Through my work at ZS, I’ve spent a considerable amount of time analyzing revenue leakage. Recently, I contributed to a ZS POV on revenue leakage within the pharmaceutical industry, showcasing public examples of major fraud faced by pharma (and how to mitigate it).
I want to highlight a few key points from the article, and expand the discussion to all industries. After all, while pharma has some unique revenue leakage points, revenue leakage isn’t itself unique.
What is revenue leakage? We can classify it with 2 buckets:
- Internal leakage: poor validation processes, outdated systems, and assumptions that may not match reality all contribute to substantial dollars lost
- E.g. two analysts at an EdTech company process a negotiated discount on a contract that was accidentally assigned to both of them for review, leading to a duplicate discount dispensed
- External fraud: a malicious actor seeks to siphon money and/or other things of value (products, intelligence) from a company, oftentimes by exploiting loopholes and blindspots in systems
- E.g. a $46 million copay fraud scheme in Detroit operated through a system of fly-by-night pharmacies
Whether internal or external, both types of leakage drain the bottom line from companies and organizations. These ill-gotten gains can also go on to foster more societal harm (for a relevant pharma example, false prescriptions can be used to defraud medical insurance payers and obtain pills that fuel the opioid epidemic).
Many companies are aware about the revenue leakage threats they face, but often underestimate the threats, or write them off as “unavoidable loss”. With a 3-step process, companies can instead take charge and systematically seal off a lot of leakage:
- Step 1: Get your data in order
- If I want to review your returns data across products, can I access everything through an easily-queryable data lake, or do I have to jump through hoops and silos across your organization for each piece of data?
- Many organizations are disorganized at best when it comes to their data. If you want to actually catch fraud, an organized set of data is necessary to enable analysis, and to make sure you’re correctly analyzing things together.
- Step 2: Advance your analytics capabilities
- The start of this stage can be as simple as talking to people and gathering business rules; this lets you do an initial sweep of low-hanging fraud that just hasn’t been prioritized or able to be detected until now. Even if some rules are known, there may be other useful information to glean from these chats.
- (Side note: A former professor of mine always emphasized learning the truth about systems by talking to the people closest to the actual work being done. She used this method to prevent one potential disaster to the Columbia space shuttle, but when they ignored her the second time around, it led to the deaths of 7 astronauts.)
- As things progress, implementing a more algorithmic approach to data will enable more comprehensive leakage detection
- Augment rules-based approaches with supervised models
- Make use of unsupervised anomaly detection to find previously unknown patterns of fraud
- The start of this stage can be as simple as talking to people and gathering business rules; this lets you do an initial sweep of low-hanging fraud that just hasn’t been prioritized or able to be detected until now. Even if some rules are known, there may be other useful information to glean from these chats.
- Step 3: Operationalize with intelligent process automation
- As your analytics pipeline and processes are more concretely established, some of the manual aspects can themselves be automated
- This includes building dashboards with automatically refreshing data, improved alerts capabilities for unusual activity, and dynamically applying new insights and suggestions in validation processes
- “Hot tips” and whistleblower information can be valuable data themselves
- With a more streamlined process to categorize and prioritize this ad-hoc data, there can be substantial breakthroughs in revenue leakage analyses
- Freeing up analysts’ time spent on repetitive and established workflows will allow them to venture ahead to the forefront of the ever-evolving game of revenue leakage
- As your analytics pipeline and processes are more concretely established, some of the manual aspects can themselves be automated
We live in a world that’s more complex than ever before, and fraud flourishes in its shadow. There are trillions of dollars that are wasted and stolen in our global economy, so there’s no time like the present to take control and staunch the flow of revenue leakage.