Measuring Legal Workflow Improvements

A chamber adopts a new drafting tool and everyone reports that drafts come together faster. Six months on, the seniors are more tired than before, because every quick draft now needs heavier review to catch what the speed introduced. The first draft got faster; the matter did not. This is the trap in judging any change to how legal work is done: the improvement is visible where the change happens, but the cost often lands somewhere else, later, on someone else's desk. Measuring workflow honestly means following the work through the whole chain, not admiring the fast bit.

Pick an outcome the team actually feels

Vague goals like "be more efficient" cannot be measured and cannot be disproved, which makes them useless. Choose something concrete and observable that the team already notices:

• Time from instruction to a draft a senior is willing to review.
• Number of correction rounds before a document is filed.
• Time spent locating a cited page or source document.
• How complete a handover is when a matter changes hands.

Each of these is a real experience of friction, and each can be compared before and after a change with honest eyes.

Establish the baseline before you change anything

You cannot measure an improvement you never sized. Before adopting a new practice, capture how the current process performs on the chosen outcome, roughly is fine, using a handful of recent, comparable matters. Without this, every later assessment collapses into impression, and impressions favour whatever is new. A modest baseline drawn from real files beats a precise measurement of nothing.

Follow the work downstream

The central discipline is to look past the point of change. If a step got faster, ask what happened next. Did the saved time reappear as extra review? Did a cleaner intake reduce corrections at filing, or just move the same corrections later? Rework is like water; it does not vanish when you speed up one step, it flows to the lowest point in the process. Ask the people downstream where their effort went. They will tell you things the upstream metric hides.

Compare like with like

A change can look brilliant or disastrous depending on which matters you compare. A complex commercial dispute and a routine notice are not the same measurement, and comparing an easy new matter against a hard old one proves nothing. Hold the type of work constant. Look at similar documents, similar forums, similar complexity, before and after, so the difference you see is the practice, not the luck of the sample.

Ask people, not only the dashboard

Usage statistics tell you what happened, not whether it helped. A tool can show high adoption while quietly making the work worse, or low adoption while the few who use it well transform their output. Talk to the team. Where did the annoyance go? What now takes less thought, and what takes more? Where did a new kind of error appear? These conversations catch the effects no counter records, and they are where the real signal usually lives.

Change the process, then measure again

Measurement is not a verdict; it is a loop. Once you see where the work actually moved, adjust, and measure the same outcome again. A practice that improved intake but overloaded review needs the review step redesigned, not the intake abandoned. The point is never the number for its own sake. It is a truthful picture of where effort and error live now, so the next change is aimed at the real constraint rather than the one that was easiest to see.