Classic business rule: develop the product then unleash it on customers. Modern business rule: find the skinniest expression of your product, unleash it on your customers, then lever up. Startup thinking has invaded the traditional business model.
Frank Robinson, the founder of SyncDev, claims founder’s rights to the term “minimum viable product.” Even so, Steve Blank and Eric Ries popularized the term. It is a key component of Ries’ The Lean Startup and has entered everyday speech for those wanting to show their startup chops.
If you are not familiar with startup lingo, minimum viable product (or MVP as insiders know it), sounds like a recipe for disaster. Ries gives it the cryptic definition, “The MBP is that version of the product that enables a full turn of the Build-Measure-Learn loop with a minimum amount of effort and the least amount of development time.”¹ Robinson describes it as, “the product with the highest return-on-risk.”
Without business-speak, MVP can be described as starting from the simplest, easiest to build version of your product or service. Imagine you were starting an online bookstore. Most of us would want to build the store, complete with a nice user interface and a slick checkout process. The online catalogue would have lots of information about the books, including publisher’s specs and copies of book reviews from newspapers and magazines.
If we followed the MVP approach, we would start at a simple place. The online site might consist of a many page pdf you could flip through onscreen. Each entry would have a picture of the book’s cover, basic information, and the price. We would give users a “1–800” number to call and place an order (whether they would trust us and give us their credit card number over the phone today is a separate issue, but you can see where this is going).
Behind the scenes, we take that order and buy the book the customer wants. We package it and ship it to them and then collect payment. The fundamental question we want to answer is: will people buy books online? When I was in the shoe industry and part of a team that built the company’s first online store, we had a big question: would people buy shoes online without the chance to try them on first?
At this point, we are in Ries’ Build-Measure-Learn loop. We built a simple version of the product, we measured its performance with data that told us whether the basic idea worked, and we learned from that data. We could have invested the time and money to build a slicker user interface, a prettier catalogue, or a more secure checkout system. But if we invested that time and money and the basic idea flopped, we would have problems. We used valuable resources, not gathered useful information, and wasted time.
MVP tells us to start small and simple, iterate rapidly, and pivot quickly. All of those steps should be informed by data. None of those steps exist in the legal industry.
AI Gives Us A New Tool
Artificial intelligence has introduced an element that, coupled with the MVP idea, could change for the better how we approach the access to justice problem. In a broader way, it could change how we solve many routine legal problems. The key is to think small.
Traditional legal thinking encourages finding every possible risk source and patching the hole. Over decades, this view has driven us from one or two page contracts to 100-plus page contracts. While the goal is noble, it is doomed to failure. A contract cannot cover every possibility.
Still, the same goal exists for legal work that we do for those who cannot afford to pay $232 per hour, the average rate for individual legal services. In startup terms, we go directly to a scaled up service.
While we go big with documents, we go small with data. Rather, we don’t do data. Each interaction with a client is filled with data collection points. The legal industry’s general view is to skip the data and go right to the pre-determined solution.
Given the current state of AI and what we expect in our future, this “data lite” approach is a mistake. One key to solving problems is in finding the patterns. What recurring themes accompany our clients’ needs? What data points could tip us off that an issue will need attention? When is a solution enough, or not sufficient? AI loves data and we should structure our processes to feed the data beast.
Let’s mash up the AI concept with MVP, which we can re-name Minimum Viable Legal Services or MVLS, and see what we get for access to justice. First, we know that the demand in the A2J sector is enormous. Around the world, we have somewhere between 3 and 4 billion clients whose needs go unmet. In the US alone, we can estimate the number of legal issues underserved individuals have at any one time at 232.5 million.
Assume we started capturing that data using the tools we have available today. We can use online Q&A tools — with the A’s supplied by the person in need or a helper — to construct a data set of issues. With that data set and other data we can use from existing databases (let your imagination run wild a bit), we can build “pictures” of legal needs.
Armed with the MVLS approach, we can start small with simple solutions. Each solution provides a feedback loop (the Measure and Learn steps in Ries’ definition). Using our AI and the volume of data points flowing into our model, we will quickly generate some boundaries. When a client is in this situation, a solution at that level works best.
Seeking Perfection But Accepting Good Enough
We feel the pull of old habits. A simple solution may leave our client exposed to many risks. But this is binary thinking. Before, our client had no legal assistance. The only alternative cannot be too much legal assistance. We want to tailor the assistance to what the client needs, and bring the cost down from $232 per hour to $2.32 per hour, or better yet, to a one time small fee.
Too many of our proposed legal solutions to A2J issues involve layering today’s solutions on problems. Volume, data, and a ground-up approach to solutions gives us a way to build solutions tailored to client needs and risk levels.
Disruptive ideas feel uncomfortable when we confront them. They should. They cause us to re-think those things we believe are settled. But those disruptive ideas can also cause us to re-think things not directly in their path. The AI and MVLS mashup is not solely for A2J, it will work other areas, such as corporate legal services. Through design thinking, process improvement, and technology, we can bring law into the modern era. Addressing A2J isn’t the first or last step, but it is an important step.
¹ Eric Ries. The Lean Startup (2011), at p. 77.
Ken is an author writing about innovation, leadership, and on the future of people, processes, and technology in the legal industry.
He is a Top Writer on Medium in Innovation and Leadership. Ken is an Adjunct Professor and Research Fellow at Michigan State University College of Law; and on the Advisory Boards for Elevate Services, MDR Lab and LARI, Ltd.
Ken first published this post as Minimum Viable Legal Services, AI, And A2J on Medium in Just 2017.