Editor’s Note: this is the first part in a three-part series looking at issues related to peer selection. Part II presented evidence that peer sets should be updated regularly. Part III looked at the drivers of profit per equity partner.
Selecting appropriate peers is a critical exercise for any law firm. Peer sets are used for benchmarking exercises, for annual rate-setting, and in understanding how wider trends in the industry are impacting the firm’s specific corner of the market. Often, lawyer and staff headcount decisions take into account peer ratios. Many law firms perform analyses to better understand differences with peers. Strategic growth goals are influenced by performance levels of peer firms.
Law firms typically choose peers using a manual process. Some firms look for peers with similar geographic focus. Other firms select peers based on perceived practice area strengths, faced in competitive situations or serving common set of clients. While this thoughtful exercise is useful and has served many firms well, it has some shortcomings.
The most obvious downside of law firms’ current peer selection methodology is that it is, all too often, subjective. In many cases, firms don’t specify detailed numerical criteria when identifying peers. In the absence of hard data, opinions seem to dominate the discussion. Many law firm CFOs and COOs tells stories of how senior partners insist that a particular firm be classified as a peer since they were competing for client business. Others law firm leaders add aspirational firms to their peer set, simply because their partners believe their firm ‘should be’ competing with such firms. The downsides to these approaches are obvious. Benchmarking exercises are meant to compare a firm to an appropriate group of competitors, based on financial or operational considerations. If the selected peer set does not accurately reflect the firm’s true financial position, the outcomes of the benchmarking expertise will, likely be misleading. As this information is typically meant to inform strategic and operational decisions, such an outcome could be costly for the firm.
A New Approach to Peer Selection
There is clearly a need for an objective, data-driven method to identify law firm peers. But how can this be done? A good place to start is to find firms of similar revenue. Unfortunately, this does not always work well. Consider Quinn Emmanuel, Davis Polk and K&L Gates. Each earned approximately $1.2 billion in revenue last year. That, however, is about all these firms have in common. As Figure 1 shows, each firm has drastically different profit margin, leverage, and geographic focus.
Clearly a more robust methodology is needed. The challenge is to add other appropriate metrics to revenue; ideally these metrics should be independent of each other. This is problematic as many law firm metrics are strongly correlated. For example, revenue and total lawyers are strongly related – firms with more lawyers typically have higher revenue. Similarly, revenue per lawyer and profit per lawyer are interrelated. The strong relationship between these metrics creates a problem. If both are included into the peer selection methodology it is likely to overweight the underlying shared characteristics.
After significant analysis, the authors of this article would recommend a methodology which combines revenue with two additional reasonably non-correlated metrics – revenue per lawyer and profit margin. This methodology of incorporating all three dimensions to find financial peers appears to work for most firms. Profit per equity partner (PPP) as a factor to identify peers was purposefully left out. Why? For two reasons. Firstly, PPP reflects how a firm’s partnership is internally organized. Secondly, PPP is not a component of a firm’s financial model – it is the result of the model. The authors of this article felt that peer groups should be based on these three fundamental building blocks of a firm’s financial model.
Put simply, this approach identifies law firms which are closest to each other on all three metrics simultaneously when mapped in three dimensions. To illustrate, let’s look at the results for Cooley, a mid-sized Palo Alto based law firm. The top ten peers which the methodology identifies are fairly similar to Cooley over a broad range of metrics – from lawyer headcount to more financially oriented metrics such as revenue per lawyer and profit per lawyer (see Figure 2).
Interestingly the methodology, which does not incorporate any geographical criteria, identified firms with geographic footprints similar to Cooley. This makes sense –firms with similar geographic footprints generally have similar financial performance. In Cooley’s case, all of the identified firms have an international presence, but a fairly limited one, with less than 25% of their lawyers outside of the US (see Figure 3). Their domestic presence is also similar. All of the firms, with the exception of Debevoise, have a broad footprint in the United States. Most have offices in two or more geographic regions across the United States and a significant or mid-sized presence in New York City, the Mid-Atlantic, the West Coast, and in New England. Cooley’s geographic focus, while not exactly the same, is fairly close to the average of the selected peer group (see Figure 3). This suggests the model is identifying firms fairly similar to Cooley.
No Model Is Perfect
Does the model perform equally well for every law firm? The simple answer is no. Not every firm within the Am Law 200 has a large set of financially equivalent peers. Some firms are outliers. Let’s look at Baker Mckenzie, which is ranked highly on revenue but among the bottom twenty percent of Am Law 200 firms in revenue per lawyer. Some firms share these characteristics, but not many. The model identifies ten firms which could be considered peers to Baker Mckenzie (see Figure 4). The first potential peer, DLA Piper, is a good match. DLA Piper is similar to Baker Mckenzie in size, revenue per lawyer, profit margin, and even in profit per equity partner. The next set of firms fall in two categories, both of which are not perfect peers. Large global firms like Hogan Lovells, Sidley Austin and Norton Rose Fulbright have relatively lower revenue per lawyer and profitability, rather similar to Baker Mckenzie, but are significantly smaller in revenue, headcount, and geographic coverage. The second group includes firms like Latham, Skadden and Kirkland, which have similar revenue but significantly higher revenue per lawyer and profit margins.
While the model produces robust results for most law firms which are clustered together, there are some notable outliers like Baker Mckenzie. Quinn Emmanuel is another example. Apart from Gibson Dunn, the firm’s high RPL and profit margins make finding other peers quite difficult. Sullivan and Cromwell has a relatively high revenue per lawyer for a firm of its size and profitability. Cadwalader’s low profitability is out of line with its size and RPL. Latham & Watkins is also a tough case. Beyond the two natural peers, Kirkland and Skadden, which may be intuitive, other Big Law firms have the same size but lower profitability; or lower size and comparable profitability. For these reasons the model presents a wide range of potential peers.
What should we make of the difficulties the model faces with these outlier firms? Two things should be kept in mind about a data-based approach. First, applying the algorithm uniformly does typically produce better results than the manual methodology that most firms would be using. Manually selecting peers through a subjective process is very likely to have biases. The model struggles with firms like Baker Mckenzie, Quinn Emmanuel, and Latham & Watkins since these are outlier firms. Finding a relevant group of peers for these firms is extremely difficult. It is not clear a subjective process could do better.
The second thing to keep mind about the quantitative approach the authors are suggesting is that it is meant to inform – not dictate – the process of selecting a peer set. The manual selection process which most law firms are currently using can be enriched with the inclusion of data and algorithms. The model is meant to provide guidance and help motivate a deeper data-driven discussion. Opinions and other subjective information can be supplemented with objective criteria to justify the inclusion or exclusion of a firm in a peer set.
The authors believe this new, data-driven methodology to identify peers could be of value to many law firms, to banks, to industry publications, and, potentially, even to purchasers of legal services. It can provide valuable insights and complements subjective identification. The next article in this series will explore how peer sets change over time and what those changes mean for firms.
Madhav Srinivasan is the Chief Financial Officer at Hunton Andrews Kurth LLP, leading the global finance and pricing competencies. Madhav is also an adjunct faculty at Columbia Law School in New York and University of Texas at Austin School of Law.
More information on the ALM Intelligence Fellows Program can be found here.