Big data: too much information

Big data analytics have long been used in retail, advertising and financial services, to name but a few sectors where data scientists routinely analyse large volumes of anonymised unstructured data to identify trends and patterns, and forecast future market developments and consumer behaviour. The information is then used to guide strategic decision making and maximise profitability.

The main features of big data are the five Vs: volume, variety, velocity, veracity and visualisation.

In the legal sector, while regulatory compliance requires firms to retain and manage exponentially increasing amounts of data, online publishing has also brought lawyers more work in terms of cases around privacy and the right to be forgotten. Big data analytics are also being applied to the legislative system, dispute resolution and legal services provision.

Digitising the statute book

In 2014 big data moved centre stage, with the Arts and Humanities Research Council-funded Data for Law project, which was launched to facilitate socio-economic research and identify patterns in the way we legislate. This involves digitising the entire statute book and presenting the data so that all UK legislation can be analysed, together with publicly available data from legal publishers such as LexisNexis and Westlaw.

‘It’s better to count than to guess,’ observes David Howarth, a legal academic and former MP who co-leads the project. ‘Big Data for Law will provide the most comprehensive record of all UK legislation ever created, together with analytical tools. Although these will be most useful for the public sector, government, researchers and policy-makers, it is also useful to law firms, particularly those with public policy practices who will be able to gain insights into the changing regulatory environment.’

Firms and lawyers will be able to use the Big Data for Law resources to identify patterns and trends in particular areas of legislation. An important consideration is to publish the data in a form that will enable researchers to work with it effectively – and present their findings clearly.

Another part of the project uses big data as a design tool, to identify patterns within statutes – combinations of rules that are used repeatedly to meet policy goals. This is effectively extracting the structure of legislation and thinking about how it can be reused. Howarth suggests that law firms could also apply this approach to legal documents.

John Sheridan, head of legislation services for the National Archives and senior investigator for the Big Data for Law project, highlights the importance of creating tools and methods that are accessible to those without deep statistical knowledge as well as developing pre-packaged analysis that researchers and others can use and cite in documents: ‘For example, we can plot fluctuations in the number of laws made year on year, and in the length of the text.

‘We can discover how modular legislation is in particular areas and frequency of legislative change. We can also examine the language of law which uncovers the topics of the day and reflects major global events and political and economic trends.’

Mapping the statute book identifies commonly recurring themes and language patterns. Sheridan and his team have identified patterns for licensing, prohibition, regulators, tax and so on, with the purpose of enabling interested parties to quickly gain an understanding of a particular legislative development and identify commonly occurring solutions to particular issues.

The ability to search the entirety of UK legislation using time-based parameters and particular words and phrases makes it straightforward to find relevant laws pertaining to a specific issue, which is particularly useful to parties involved in litigation. Sheridan underlines the importance of visual presentation to plot patterns and trends, and the ability to drill down into the detail.

Making data ‘visual’

Big data analysis helps law firms to leverage internal and external data by turning it into information and insights into their own operations, clients and markets. Visual representation is the only clear and comprehensive way to present the outcomes of big data analysis. This has helped to make big data analytics practical and accessible. It has also contributed to the design and widespread takeup of business intelligence applications that use graphical representation and dashboard-style user interfaces. These enable law firm management, partners, fee-earners and business support functions to work in real time with complex data sets.  

It is, however, important to differentiate between big data analytics and business intelligence. Whereas business intelligence applications predominantly help firms leverage financial and management information, big data analytics uses algorithms to interrogate large volumes of unstructured, anonymised data to identify correlations, patterns and trends. The data, which can include transactional data held by the firm and data from external sources, can be cut in different ways – for instance, by geography, market, sector, product or any combination of these.

One of big data’s value propositions is the potential to uncover patterns – and opportunities – that are not immediately obvious. This requires human input: analysts interpret the outcomes of big data analytics to determine the context of trends and correlations.

Litigation technology

An obvious practical application of big data to legal practice is e-discovery. Here, big data analytics support machine learning and natural language processing to reduce the time and cost of trawling through massive volumes of structured and unstructured data held in different media and platforms, in order to extract relevant evidence.

Tracey Stretton, legal consultant and e-disclosure specialist at Kroll Ontrack, has seen more takeup of this technology in the UK as the volume of data involved in big corporate cases has grown. ‘Every case we work on now includes some form of predictive coding,’ she observes. UK firms focus on prioritisation, whereby, having learned from a human reader, the technology scans all the documents to bring the lawyers the most important ones first. ‘There is a lot of talk about using predictive coding to speed up the review process but the real advantage is strategic. Early identification of the most important documents helps lawyers build up the bigger picture about a case much earlier in the litigation process.’

The US is ahead of the curve when it comes to e-discovery. A current hot topic is the ability to apply algorithms to relevant financial data to extract information relating to corporate cases dealing with financial dealings. These complex data sets tend to be held in relational databases such as SAP. So when the figures are extracted they need to be presented in a way that highlights their relevance to the case. Here we are seeing the emergence of graphical visualisation tools to present financial evidence – because the numbers on their own don’t always tell the story. Mining company data and presenting it graphically helps law firms assess the risk involved in a case and decide how to handle it.

Big data analytics are increasingly being used to identify patterns and trends in dispute resolution to predict the probability of case outcomes. In the US, Lex Machina applies analytics to IP transactions and litigation bringing together case information, court records and other public data in order to predict case outcomes.

The legal sector is also following the financial services and retail industries in applying big data to day-to-day work. A key consideration here is that big data analytics must produce a return on investment for the firm, in relation to winning cases, expanding client services, boosting margins and driving competitive advantage.

US litigation firms were the first to work with big data. Eric Hunter, director of knowledge, innovation and technology strategies at Bradford & Barthel in San Diego, transformed the firm’s operations by moving to a consumer platform with a social-media style client interface to share case and billing information in real time. The firm’s spin-off consultancy, Spherical Models, mines client and case data to predict probable case outcomes. This facilitates decision-making – for example, whether to pursue a case or to settle – and shortens the litigation process, allowing the firm to handle more work and price competitively. Again, visualisation creates a user-friendly interface.

Cranking up the volume

In the UK, the most successful deployment of big data technology is in the volume sector, where a key driver is competition from alternative business structures offering online legal services that include self-service applications supported by automated workflow technology. This is becoming more sophisticated, with robotic process automation combining workflow technology with data analytics and artificial intelligence engines to create automated processes that include basic decision-making.  

Firms that handle volume work are investing heavily in big data analytics to facilitate decision-making and produce additional market-, sector- and case-specific intelligence to feed back to their clients. This enhances client services in two ways: the information can guide client services, and the client organisation gains extra insights into their own business and into their sector generally.

DWF uses big data analytics to offer the firm’s insurance clients a service identifying potentially fraudulent claims, using IBM i2 Analyst’s Notebook visual intelligence analysis software. As chief technology officer Richard Hodkinson explains, the software identifies connections between different entities to identify potentially fraudulent patterns of activity visualised on the screen.

DWF is also applying IBM Watson Content Analytics to volume case management. A complex algorithm will analyse the documents associated with each matter and give it a score between one and 10, which will indicate how much or little lawyer input the case is likely to need. If a matter scores 10, it will require a lawyer right away, and if it scores one it will be a routine case that will require minimal lawyer input. ‘At the moment, every case that comes in is reviewed by a lawyer,’ says Hodkinson.

‘The new system will review each case and suggest a potential route. The software doesn’t just look at fields and keywords; it reads the text and analyses words and phrases as well as context, and reaches an informed decision.’ There are clear advantages in terms of speed, efficiency and cost. Watson Content Analytics works with structured and unstructured data, meaning that it will look at databases as well as contextual information. Hodkinson adds that lawyers are happier spending their time on quality work rather than routine tasks.

Berrymans Lace Mawer (BLM), which handles volume insurance work, is the first law firm to invest in Board’s data analytics tool, which analyses clients’ data to produce management intelligence and insight as an additional service for its clients. IT director Abby Ewen explains that most large insurers have undergone multiple mergers and have numerous legacy systems. ‘Insurers are looking for an overview of their claims business,’ she explains. ‘They look to us to transform their transactional claims data into management information and insights.’ Transactional data that BLM reports back to clients includes numbers of cases handled by different areas, offices and people, and settlement times and amounts.

Board operates in real time, identifying patterns and trends and presenting business intelligence back in a visual way using heat maps and graphs. ‘The software sits over our data warehouse, so it has access to our case management, practice management and financial data,’ explains Ewen. ‘Clients simply log into the system online to access and analyse their own transactional data. Features include the ability to build models upwards from small details or drill down from broader parameters. Predictive modelling allows users to apply scenario planning to identify significant issues and support business decisions. It is about transforming data to insight.’

However, the process does require a human element to interpret what the data means for each organisation – BLM employs 26 people in its management information and business intelligence team.

‘In a non-volume law firm, lawyers’ use of technology is almost optional,’ observes Ewen. ‘But in the volume space, it is the differentiator. Firms compete for business on their ability to run volume cases efficiently and profitably, and provide clients with accurate management information, which is why everyone is chasing the holy grail of data insight and management information.

‘Of course some firms are further along the evolutionary curve than others,’ she adds. Next steps include anonymising and analysing all BLM’s transactional data to produce industry benchmark data and insights for its clients based on real-time analytics.

Other uses of big data analytics are predicated on data visualisation. Wright Hassall deploys C24’s Bi24 business intelligence and analytics platform to geo-demographic data, to support the firm’s debt recovery business. A geo-demographic heat map identifies which locations have a higher propensity to pay and the team concentrates its efforts on these areas, significantly improving its success rates. This also drives decision-making on how to proceed with the case.

‘There are compliance constraints around making the information actionable. Nonetheless, it is a useful resource,’ chief information officer Martyn Wells says. There are also compliance and confidentiality considerations associated with working with personal data that belongs to the firm’s financial services and other clients. ‘As the firm acts for numerous clients with a common experience and a common set of data that determines our success rate in a particular area, we work with anonymised data,’ explains Wells. The firm can then work with the anonymised data to establish a holistic view of UK payment profiles in different areas.

David Ricketts, head of marketing at C24, highlights another application of big data visualisation technology in the legal sector: competitor analysis and sentiment analysis – monitoring social media to identify positive and negative trends. This is where big data analysis becomes applicable to the major international law firms who use big data analysis to advise their global clients on IP and branding issues.

Most firms featured deploy data analysis software and services that are not supplied by legal IT vendors. A notable exception is Manzama’s cloud-based current awareness product: a software engine that collects and aggregates publicly available data from 70,000 news and business sources, and integrates this with firms’ own information resources, including subscription-based content from legal publishers such as LexisNexis and Westlaw. This provides firms with information about their clients and competitors.

As Miles McGoun, head of global sales and business development explains, users can add and remove parameters, adjusting the algorithm to reflect the changing interests of individuals and practice groups. Firms use Manzama to understand their clients’ businesses and the environment in which they operate, and increasingly to monitor their peers and competitors – trawling announcements for instructions, transactions and mergers. Results are presented dashboard-style in a variety of ways including a smartphone app.

Skilling up

Big data brings the legal sector opportunities to leverage the growing volume of data they handle by transforming it into information and insight that guides decision-making – around their own strategy and that of their clients. However, as in other sector, firms are recognising that big data analytics also require specific skills to interpret and present the findings and outcomes. They increasingly employ data analysts and statisticians.

There are also challenges. Combining big data analytics and artificial intelligence to handle routine tasks and decisions will free up lawyers for more challenging work. However, this would create issues around training – trainee lawyers have traditionally learned ‘on the job’ by taking on straightforward and routine cases before progressing to more complex work. Data privacy is also problematic – the data that law firms are analysing belongs to their clients, so they need to ensure that it is properly anonymised before being analysed and repurposed. There are clearly opportunities for legal IT vendors too.

Big data cannot find the answer to every problem, and it is certainly not applicable to every law firm or practice area. But it is this year’s disruptive technology, affecting legislation, policy and legal services delivery.

Joanna Goodman is a freelance journalist and legal IT specialist