The official 2016 ILTACON conference theme may have been “Embracing Change,” but the hot topic at the conference was data.
Two keynote presentations set the data-driven tone, starting with author and futurist Mike Walsh’s opening keynote on “Re-Imagining Legal Technology for the 21st Century”, in which he described how data is being used in real time by entities such as Walt Disney World to deliver customized, predictive services and products to visitors. He anticipates a future driven by in-the-moment delivery of information that influences behavior in real time, in which law firm clients demand real-time data and visualization tools that augment their decision-making process around legal services and issues.
Chicago Kent College of Law Professor Dan Katz’s keynote on “Solving the Legal Profession’s Biggest Problems” positioned data as an essential to the success of legal practice going forward, stressing that a lawyer or enterprise must capture, clean, and regularize data in order to then offer the predictive services (substantive, procedural, risk management) that clients will expect and demand to receive in the next decade. Other sessions applied data more specifically to legal knowledge management endeavors, including:
- A Picture is Worth More than a Thousand Words: Data Visualization Best Practices (moderated by David Hobbie, KM’er and ILTACON 2016 Team Coordinator, Information Management),
- Data Analytics You Can Do Now in Legal,
- Using the Right Data to Drive Your KM Program,
- New Tools for Old Data: Computational Linguistics and the Practice of Law, and
- Data Mining: Leveraging Information to Make Strategic Decisions
as well as several others.
The range of discussion points included the practical nuts and bolts of using data in law firms and legal departments: what data to capture, tips on how to represent data visually, and some “show-and-tell” examples of data dashboards and analytics. Other sessions followed on the themes of the keynote presentations, such as current and potential use of artificial intelligence and expert systems in legal and predictive analytics. Conference attendees could not help but receive the message: Data is important, it is the future, and law firms need to capitalize on it. Less clear, however, was exactly how to implement the data directive, what challenges a firm looking to build a data culture may face, and how to overcome them.
Winning with Data
A slim book published in May, titled Winning with Data, fills some of these gaps. The book addresses many of the challenges organizations face as they evolve from data silos to data-driven enterprises that use data to operationalize their business processes. At less than 200 pages, it is a quick read that cites real-world examples of how companies like Facebook and Zendesk are cultivating a data culture that empowers everyone in the organization to access data and perform their own analyses. Authors Tomasz Tunguz and Frank Bien speak from their own experiences, the former having worked at Google and now a partner at a VC fund (and prolific blogger), and the latter an industry veteran with more than 20 years’ experience working with various database and BI companies. (Disclosure: My employer has had, and does have, relationships with certain of the authors’ employers and entities referenced in the book.)
Even though law firms and legal departments do not have the same business model as technology product and services-driven businesses, several observations and recommendations in the book about the progression of data-driven organizations should be relevant to most professional services organizations.
The authors identify several hurdles along the path to what they hold out as the ultimate goal of “operationalizing data” – meaning organizations using data in real time to enable immediate and predictive changes in operations. Uber is a clear example of a company that has operationalized data – unlike taxi drivers who had to guess where to find fares, Uber uses data to deliver customers to drivers in the most time-efficient way possible and even increase the fare when demand hits certain thresholds.
But organizations do not reach data maturity overnight. The journey to data maturity is difficult, and the authors quickly identify scenarios and stumbling blocks that are both recognizable and humorous.
- The “data breadline.” Much as there were lines of people waiting for free bread during the Great Depression, data-poor employees wait at the end of data breadlines. Unable to directly tap the source of the data, they are dependent on data gatekeepers who are overwhelmed with requests for data from all departments.
- Data obscurity. Once employees get to the front of the line, then they must be able to specify what data they need and where it is — using the same terminology as the data-gatekeeper. While routine requests may be well documented, the huge increase in the amount of data available to an enterprise, and the uses for that data, increases the potential for obscurity.
- Rogue databases and shadow data analysts. Rather than suffer standing in a data breadline, employees circumvent chains of command and find data other ways. Data obtained through personal relationships or favors is then quietly hoarded by the employee or team and used for analysis as a substitute for “official” data, with obvious risk for misinterpretation.
- The “bastardize an existing solution” phenomenon. While the authors use a different example, we might just call this SharePoint – or more accurately, the temptation to clone a SharePoint solution created for one team, and shoehorn different data into it.
So what are the building blocks of mature data teams and data-driven organizations? The authors cite three foundational elements.
The Building Blocks
The first building block is a culture of data literacy. Everyone within the organization needs to know what data is available, what it means, and how it is used internally. New Facebook employees attend a two-week data camp run by the Facebook data team. AvantCredit, a growing lending institution, requires all of its employees to attend a two-week seminar on the data infrastructure and data tools the company uses. Zendesk, a help desk customer service platform, has a 30-person data team that hosts weekly office hours for the 1,000 person organization.
Second, a functional data supply chain is needed. Eliminate, or at least shorten, the data breadline. Allow direct access to as much data as possible. Teach employees how, when and why to tap different data supply chains.
The third foundational element is a shared data language. Establish an institutional “Data Dictionary” that contains universal definitions of particular metrics within a company. A data dictionary ensures that everyone within the organization speaks the same data language. And perhaps more importantly, it requires that the organization clearly define its key performance indicators (KPIs) and the equations that define the business.
The book concludes with an appendix of revenue metrics, some of which are adaptable to legal, but which also provide concrete examples of what a data dictionary may look like.
Takeaways for Legal KM
While not all lessons from the start-up case studies provided in Winning with Data are easily transferred to law firms and legal departments, many are. In particular, the descriptions of human behavior around data seem particularly apt. Data access and supply chains likely need to be far more controlled in our organizations due to confidentiality concerns. But taken in context with the aspirational ILTACON sessions, the book does highlight a few subtle but potentially significant aspects of successful data-driven organizations.
- Develop a data dictionary for your firm or department — and make it widely available. Otherwise the “shadow data analyst” and KM may inadvertently be working from different numbers or with inconsistent formulas.
- Learn, teach and encourage data literacy within the organization. This may be challenging, given it is a complex skill set, but look for ways to create shortcuts for users by making things such as Excel templates with pre-populated cells available for budget tracking and projecting.
- Reporting dashboards are nice, but aim for operationalized data and that changes behavior. For example, by tracking and logging the types of requests a KM department receives from users from month to month or year to year, one may spot trends that will allow your department to predict needs and push the needed materials to users when they need it.
- Impose discipline on how data is defined when exploring out of-the-box BI dashboards and reporting. Does the proffered output address the equation that is most important to the business? Does the output inform your firm or legal department’s strategic decision making?
Be an Explorer
While the authors cite to Gartner’s Data Sophistication Journey, which diagrams the path from descriptive analytics (such as monthly budget reports) to prescriptive analytics (such as risk mitigation), they suggest that the diagram is missing an important step between the retrospective analysis (steps 1 and 2) and prescriptive analysis (steps 3 and 4). Exploratory analytics — using data to search for hypotheses — should occur at the halfway point of the ascending line. Exploratory data analysis doesn’t seek to prove or disprove a particular idea using data. Rather, it is the use of data to test a hypothesis for patterns observed in one’s business. Exploratory analytics allows one to see what is happening in real time, and the authors suggest it is the key to reaching the ultimate goal of operationalizing data.