MCC: What types of engagements do you typically handle on the data analytics side?
Sehgal: Our engagements can range from smaller dispute analyses to major high-profile investigations. In all types of cases, our clients are often facing a seemingly insurmountable amount of data that needs to be extracted, validated and analyzed. Due to an exponential increase in data volume and increased complexities of systems that store this information, clients are quickly realizing that data analytics professionals are critical to the success of any engagement. We have years of experience along with proven methodologies that enable us to help our clients successfully navigate through challenges posed by complex data sets.
MCC: How do you define data analytics?
Sehgal: As you can imagine, the definition of “data analytics” is wide and multifaceted, but in my 18 years of experience in this area, the overarching theme comes down to collecting and analyzing vast amounts of data in order to make the most sense of it. The process consists of inspecting, cleaning, transforming and modeling data with the goal of highlighting actionable information, suggesting conclusions, and supporting decision making.
Data analytics has always been crucial, but is now a critical aspect of any investigation. This phenomenon started because data storage capacity suddenly became very cheap and companies accelerated generation of vast amounts of information – I read that approximately 90 percent of all corporate data has been created over the last two years. The good news is that the data analytics technologies have kept pace as well. Long ago, we had to manually write code and develop procedures to analyze data and build reports and visualizations, but the tools we utilize today have a number of built-in analytics and reporting capabilities. The key is to leverage our experience and these capabilities together to efficiently manage your data challenges.
MCC: So timing is an important factor.
Sehgal: That’s exactly right. I’m a firm believer in getting information to counsel as soon as possible, usually within the first few days of the start of an engagement. It helps them make strategic decisions and enables us to very quickly glean the key points we can share in helping to drive the project. If the preliminary results are what you expected, your strategy is on point, but if they are not, you have the ability to effectuate changes to your strategy. I've found it is very helpful to share as much information as possible up front. Our clients obviously have requirements to provide updates to key stakeholders, and arming them with the right information as soon as possible is our focus.
To accomplish this, we clearly define the scope of work and a plan of attack to minimize the time between getting the data, validating it and delivering a “first cut” of analyzed information to counsel.
MCC: Please expand on the validation process and your approach to data analytics. It seems this cuts to the heart of delivering an excellent work product.
Sehgal: It does. It is essential to validate and quality review the integrated data against the source data to ensure completeness and accuracy. As you can imagine, extracting and integrating data from multiple complex systems and sources can be challenging. If not done correctly, it can have dire consequences. We follow specific methodologies to help mitigate these risks. Data validation may include cardinality checks, counts of rows, totals of financials that may exist in the data, date range checks, comparison to existing reports, etc. We then review this information and look at how it plays out against our expectations. When validation results are contrary to expectations, we’ll take a deeper dive to ensure an accurate understanding of the data. Take as an example a financial investigation of a broker-dealer, which requires looking at transactions in one data set and positions in another. During the validation process, we may discover that certain transactions don’t have matching positions and vice versa. This may be fine, let’s say because of cancellations or stop payments, but this is the kind of test we’ll run right away because the results can significantly affect the outcome of the analysis. As we proceed past the preliminary stage and increase our knowledge of the data, slight changes are common, and we will advise immediately if we encounter an unexpected large swing.
Much of what we do is the result of experience, in my case nearly two decades in the field. Our comprehensive methodology for executing a data analytics engagement follows a series of main steps, with a lot of internal work behind them.
The first step is extraction, which involves a lot of back-end planning to develop a targeted approach. We first learn what the client is trying to accomplish, and then we can think about what meaningful data to include and, importantly, what not to include. The more you collect, the more you have to cleanse, validate and maintain – so it is critical to define the scope of projects right from the start. Expanding on my prior example, if we are dealing with a financial investigation, we most likely don’t need HR data, or if the investigation is limited to a certain time frame, that’s precisely where we will target the extraction.
From here, we validate the data to acertain the completeness and accuracy. We are very mindful of the fact that if our integrated data is wrong, anything we produce will be wrong. Again, this is a critical step, and it’s a step that many overlook.
The next step is analysis of this data. There are certainly a number of steps involved here, but the overarching theme is, what exactly is the data telling us? What is the story behind all of this information? We always work towards providing responses to our main goal but are also always cognizant of not overlooking details. What I have learned over the last two decades is to pay just as much attention to things that may not seem as relevant. The devil really is in the details in data analytics.
Another major aspect of investigations, and, frankly, the most overlooked aspect, is client communication. I like to put myself in my clients’ shoes. We expect our clients to completely rely on our work and trust what we did is accurate. This can be nerve-racking for many. We try to alleviate their concerns by keeping them apprised at each step of the process. Some of our clients have worked with us before and understand the process, but for our newer clients, I want them to be comfortable about the fact that we are not a “black box” service provider. Transparency as to process and results underlies all that we do.
MCC: How does big data tie in with data analytics being used in investigations?
Sehgal: In the context of accounting, financial or enterprise-wide systems, although the data has grown exponentially, it does not refer to the big data everyone is talking about. The term "big data" relates to semi-structured data being used in areas like marketing and retail, and has already had a significant impact on these businesses. It requires new forms of integration and analysis to uncover hidden values from large data sets that are diverse, complex and of a massive scale. Big data, however, won’t replace transactional data, which we are used to seeing and will continue to see in investigations. Transactional data is used when your information needs to be completely accurate, like your banking systems. Big data systems will not replace these transaction-based systems, but it will play a role in future investigations. Right now, we are in the process of determining the effect on litigation and investigations. While the majority of my work remains on the transactional side, we know that big data is here to stay, and we recognize that it’s important to get on board now and embrace the current technologies so information can be used to our benefit.
MCC: How does AlixPartners ensure global compliance around data protection?
Sehgal: On the compliance side, we are constantly facing issues about data protection and privacy. These issues are only magnified with the high-profile investigations we read about in the Journal or the Times. To address these and other needs, AlixPartners has robust data processing centers around the world to ensure data security and deal with jurisdictional issues. In many instances, we can't just take data from Asia or a European Union country and move it to the U.S. To resolve this, we have established secure data centers in Asia and the European Union.
From a security standpoint, our data centers are routinely and rigorously tested. We regularly perform penetration testing to ensure our data centers are up to date and secure.
MCC: Has the business community truly recognized the impact of big data? Are these issues here to stay?
Sehgal: Yes. In fact, big data is only getting bigger. Companies that put their proverbial heads in the sand are doing themselves a huge disservice. This is no ordinary trend; it's literally what's happening in the industry right now. As storage costs continue to decrease, we will generate more and more data. It’s obvious that we're going to need to deal with this issue indefinitely into the future.
For corporations, the solution is to get in front of their information and embrace the available technologies, which also have grown by leaps and bounds. It is not just about data but also about the need to have the right tools and expertise to tackle this head-on. Big data is not necessarily scary, but it has made people nervous. If you think positively and with proper guidance, it’s easier to plan the path towards taming and then leveraging data, first by making sense of it and then by utilizing it to your benefit. Business intelligence is a powerful tool, but it requires a determined mindset and, again, the right tools and expertise. And this is exactly where AlixPartners can help.
Published June 3, 2015.