Standing on the Shoulders of Giants: The Future Opportunities in Data Analytics

Data analytics changed my career once, and I think it is going to do so at least once or twice again in the coming few years. When I was a staff auditor at Ernst & Young, the firm realized we had blown an audit. We had issued a clean opinion for a large call center operator, but found out after the fact that it seemed customers had been getting overbilled in many cases. It turned out that in assessing how the minutes of calls taken in the call center actually landed on a customer’s bill during the audit we had tested just a single call. The auditor watched the agent take a call, validated that the minutes had been captured on the phone switch, and then traced the time through MANY complex data warehouses and data transformations before seeing it land on the bill to a customer in SAP. After verifying that one call (out of hundreds of millions), we had declared the call data processing an “application control” and therefore we could rely on the call processing for all calls based on the successful walkthrough. Of course, there were actually millions of calls being “accidentally” (according to the client) inflated in their billable duration during those complex data transformations, resulting in overbilled customers and, in turn, misstated financial statements.

To address this issue, I was asked to learn how to use ACL for data analytics in a single week, and unwind the mess to find the overall error. I nearly died of sleep deprivation that week, but in just those seven days, I built analytics that replicated the billing logic defined in the customer contracts, extracted hundreds of millions of phone calls from 40+ phone switch servers around the world, and completely validate the correct bill amounts. In aggregate, we found customers were overbilled and revenues inflated by almost $40 million. After almost two years in the job, I realized then that I could love being an auditor if I focused on using data analysis to audit right through the computer (rather than manually testing and auditing around it).

That was 12 years ago, and I’m still saying the same thing… one day data analytics won’t just be a skill some technical auditors have, data analytics will be the backbone on which every audit is primarily based and will eliminate the need for 80% of what auditors do in the field today.

I am thankful for is the opportunity to work at ACL as Chief Product Officer, thus having an incredible platform to imagine what that future looks like, shape the technology that will deliver it, and work with an incredible community of people like Dave Coderre who are leading that change. Dave once presented to our team at ACL a very simple slide that said the sweet spot for success with audit analytics was simply being able to combine 1) knowledge of audit methodology, 2) technical ability to use data analysis, and 3) a real understanding of the business’ strategy and operations. I could not agree more. Unfortunately however, I would suggest that maybe only 10% (if that) of professionals in the field do have a reasonable level of acumen in all three. Therefore, it stands out to me that my job developing technology products is to find ways to fill that gap.

Today, the world has more analytics tools than it could probably ever use. I walked around the business intelligence section of the vendor show floor at a large Gartner Summit event recently. I must have talked to 50 vendors, each with some variation of the exact same pitch, “we will find the insights in your data” followed by some variation of the same question, “what are your analysis requirements?”. Therein lies the problem.  As Dave had told us, the user must combine all three components to answer that question. I think we need to get people over that hurdle, especially in audit and risk related professions. Up front, most users don’t really, precisely know what they are looking for in data or what the results may mean in the broader context of their business, and that’s probably OK. If I go to a furniture store to purchase a new sofa, I don’t know going in precisely what I want the sofa to look like or how it might fit within the overall feel of my home. However, I know the one I want once I see it. That is how I see us continuing to evolve analytics forward – help users find those things they didn’t know they needed to be looking for and help them understand it when they see it.  I believe we are going to bridge that gap in three ways:

1. Technology evolution – Advanced analysis techniques, machine learning in particular, need to become accessible and understandable to front line users. Machine learning holds the potential for a revolution that we have only begun to comprehend. IN a similar way that data analytics completely transformed my ability to audit hundreds of millions of phone calls, machine learning has the potential to completely transform that experience again by replacing the need for me to spend seven sleepless days mapping out data tables and defining queries. Instead, the computer will be able to train its models to look at a data lake of phone call information and self-define the data structures and relationships necessary to find the anomalous call activity.
2. Content – There is more financial opportunity in front of all of us today in packaging and productizing analytic applications than there is in selling the analytics software tools themselves, and anyone could benefit from this. We need “App Store” like marketplaces where practitioners can distribute their high impact data analytics to a global audience. In other words, rather than spending seven days self-discovering how to connect to the phone switches, pull the call data, and normalize it into something that could be analyzed, I will go to the marketplace, find an “analysis app” built by someone else who already had that problem, and purchase their solution and avoid reinventing the wheel.
3. Community – We find the big value in being an ACL customer is in being part of a community of 14,000 other organizations that are working on similar challenges. Conferences, user groups, and online forums are historically where this happens. However, as we look to the future, we’ll see the products working to connect a user to the knowledge in the community around them by providing intelligent benchmarks from peer users, making social connections based on similarity of profile and usage interactions, and enabling peer insight discovery. In other words, users will have at their fingertips the ability to finish sentences like “other call centers like us tend to have [number] key controls in their billing process to mitigate billing errors”.

I look forward to building on the successes of data analytics over the years and the work of pioneers like Dave Coderre to figure out how we can empower a million “mini-Daves” around the world. Leveraging these future trends, we will enrich the perspective and value of data analysts while easing the skill breadth burden necessary to achieve impactful and sustainable successes with data.