PortfolioCo*, a large enterprise software company with a mature product portfolio formed by a merger, had declining revenue. The company believed that revamping its license compliance (LC) processes could compensate for falling revenue by ensuring that customers were paying appropriately for the licenses and service agreements they already had. It knew that a significant number of customers had fallen out of compliance, either by deploying more licenses than they had purchased or by violating usage agreements. As a result, PortfolioCo was missing out on revenue from licenses and support agreements.
The LC process is relatively straightforward: Identify customers believed to be out of compliance, perform an audit, and, if there is evidence of noncompliance (that is, a “hit”), negotiate a settlement. To meet its goals for increased revenue, the company ideally wanted to have a high hit rate, preferably targeting companies that were likely to yield bigger settlements, and resolve issues in the shortest possible time.
But as a result of the merger, PortfolioCo found itself with two distinct LC teams, each with its own way of operating. The company also had problems tracking audit activity, challenges in maintaining data quality and difficulties in forecasting.
PortfolioCo partnered with Bain to address these and related issues. The Bain team worked with the company’s leadership to define the future plan—namely, how many reviews would need to be started, what operational capacity would be required to do the work, and whether yield from those starts would meet revenue goals. The Bain team also provided a deep dive on historical data to help PortfolioCo’s leadership understand business trends and prior performance. Bain then developed a dashboard that provided visibility into the LC pipeline, much like a traditional sales pipeline, and monitored the achievement of targets.
The consulting team also tapped the capabilities of Bain’s Advanced Analytics Group (AAG) to create a tool that forecasts customers’ propensity to be out of compliance, which would help identify and prioritize which customers to review. The AAG team was able to develop the tool in less than 12 weeks, using its proprietary use case factory production methods.
PortfolioCo and Bain kicked off the effort by answering a few key questions: Which products will be audited? How much time should elapse between audits of a given customer? Should certain customers be exempt from audits?
The team then categorized different types of noncompliance:
- Violating license usage terms;
- Giving inappropriate access to third parties;
- Using the software outside the geographies for which it’s licensed;
- Using more seats than were paid for;
- Illegally copying the software; and
- Continuing to use support services after the maintenance contract has expired.
To increase the hit rate, the Bain team identified drivers that correlated with each of the varieties of noncompliance. For example, different license versions in support tickets might indicate illegally copied and inconsistently updated software; if support growth outpaces license purchases that could indicate unauthorized over-deployment, or granting third-party access to the software.
Bain collected, cleaned and analyzed a range of internal and external data. The AAG team addressed a number of data challenges, ranging from a low sample size for certain kinds of audits to missing and contradictory data to missing or multiple identifiers, and applied its analytics expertise to find reliable workarounds and viable substitutions. The propensity tool focused on producing three distinct outputs: findings value (that is, a predictive model for calculating the dollar amount of noncompliant licenses), hit rate (the likelihood that an audit would result in revenue), and audit length (the time from audit start to audit close).
The team tested variables for correlation with historical audit outcomes and key metric baselines, and they developed individual regression algorithms for each output. The company (including LC leadership, the operations team and, potentially, field personnel) then tested the drivers to account for business context and further iterate the model.
The final version of the propensity tool delivered three outputs per PortfolioCo client: likelihood of a hit, likelihood that the hit would exceed a certain revenue threshold (that is, avoid audits of small deals) and likelihood that the audit could be completed in less than six months.
Good to Go
The Bain team prioritized a seamless handoff of the propensity tool to PortfolioCo, with both the consulting team and the experts from AAG providing training. The goal was to ensure that the company could run the tool without Bain involvement going forward. Toward that end, Bain addressed key issues running the tool (including the creation of an initial target list, the data inputs needed and generating outputs) and updating the tool (including the creation of a product roadmap, testing correlations and updating the SQL code). Bain provided significant coaching and a playbook to help the company develop a high level of internal capability.
Ultimately, PortfolioCo was well positioned to not only use the propensity tool effectively but also to improve and refine it on its own through a process of Agile sprints. The company’s LC team now uses a dashboard to review progress weekly and spot bottlenecks, and has increased the targeted number of audit starts in each region, meeting its goal of boosting LC revenue by 60% within the first year of full implementation
* We take our clients' confidentiality seriously. While we've changed their names, the results are real.
