Disco - 10.5.2022

Disco sells e-discovery software for use in litigation.  In the discovery phase of litigation, both sides dump huge amounts of documents on each other, and it’s the responsibility of each side to sift through all the documents, the vast majority of which are useless, and find the few key ones that would make the case (either for trial, or, much more often, for settlement).  The volume of data involved constantly grows (now including sources like Slack), and software becomes essential for sorting through it.  This is also where huge legal bills come in – each side has teams of attorneys reviewing documents for many months.

Disco was founded in 2013 and is creating beautiful, fast, cloud-based software for the modern age.  Disco’s main competitor and the incumbent in the e-discovery space, Relativity, was founded in 2001 and has software that is universally hated by litigation attorneys.  Relativity e-discovery software is non-performant in a way that is almost hard for us to imagine – it’s extremely slow (spinning icon comes up when trying to open or change documents), and difficult to use to a degree that lawyers require an in-house Relativity specialist to conduct queries on their behalf.  It’s not an exaggeration to say that Relativity is like Windows 95 and Disco is like Google.

Lawyer feedback for Disco’s product is exceptional.  We’ve spoken to lots of lawyers about it, both formally through experts and through surveying our litigation attorney friends.  The UI is beautiful, document loading is extremely fast, conducting searches is intuitive and fast, and the product contains AI-assist features that, for example, highlight documents with the highest potential for relevance.  Using Disco is so intuitive for lawyers that no internal administrator is needed – they can conduct queries themselves instead of sending out requests to a specialist and waiting hours for a response.  Disco reports that its NPS is 63.

Furthermore, Disco’s superior product leads to cost savings.  Disco saves attorney time by allowing them to get through documents faster and find the right ones faster; it also saves direct cost by being much simpler to run and thus removing the necessity of having a Relativity administrator (internal and external). 

Given these dynamics, Disco is rapidly gaining share in the e-discovery software market.  Revenue grew 41% in 2020 and 67% in 2021, before slowing this year to 17% (which we will discuss below).  As is standard in the industry, Disco prices its software on a dollars per GB hosted basis, with everything else included in that price.  A lot of data is dumped into Disco in the discovery phase of each matter, where it sits (and generates fees for Disco) until ultimately the matter is concluded and the data is removed from Disco and hosted in cheaper cold storage.  This pricing model means that Disco generates recurring revenue and benefits from the explosive growth of data per case.

Legaltech has traditionally been a troubled industry, for two main reasons: 1) legal departments are cost centers in corporations and are not given much budget; and 2) selling to law firms instead has a higher rate of success but is not a large market.  E-discovery is a rare exception that sidesteps these two problems.  Most often, it is the law firm choosing the e-discovery software, but the cost is billed to clients as a pass-through, so there is less incentive for the law firm to put pressure on price.  On the client (corporate) side, litigation is treated seriously and usually given its own budget when cases come up – i.e., litigation is a separate budget from the usually constrained legal department budget.

In theory, switching costs are low, which benefits Disco as a challenger vs. the established Relativity.  It is extremely difficult to switch e-discovery software mid-case, but for each new case, a new vendor can be swapped in with little trouble.  It’s not uncommon for a law firm to simultaneously have both Relativity and Disco running in different cases, with the decision on which to use for each case being decided by the preferences of the partner leading the case or, in some cases, the client.

Given Disco’s superiority on multiple dimensions, our conversations with lawyers suggest that the vast majority of legal cases would be better run on Disco than on Relativity.  The exceptions are for highly complex matters where rare features of Relativity would be used (e.g. where unusual foreign languages are involved, or involving corner case rare document types).  Nonetheless, Relativity still dominates the market.  The switching costs in practice are: 1) inertia, especially among older lawyers that don’t handle the document review themselves and don’t want to change their process; 2) limited motivation for some lawyers to switch, given that e-discovery cost and attorney time wastage are a pass-through to clients; 3) hand-holding that Relativity and its services ecosystem provides in the event of corner cases that might arise; 4) relationships that Relativity have built with law firms, through two decades of wining and dining key decisionmakers. 

Disco has been and should continue to overcome these barriers.  For #2, cost insensitivity for law firms is an interesting one that is not as insurmountable as it may seem.  With the explosion in discovery data volume and legal bills, clients are becoming more discerning about managing legal bills and controlling cost.  A law firm pitching for business with a lower pricing/cost structure has an advantage.  For #3, lawyers who have worked with Disco say that Relativity’s handholding is only of minimal value for most cases, and that Disco’s product is steadily iterating and adding more features.  Factors #1 and #4 are just variations of inertia. 

In addition to selling to law firms, Disco is also selling directly to corporate clients, educating them on Disco’s cost saving advantages and getting them to sign deals such that they agree to have their law firms use Disco on their matters.  In its prospectus, Disco lists Accenture, DISH, Lyft, Peloton, Southwest Airlines, WeWork, and others as corporate customers.

Valuation

Disco currently trades at 2.6x NTM revenue, down from a peak of 34x NTM revenue shortly after IPO in late 2021.  I estimate that Disco can grow revenues in the mid-to-low 20s for a long time to come, based on rough estimates of its industry penetration in terms of % of GBs hosted.  In terms of margins, this is a 74%+ gross margin software business where revenues (driven by GBs hosted) and costs ex-S&M are fairly disconnected; ultimately this should be able to get to 20%s type profitability.  Thus, I believe an exit multiple in the 4-5x range would be warranted, which would add to the IRR.

What happened in second quarter results?

Disco had disastrous second quarter results.  2Q22 revenues met the high end of guidance, but updated full year guidance implied a major cut to back-half numbers (with Q3 now +11% y/y, Q4 -3% y/y).  The company blamed this on the underperformance of its lumpy services business (where it does outsourced document review using its team of in-house attorneys), the magnitude of which it disclosed for the first time (~$100 mm e-discovery revenue, vs. $34 mm other services revenue in 2022).  Management said that e-discovery revenue alone was expected to grow 30% in 2022 vs. overall revenue growth of 17%, which was dragged by services.

The earnings call was strange and not well handled by the company.  Founder/CEO Kiwi Camara (who is a brilliant but controversial character for some statements he made while a teenager in law school – look him up) answered robotically and repetitively to some earnest analyst questions, and the CFO did not speak in the Q&A section at all.  Despite the decline in forecasted revenue, the company would not alter its spending plans, and thus Adj. EBITDA was expected to blow out to -57% margin in the back half, up from negative single digit margins at its highest point (Q1 and Q2 2021).  The company has a large cash balance ($228 mm) relative to the loss, so liquidity or solvency is not the issue, but the magnitude of the negative margin was nonetheless alarming.  The stock fell 50%. 

Nonetheless, checks with customers we did after this call reaffirmed the clear superiority of its product and its potential in the market.  Conversations with former employees reinforced that this management team still needed to “grow up,” especially in the context of handling the public markets.  But Kiwi is smart and by all accounts appears to be a learning machine.  We think this is a hiccup that doesn’t change the long-term potential of the company, and are excited by its current low valuation which we believe compensates for the risks.