Why Clients Are Insourcing Legal Work

In-house counsel are increasingly doing work that outside counsel used to do. According to Altman Weil, about 80% of large law firms have reported losing business to in-house counsel for the past three years.

We can make sense of this “insourcing” trend in many ways. Hugh Simons, in a post at Legaltech News, attributes it to a not-so-great growth strategy on the part of Big Law, plus a simple economic calculus on the part of clients:

…in response to firms overshooting the market’s needs in terms of the [high]sophistication of the services offered and their [commensurately high]price point…, clients have grown frustrated … and have slowly, steadily and irreparably taken legal services in house…

Insourcing makes sense only if in-house counsel save money, which they do, and the reason for this is essentially better technological literacy driven by in-house counsel’s inherently different incentives:

While law firms’ sluggishness in new technology adoption has been ascribed to traditional conservatism and low technological literacy, the true root cause is more basic: lawyers are measured on their billed hours; technology reduces billed hours…. In-house attorneys have the obverse incentives. They are pressured relentlessly to do more with less. Technology helps them do so.

Richard Susskind, in Tomorrow’s Lawyers: An Introduction to Your Future, views the trend toward insourcing less as a deliberate act by in-house counsel who fully understand their own economic interests, and more as an obstacle (the “more-for-less challenge”) created by external factors:

General Counsel … say that they face three problems. First of all, because of difficult economic conditions, they are under pressure to reduce the number of lawyers in their teams. Second, they are being asked by their chief executives, chief finance officers, and boards to reduce the amount they spend on external law firms. … [T]hird, they say they have more legal and compliance work to undertake than ever before; and that the work is riskier too.

Another way to make sense of the trend is that, for certain kinds of work, insourcing inherently increases efficiency. This phenomenon isn’t limited to the legal industry. It’s one of the factors behind a recent trend among U.S. software developers to do more development in the U.S.

Three perspectives from different decades and industries help explain how simply doing work yourself can help you do it more efficiently.

Parkinson’s Law of Multiplication of Work

Cyril Northcote Parkinson observed that the staff of the UK Colonial Office increased from 372 people in 1932 to 1,661 people in 1954, even though there weren’t a lot of colonies to administer by the 1950s. What were all of these people doing? How did this happen?

Parkinson’s Law of Multiplication of Work depends on a cynical but probably accurate observation about organizational behavior that Parkinson dubs the Law of Multiplication of Subordinates: if you decide you need a subordinate, you’d better get at least two. Otherwise you risk your subordinate taking your job, stealing your next promotion, and so on.

The result of lots of subordinates is lots of make-work, and Parkinson dubs this the Law of Multiplication of Work. Parkinson illustrates this “law”’ using seven government officials named with the letters A and C to H:

…these seven make so much work for each other that all are fully occupied and A is actually working harder than ever. An incoming document may well come before each of them in turn. Official E decides that it falls within the province of F, who places a draft reply before C, who amends it drastically before consulting D, who asks G to deal with it. But G goes on leave at this point, handing the file over to H, who drafts a minute, which is signed by D and returned to C, who revises his draft accordingly and lays the new version before A. … [Official A] finally produces the same reply he would have written if officials C to H had never been born. Far more people have taken far longer to produce the same result. No one has been idle. All have done their best.

If you lost track of who was doing what, that’s the point. Everyone was doing unnecessary make-work. Official A simply gluing himself to his chair and getting the document done probably would have saved several days’ worth of time.

The Mythical Man-Month

Fred Brooks was overseeing development of an IBM mainframe operating system in the 1960s, and it was behind schedule. Brooks added programmers to the project, but this didn’t help. In “The Mythical Man-Month,” Brooks concludes that adding programmers had actually put the project further behind schedule. This seems counterintuitive, as many hands make light work, right? What happened?

As Brooks explains, for some kinds of work (having a baby, for example), you can assign all the people in the world to it, and it’s still going to take nine months. For these kinds of work, m = 9 (m is months to completion):


Other kinds of work (purely manual labor, for example) can be split evenly among any number of people: 1 person can do it in 9 months, 2 people in 4.5 months, 3 people in 3 months, and so on. In other words, months and people are interchangeable. (You might try to measure the amount of work to be done in man-months.) For this kind of work, m = 9 / p (p is number of people):


In real life, practically no work can be split absolutely evenly (thus the man-month is just a myth). For most kinds of work, people must be brought up to speed, trained, and so on, and this takes time. Say this training takes 0.2 months per person. For this kind of work, m = 9 / p + 0.2p:


In information-heavy work, communication is essential. Communication also takes time, and the more people who need to communicate, the longer it takes. Three people means three lines of communication; four people means six lines, and so on.

Put another way, the number of lines of communication c is p (p – 1) / 2.

For this kind of work, m = 9 / p + c:


The reasoning here is practically the same as the reasoning behind Parkinson’s Law: in the course of doing work, people communicate, communication takes time, and the more lines of communication you have, the more time you spend communicating instead of working. (And if you have enough people, you may actually spend more time communicating than working.)


Reengineering—a process improvement paradigm that was that was popularized by Hammer & Champy’s “Reengineering the Corporation” in the 1990s—takes a critical view of disaggregating/decomposing/unbundling work:

Inflexibility, unresponsiveness, the absence of customer focus, an obsession with activity rather than result, bureaucratic paralysis, lack of innovation, high overhead—these are the legacies of [Adam Smith–style production processes].

“Reengineering” includes several case studies that illustrate the benefits of reengineered processes, but IBM’s financing department sticks out. IBM Credit reduced its turnaround time on financing customer purchases from seven days to four hours—a 97.6% reduction. To do this, IBM Credit realized that the information needed to approve a financing request had been spread among a whole department of people in a six-step process, and a small group with the right information could approve a financing request much more quickly. What proportion of the 97.6% reduction in turnaround time was attributable to overall process improvement and better information flow versus just having fewer people involved is open to debate, but without doubt, some of the seven wasted days were spent on communication-oriented make-work: getting people up to speed, following up on approvals, editing memos to your boss, and so on.

The underlying reasoning of reengineering is practically identical to Parkinson’s Law and the mythical man-month: involving more people in a task can increase the amount of time it takes.

Applied to insourcing legal work, delegating work to anyone (outside counsel included) means involving more people. Involving more people means more communication: checking in, following up, circling back, and so on. These activities take time (and are much more difficult to manage if the people to whom you delegate are incentivized to maximize the amount of time these activities take). And if the real-life behavior of in-house counsel is any indicator, it’s often easier, faster, and cheaper to just do the work yourself.

About the Author

Benjamin Whetsell is the co-founder of Paper Software, and previously was an associate at Fried Frank in New York. Contact him at bwhetsell@papersoftware.com.

Send this to a friend