Trexin brings analytics to the front lines

George Smirnoff speaks about being reassigned to Trexin Consulting, a new management and business technology consultancy. George explained his history with Tableau and why Tableau is a critical tool for his new company. “It’s really about exploratory analysis, raucous analysis, problem-solving analysis – that’s where Tableau is head and shoulders above everybody else.”

It’s really about exploratory analysis, raucous analysis, problem-solving analysis – that’s where Tableau is head and shoulders above everybody else.

Tableau: How are you using Tableau at Trexin?
George: We focus heavily on strategy and analytics, and we use Tableau in most of our engagements.
Tableau: And why did you choose Tableau as your primary tool for the new company? Did you use it before, in your work at Ernst & Young?
George: Yes, I’ve been using Tableau since version 1.2 or 1.3. I’ve done a lot of work in trade data analysis on Wall Street and a lot of work with the SEC. Trade data analysis involves millions, sometimes billions, of transactions, depending on your time period, and compliance is a big deal with FINRA and the SEC. In the New York Stock Exchange, sometimes sub-exchanges are actually where the biggest things happen, so the visualization got introduced to be able to spot anomalies, whether they were the types of errors that would be from trade ahead violations or different types of rules and groupings that would get dropped. If you trace trades throughout the different trading desks as they go through their whole lives of executions and orders, you can visually see everything happen. We used it for compliance. So that’s what really brought Tableau into us. We did that at Ernst & Young, and since then we’ve published an article on it and we’ve done a lot. We’ve grown it, and we use Tableau in almost everything now. 
Tableau: Are you typically working with one source or a lot of different sources when pulling information?
George: We pull from multiple sources because, especially an example like trade data, it comes from many different trading desks. It can originate from many different areas, but it usually needs to be consolidated for analytics. Very rarely is Tableau the only tool used. It’s really used for the demonstration and the testing and the understanding of data. We use it for exploratory purposes.
Tableau: When you first started using Tableau, what were the things that surprised or impressed you?
George: Tableau is really an exploratory tool, so the biggest surprise was just the ease of use. You can quickly pull up the visualizations and drill down, and if something doesn’t work out you just undo it and redo it. We use calculative fields pretty heavily. We also use multiple tables quite a bit. You could write a command that could do the same work, but with Tableau you can do it much faster and just shortcut it. You can see patterns and trends that you want to drill down on much easier, especially when you’re looking at thousands and thousands of lines of data. You can pick out that one thing that really makes a difference.
Tableau: Have you gotten feedback about Tableau from your clients?
George: Pretty much every place we’ve taken it, the client ends up becoming a Tableau customer. The last client I worked with actually had Tableau, but it was a rogue installation by an IT guy who managed to get approval to buy it, and he was using it for his own purpose. We’ll take it from the deep depths of IT, and we’ll use it with the senior executives and do an interactive session.
Tableau: How has your history of working with Tableau influenced your vision for Trexin Consulting?
George: In setting up the new company, we have a concept we use “analytics to action,” which is really picking up from the “Competing on Analytics” article that came out in January of 2006 in the Harvard Business Review. The article demonstrated how companies can use analytics for competitive advantage, but what’s really been happening out there is that the business intelligence movement has created big consolidations of data with very little ROI. Companies are not really using the data in day-to-day business. The idea of analytics to action – the whole reason that we’ve set up this area – is so that we can push analytics out to the front lines of business. It’s really about business users. Getting out to the front end, you make it exploratory. You put the data at their fingertips.

Tableau is one of the key pieces of software that we push out there. We’re going through our clients’ processes, optimizing the way that they use their data, and giving their users the power of the tools. It might be in compliance and risk analytics. We’ve also done a lot with customer analytics and then the executives use it quite a bit. For people that become super users, Tableau becomes ubiquitous with Excel. They use Excel for stuff, but very rarely will they only use Excel. They’ll pick up Tableau. I’ve even seen people analyzing their project plans in Tableau because once you get used to it, it’s such a fast way to bring up everything.

Now there’s another whole series in the Harvard Business Review about making strategy real. The top two reasons that strategies don’t become real are information barriers and decision control. Tableau is all about breaking down information barriers, bringing in multiple data sources, and putting the analytics in the hands of the people that are actually making the decisions. Those are keys to strategy execution.

Tableau: How about an example of how you’re using Tableau in this way?
George: The main example is raucous analysis. It’s a great example because it brings in multidisciplinary teams. Very rarely can an IT guy understand why some coding thing isn’t working or why some data is getting lost. It’s all about the business use. So, we use the visual analytics with these multidisciplinary teams, where you have the developers, the data owner, and often the business process owner. You’re trying to find anomalies and understand why they’re happening. Rarely does one person really know, because they don’t know all the business imperatives that led to that error.

Raucous analysis becomes a very interactive session. I’ve done it with litigations support where I’ve had attorneys in there. I’m an attorney too, so I know these guys don’t know what’s going on, but we’ll bring it on the screen. We have a big conference room, and we’ll be bringing up the trade data. It’s usually about SEC litigation or something in that regard, and we will literally show them what’s happening in the data. They’ll say “oh yeah, that’s cool!” and they’ll all act like this is normal that they can literally drill down in real time and understand why an area of trade is happening the way it is, why they see this clustering. Then you have someone else in the room that says, “oh, that’s when this got caught in the outage” or “that’s why we had all those volume issues” and they’ll pull it together. Another thing might be, “well, that trader is a specialist – he’s also making his book.” Which means, he’s not just trading on behalf of the customer, but also on behalf of his company at that time.

The examples are really endless, but it becomes a very interactive tool. It’s so powerful because when you’re dealing with areas like litigations, you have various executives in the room. When they see what visualization can do, then all of a sudden they’re into “let’s get this more into the front lines of business.” It takes those kind of examples to get them to buy in. Otherwise the data is just reports to them, things that are static, and that’s not really what Tableau is all about – at least not the way we use it.

Tableau: So it’s not just an analytics tool – it’s a problem solver.
George: That’s what we use it for. In fact, the static reports – most static reports and dashboards or anything you get out of a warehouse in all these great business intelligence tools – have a declining value. When they first come out, they’re very insightful, but then over time, people get used to them. The data becomes more stagnant and they tend to look past them. There are lots of tools out there that quite frankly do great graphics. If you take the time and write the code, you can create great dashboards and make things pretty in a number of products. What you can’t do is interactively explore and break down the data. It’s really about exploratory analysis, raucous analysis, problem-solving analysis – that’s where Tableau is head and shoulders above everybody else. The other tools just don’t do it as easily or as quickly, and make it harder to bring in to the business users.
Tableau: Tell me about your first “aha” moment with Tableau.
George: I think I was the first one to reach out from Ernst & Young. I downloaded the demo, and when I first pulled it in, it was looking at trade data for a compliance project. The intuitiveness of the tool just made sense. Early in my career, I built analytics tools and one of the problems that we always had was how to solve quickly moving data, quickly visualizing it and then saving that work in your work books. Usually it required complex equal statements to get it done. Tableau usually guesses right when you just drag the data across, and that’s when it’s really exciting. You see it happen so quickly.
Tableau: We’ve talked a lot about how you use Tableau to find anomalies in trade data. Tell me about some other ways you’re helping clients with the software.
George: We talk about compliance in trading, but the same principles apply in multiple industries. We have used it for customer data. We’ve also used it for claims data in medical environments. You have heavy compliance in medical claims coding. Tableau is extremely powerful there because there you have to group different types of medical claims under different codes and procedures, and then are things that require overwrites and different subcodes. Many times we find providers don’t get paid because of coding errors.

It’s all about compliance within the data. In financial services, that keeps people out of jail, for the most part. In medical claims, it’s money that hits the bottom line every time you find those anomalies. Usually when something got spit out for not following the proper rules and the proper coding, it didn’t get paid. So every time you find that when you’re dealing with a health care provider, they end up resubmitting it and making money.