As with a number of start-ups, there’s an anarchistic bent to GitHub – flirting as it does between open-source hosting, building a community among software engineers and generating profits. This nonconformity is given a playful outlet in its office reception which is a recreation of the Oval Office. Otherwise, the company’s spacious headquarters in San Francisco ticks a lot of the boxes for a Hollywood (California’s other iconic industry) imagining of a tech start-up.
Large play area with pool table – check; communal working – check; everyone in denim – check. When Avalos comes to greet me he’s clad in jeans, a casual shirt and a beanie. But beyond adopting their fashion sense, Avalos feels he has learned a lot from the software engineers who make up the bulk of GitHub’s staff. What he has learned has caused him to fundamentally question the way he delivers legal services, to begin to look through an engineer’s eyes and question macro assumptions around what being a company lawyer means.
Avalos first became immersed in the tech start-up world while based at a law firm. He did a lot of work for Facebook and noticed how the legal function shifted as the company grew. In common with a lot of promising associates in the Bay Area he soon left his firm to go in-house – to Yelp. Still strictly a start-up, Yelp was nevertheless formed along traditional company lines. In 2012 he made the leap to GitHub, a start-up in all senses of the word. Founded in 2008, it was self-funding until just before Avalos’ arrival, when it received its first round of outside investment.
In part, his appointment was made possible by this first cash injection. But as an owner-founded, engineer-centric business, the notion of a lawyer coming to join them was quite alien, as he explains. ‘GitHub was a paradigmatic start-up – no managers and no titles – in a way that all sub-50 employee companies naturally have some version of, but with which we were philosophically committed to experimenting and scaling.’
As the first lawyer, Avalos could have remained out on a limb and isolated. Instead he decided to go native: ‘I focused on the culture of the company, which emphasised humans and human endeavour, tackling problems in a logical, engineer-centric way, working from first principles, without assumptions, and with a constant questioning of “how do we build what works?’”
Going native
Getting to grips with his company’s distinct tech culture permeated the way Avalos thought about building GitHub’s legal team, he says. ‘Unlike the typical company structure where legal spills off from the rest of the company and is run more or less as a silo, it was really self-preservation for me to grasp the technology and to look, talk, and work like everyone else.’ He explains that, ‘ulimately, it was really about rethinking what it meant to be a lawyer within a technology start-up and viewing myself as a GitHub employee that happened to have legal skills but otherwise was no different or better than anyone else.’ As he became more aligned with his engineer co-workers, he began to think more like them, leading him to question and refine the first principles of the legal department.
With over nine million users, GitHub is the largest code-sharing community in the world. It is based on ‘git’ a widely-used software version control system. Unlike classic git, GitHub takes text-based code or html and merges it with a web-based graphical and social media-like interface that can work across mobile devices. While the company offers paid-for plans for private code repositories so that companies can develop proprietary code, as with a number of software companies, it also co-operates with its potential consumers via open sourcing and through a social networking platform for software engineers. ‘That open-sourced workflow is something that makes GitHub attractive in the tech community,’ explains Avalos, ‘and I started to see resonance in the legal space, using GitHub and working as a lawyer in the same way that the company works.’ GitHub is first and foremost a platform for tracking and working with software code, but ultimately the version control system works with any machine-readable text. If Avalos and I are both working on the same contract on GitHub, we can see each other’s solutions and suggested changes, allowing us to work together. We can work in a document flow with the code, seeing amendments and merging them seamlessly. GitHub has also become the default social networking platform for software developers. Similarly to Twitter, engineers on GitHub are able to gather followers and become stars in the community due to their coding skills. It has become a core tool in software engineer recruitment, as its platform relies on skill alone, enabling employers to ignore resumes and view exactly what code programmers have written. For smaller start-ups without a robust engineering resource, GitHub allows open sourcing so that programmers can fix glitches on others’ programming. Why would anyone do that for free? Because debugging allows them to develop their reputation, get stars on the social networking site and even, perhaps, segue into their next job.
Working in real time
Avalos saw that there was potential for this approach to be applied directly to the legal department. It became a case of ‘completely recontextualising the way we view the practice of law as ultimately code and text; to see each legal problem as a problem with the text you have written – so let’s open source this.’ Soon he stopped relating to the rest of the company as a reactive, traditional legal function would. ‘Using GitHub changed that typical siloed legal viewpoint for me,’ he explains. ‘I am able to view all projects at the company on GitHub – I can see what people are doing in real time, rather than being pulled in only when there’s an issue to escalate to legal.’ He can drop into the same workflow as the engineers, making interaction with legal a much more seamless process. The legal function has become part of the gestalt work process – not something separate. ‘We’re losing the sense of escalation of problems to legal, and developing a more lateral movement as part of product development flow,’ says Avalos. ‘For me it has been quite powerful in creating a legal repository which really reflects the company’s needs without reinventing the wheel every time.’
For the non-technical among us, here’s a real-world example. Not long after Avalos joined, the company wished to do a team-building exercise offsite with the potential for physical injury. HR naturally wished to have a release or waiver in place ahead of time. ‘Instead of emailing me and having a private conversation between HR and legal, I asked HR to create an issue in the legal repository, as if the lack of a release were a bug. It was visible to the entire company. It was frankly a way of preventing a sudden panic around culture drift, where people are suddenly saying, “we hired Julio two weeks ago and now we have contracts around everything.” I started involving the engineers in the process from the beginning, so they could see what legal was doing and why.’
They immediately got it, he says. ‘Within minutes it became a Twitter-style conversation that culminated in an employee remembering that they’d been asked to sign a release for an event earlier in the year, which in turn quickly resulted in that release being uploaded to the repository.’ So as well as working wonders culturally, this approach had a tangible impact on efficiency. ‘It was a eureka moment – I saved half a day of work.’ Multiply that out over a year or two, across teams, legal departments, across companies, across industries, and suddenly you’re saving millions of dollars rather than recreating the wheel every day – which lawyers have a penchant for doing, he points out.
Avalos was motivated by avoiding what he saw as the two competing factors which stymie the traditional legal function in-house: the separation of legal from the rest of the business, and the often unnecessary multiplication and repetition of day-to-day legal work. ‘There’s a real fetish around legal work; that it needs to be kept secret. But so often that’s unnecessary,’ he explains. ‘From the beginning I have been trying to build a legal function that reflects the business environment, the company and the culture. As an unintended consequence, I ended up with a system of real-time proactive legal support. A reactive model would really not work here. I looked at the cultural components and used these to establish good will within the existing culture of the company.’
But is it transferable?
Avalos concedes there is a way to go before the model is practical and acceptable for most lawyers.
‘GitHub is not a tool I would recommend to most legal practitioners at the moment,’ he says. ‘I tried to get my outside counsel to work in plaintext and on GitHub, and it didn’t go well.’ Why? ‘Most legal practitioners find it difficult to conceive of the text they are writing outside of the programme they use for that text. The idea of legal text as plaintext is foreign to most lawyers.’
So what might be the advantage of using plaintext for legal work or contracts?
For Avalos, it is efficiency – saving the increments of time often involved in deciphering and collating changes across multiple versions of Microsoft Word documents used by different lawyers involved in the process. ‘There is the 30 seconds where I have to figure out which version of Microsoft Word my counsel are working in, rename the document and so on. There are all these tiny microspheres where lawyers are wasting so much time and money because we’re working in silos and fetishising what we do. These turn into pervasive multimillion dollar wastes which get trickled down to clients via big bills from law firms,’ he argues.
He explains further: ‘If we sign off a deal with another company and they sign a PDF which is saved as an image, that’s locked away in our files. But if everything was text, I could easily get an engineer to write simple script to track changes to default contract language. Once we have the script for one legal contract, write me some other script to show the predictive patterns that, say, banks are asking for, or for different patents across a sector. On a macro, aggregate level, across deals, it would be really powerful.’
Democratising legal services
Another benefit of working in a more open-source model across legal would be the diversification of external legal services – or what Avalos describes as ‘the breaking of the big law firm cartels.’
While, admittedly, much of the tech community is white and male, world wide web inventor Tim Berners-Lee famously sees his creation as a force for democracy. Avalos identifies similar – currently untapped – potential for the web to democratise legal services by open sourcing. ‘You could have diverse and minority lawyers who are niche experts on, say, privacy but will never get in front of me, in the same way that you have an individual hacker developer. The difference with the individual hacker developer is that through the open-source community they can make contributions to thousands of individual coding projects.’
Clearly this might not be suitable for every legal issue, but it could certainly be interesting in more run-of-the-mill matters, with important ramifications for diversity. Avalos posits that if he took the GitHub employee handbook and turned it into an open source document that thousands of eyes could look at, the following scenario could be possible: ‘There may be a lawyer in the American Midwest who is now a stay-at-home mother but is still tuned into employment issues as she’s trying to keep her practice up to date. There might be a Supreme Court ruling that I’m unaware of as it has just come out, but this stay-at-home-mother could have that info. She could go to my handbook, and say an opinion came out yesterday which made that sentence on page six illegal, I could look at her pull request, press merge and be in almost real-time compliance.’
Open sourcing would also allow a wider variety of small and often minority-owned law firms to get a GC such as Avalos’ attention. ‘It’s allowing companies to get more individualised legal services without the overhead, without the law firm PR machine. It would be creating competition and opening up the legal services market. We have a glut of law schools in the US churning out hundreds of thousands of lawyers every year, and most of them aren’t going to big law firm people – and we have fewer and fewer big law firms.’
Avalos currently puts his money where his mouth is by actively seeking out smaller boutique options and using local counsel in international matters wherever possible. However, he acknowledges that this is a time-intensive process, because there’s often no other way than back channels and talking exhaustively with other GCs. He feels the extra effort does pay off, however. He is currently using a law firm in Oakland, California comprised of 12 to 15 lawyers who are former biglaw partners. Whereas they might have charged out at $795 per hour at a nationwide firm, they are now priced at $295 per hour. Says Avalos: ‘Over two to five years that sort of saving will really pay off, and I’m not paying $500 an hour extra to maintain some international law firm’s Shanghai office.’
Avalos’ ethos of streamlining processes extends to his own team, which by the end of the year will number seven to eight lawyers – quite a significant chunk for a 300-person company. Its size is down to his desire to conduct a lot of the legal work in-house in a bespoke way, using fewer outside counsel. He uses a simple metric: ‘As soon as my spend goes above what I would pay on a salary, I start to consider a hire, particularly if it’s an area that has longevity in regards to what the company is doing. By hiring someone I get much better legal support and better value for money in the long term.’
Obviously getting buy-in from the business for significant expansion of the legal team is something that many GCs might feel is as likely as winning the lottery, but this is where the unique culture of a start-up comes to the fore. Avalos has been able to leverage his experience in business to educate the board on the value of investing in the in-house function. ‘It’s a function of leadership in a start-up that they’re often more inexperienced – our CEO and COO have never worked in business before. But that’s an advantage because they are one-minded and will say “pitch me on what makes sense”. The fact that they have no experience of how it is done means that they are open for me to define what legal should look like in GitHub.’ He is free to build an atypical legal department.
Disrupting the market
Avalos sees himself as part of a community of disruption-minded people in the law. While his theories might seem similar to those of Richard Susskind, he feels that Susskind’s model is too focused on the automation of legal services, rather than the better and more egalitarian use of the thousands of skilled lawyers in the legal community, many of whom are under-deployed or not deployed at all in what they trained to do. ‘Susskind’s model is antagonistic to someone who is thousands of dollars in debt having gone to law school. But with a tweak to the message, it becomes less about just saving money alone. It needs to be deeper and more fundamental, retaking the profession and redefining what we do as providers of legal services – which should be helping to solve interesting legal problems and working to the betterment of society.’ Avalos tells me he often feels embarrassed admitting he is a lawyer: ‘The current legal services model stereotypes lawyers as parasites trying to leech as much money as possible from the system, as opposed to viewing lawyers as partners and advancing the value of what the company is.’
In many ways he echoes Clayton Christiansen’s theory of disruptive innovation, where an innovation may take business in an unexpected direction by creating something for a new set of consumers in a new market. Certainly an open-sourced model of legal services would reconfigure the current conception of what being a lawyer is, and resonates with many of the democratic, even anarchic, principles of technology innovators.
Avalos summarises the state of play thus: ‘What I have learnt from the changing landscape in tech is that there is the opportunity to be creative and to completely redesign something that doesn’t work.’