One general ideal of analytics is to drill down further and further into data while still maintaining a large sample size in order to gain meaningful insight into that data. When it comes to legal spend, data at the task level is more informative than data at the case level which is more informative than data at the law firm level, and so on. Therefore, the biggest benefit of the UTBMS framework is phase coding. With phase coding, line item entries of an invoice can categorized into one of five phases: case assessment, pre-trial pleadings/motions, discovery, trial, and appeal. The ability to break matters down into these subsets provides significant value in a few key areas.
Phase coding is very helpful for budgeting. Phase-based budgeting gives a more granular and informed look at where money is going and where it should go. Perhaps a law firm has had a recent stint of poor outcomes, and the phase data reveals the firm’s trial prep/trial spend has been trending downward. The client may request that the firm spend more time in trial prep/trial. So, once a sufficient amount of phase data is accumulated, the client can set budgets for each phase of a case. Notably, this can greatly assist with keeping law firms on track duringthe case. Without phase data, a client might be watching a case with the expectation that spend approaches half the budget near where the halfway point of the case. Now, the client can know that as discovery nears completion, a firm should have billed 70% of their budget.
Phase code analysis also allows clients to forecast future spend. A client can estimate total case costs on individual matters and even across entire dockets without phase-level data, but doing so at the phase level gives a client more insight and control for managing spend, which allows for more accurate predictions. For instance, let’s say phase data has revealed that a recent change in the law has led to longer, more expensive discovery. The client can now discover that wrinkle more quickly, and then the client can increase their anticipated budget based on future caseload. Armed with phase data, the client might also be able to assess how law firms should reduce spend in other phases to remain on budget.
Finally, analyzing legal spend data at the phase level benefits benchmarking. Including phases when comparing law firms makes the comparison more robust and more instructive to the client and the law firm. If two law firms have similar total case costs (on similar matters), phase analysis may reveal that Law Firm B spends extremely high amounts in the initial stages of its matters (confirmed by a larger comparison to other firms). Curtailing this spend would benefit the client by saving money to which the client would not otherwise have had access. Furthermore, it could benefit the law firm, if privy to the information, by lowering their rates/case costs to be more competitive in the marketplace.
So, the UTBMS phase framework has had the potential to be a powerful tool in legal data analytics, but that is based on the assumption that the data is reliable. In a previous article, I discussed how UTBMS codes are flawed. In short, the reliance on time keepers applying the task codes often leads to inaccurate data. Luckily, there is at least one other method for effectively capturing reportable data: phase automation using keyword search algorithms. In this model, an algorithm scans line-item entries for specific keywords and applies phase codes automatically. While they might work for some endeavors, there have been concerns that invoice line descriptions do not lend themselves to such automation—that legal work is too complex and has too many variables. These concerns are well founded and lead to two fundamental questions: How difficult will it be to develop an effective automation, and will the phase data be accurate enough?
I’ll address these questions in detail in the next piece of this series. But to summarize, while phase coding algorithms may not be perfect, they are a far cry better than relying on busy time keepers to repeatedly apply correct codes after interpreting a set of rules riddled with gray areas. Phase coding algorithms produces immediate results, reduce the burden on law firms and time keepers, and produce results that are consistent. When designed correctly, keyword algorithms can result in a data set that is accurate, reliable, and actionable.