that is slightly more general than revenue assurance, and somewhat more specific than data science. This is how Scarduzio explains it:

The role of the RevOps team — which could either sit within a business function or run as a separate entity to complement other functions — does not take overall responsibility, but provides an objective lens in the decision-making process. It does so by pulling together data sources from across the organization and uses analytics tools to derive new insights to guide business leaders in their next move.

Scarduzio quotes an IDC survey of 1,200 global organizations that said 75 percent of business leaders witnessed increased revenues and 74 percent reported increased profits as a consequence of using analytics to turn their organization’s data into useful insights. This was supported with the example of CSC ServiceWorks, the largest provider of laundry services in the USA, which was said to have saved over USD2mn each year by analyzing patterns from its customer app and a million laundry devices in order to eliminate unnecessary travel by maintenance trucks.

The resistance to treating revenue assurance as a particular form of data science appears especially bizarre when we consider who has grasped the mantle of revenue assurance leadership for want of more leaders from telcos. Here is a clue to the identity of these alternative leaders: they offer marketing disguised as leadership and they spend a lot of time saying that RA needs more artificial intelligence and/or machine learning. Most RA systems are just specialized tools for data analytics. They are sold by firms that could just as well develop software to analyze the data produced by other kinds of business, with the intention of satisfying any goal that can be measured using data. The absence of leadership from telco managers means the de facto leaders of revenue assurance are vendors of systems that crunch data in a more or less scientific fashion. Some of these firms employ people who literally refer to themselves as scientists. If revenue assurance should not be treated as a branch of data science then it is paradoxical that the people who most commonly encourage innovation in revenue assurance are those who are motivated to sell tools that examine data. For revenue assurance to be more than data science then it needs to have a development roadmap that includes more than just a single entry demanding that every RA team must implement AI/ML.

The greatest advantage in turning revenue assurance into a branch of data science is that it will give RA teams the freedom to satisfy a much wider variety of objectives within their companies. I had a hand in writing the original TM Forum definition of revenue assurance, which made it plain that the common root of the discipline lay in the methods that were used, not the goals that were pursued. This conflicted with the preoccupations of many previous commentators who insisted RA should be defined by a strictly prescribed series of objectives, with no reference to the techniques being deployed in practice. This led to the paradox that commentators would demand a definition that included a list of objectives even though each list was inconsistent with the objectives listed by others. To place tight limits on the ends of revenue assurance was short-sighted because different kinds of telco should set themselves different financial priorities and objectives.

The impact of this mistake is apparent even today, as it can be seen whenever a revenue assurance professional describes tackling margin assurance as if it is a radically different proposition to doing revenue assurance. There were some RA teams who immediately undertook margin assurance from their inception; they were not wrong to do so. Different organizations will take time to evolve through a different series of phases, but there was nothing about the methods used for revenue assurance that demanded they first be used to treat revenue leakages, and only later be deployed to identify margin issues. The rather vague call for revenue assurance to transform into business assurance is really just another way of saying the methods used by revenue assurance might be utilized to address a wider range of challenges. This tends to ignore that some RA teams have long extended their remit well beyond the more rigid and limited definitions of revenue assurance. Ambitious revenue assurance managers should treat a pivot towards data science as being beneficial for the same reasons they pursue the wider scope associated with business assurance.

Heads of RA departments have much to gain by embracing a more extensive remit, but their staff will also enjoy an important career benefit by embracing data science. Put simply, there are many more high-quality training courses available for data scientists than business assurance professionals. As those skills are more generally recognized, and can be more generally used, they make the recipient of training more attractive to a wider range of potential employers. That might be one reason that selfish businesses may refuse to train RA staff in data science; they will fear departures from team members who realize they can command a higher salary elsewhere. RA professionals should agitate for more data science training, whilst pointing out the very large variety of courses on offer, including many which are inexpensive or short. If they can make themselves more attractive to other employers then they also have the chance to improve the pay and conditions in their existing company, and that will also feed into making revenue assurance a more attractive and vibrant discipline overall.

Eric Priezkalns
Eric Priezkalns
Eric is the Editor of Commsrisk. Look here for more about the history of Commsrisk and the role played by Eric.

Eric is also the Chief Executive of the Risk & Assurance Group (RAG), a global association of professionals working in risk management and business assurance for communications providers.

Previously Eric was Director of Risk Management for Qatar Telecom and he has worked with Cable & Wireless, T‑Mobile, Sky, Worldcom and other telcos. He was lead author of Revenue Assurance: Expert Opinions for Communications Providers, published by CRC Press. He is a qualified chartered accountant, with degrees in information systems, and in mathematics and philosophy.