By identifying and ranking data sources, including accessibility, cost, and reliability, we can preserve and analyze only the issues that matter.
CCBJ: Dan, can please explain “Fact Crashing” for our audience?
Dan Regard: Fact Crashing, as a portmanteau, refers to the use of digital facts, data, and records to crash litigation. It’s a term borrowed from construction project management which means accelerating focus on one part of a project to the overall benefit of the entire project.
What does that mean in the context of litigation and discovery? We have found, over the last 25 years – and increasingly in litigation today – that focusing on alternative sources of structured data – digital facts – can rapidly accelerate the resolution of disputes. That’s how you “crash” litigation.
As a methodology, Fact Crashing identifies the sources of data that might be informative for a set of investigations or disputes. How do you qualify and prioritize data those sources? It’s important to prioritize data because we’ve moved from a world of scarcity of evidence to a world of abundance. This is important. For those of us trained in jurisprudence, we were taught how to deal with a few pieces of core evidence, how to authenticate them, move them through the system, present them, and tie them into the narrative of claims and defenses in a case. All of our discovery rules, while infamously generous in the United States, are really focused on acquiring what is often seen as a finite set of evidence. We crossed a turning point 10 or 20 years ago, but we’re only realizing the implications now. There is more evidence out there than we have the resources to deal with – and more evidence than we need to deal with.
The changes to the rules of proportionality cannot have come at a better time because today we truly have more data than we should deal with. Nonetheless, there is some amount of data that is proportionate to the needs of case. It’s informative and it’ll help us resolve the factual aspects, and sometimes the legal aspects, of a claim or defense. Fact Crashing helps us identify a broad universe of potential data sources and how can we qualify the accessibility, the cost, the reliability of these data sources, the ease of using them, their economic profile, as well as their legal reliability. We can rank those and come up with a preservation plan, a collection plan, and an analysis plan to address the issues that really matter.
That leads into my next question. Talk about what is discoverable and how the Federal Rules of Civil Procedure play into this aspect of litigation.
Under the broad interpretation of the rules, anything that’s recorded and informative toward a factual question or resolution of a dispute is potentially discoverable. That’s a layperson’s definition. We can go into Rule 26 and see what the formal rule definition of it is, which is how we look at it. If it was recorded and it’s reflective, relevant, responsive, non-prejudicial, non-privileged, it’s discoverable. That provides a scope of inclusivity, and a few exceptions that can move things out.
What does that mean for an automobile accident? Discoverability might concern your calendar for a specific day? What were you planning to do? Where were you going? Where did you come from? Where would you have gone next? Were you driving urgently? Were you not driving in a rush? That provides insight.
Next, what does the vehicle say? Do you have a black box in your car that tracks your carburetor, your speed, your gas, your use or non-use of the brake? Do you have GPS that can triangulate your speed, your location, your destination, your point of origin? Are there RFID chips embedded in the paint that allow people to understand where you’ve left or picked up paint traces? Are there traffic cams that show the accident from afar? Are there taxis with dash cams that went by? Were there potential witnesses on the street? Have we done a geofence of Twitter or Facebook to see if anybody posted messages with GPS locations that would put them close enough to the accident to be an eyewitness? Did they actually post anything about the accident itself? Does your cell phone have an accelerometer that indicates whether the car accelerated or decelerated prior to the accident?
All of these are potential sources of contributory information, but some may be more accessible than others. Even in a world where we’re creating more and more digital information, all information is not equally accessible. Pulling the accelerometer information off the chip embedded in your phone is a very difficult technical process that can be extraordinarily expensive. That information may or may not be informative and contribute to the overall analysis of how the accident happened and who’s at fault.
We should also look at the light timing sequences to find out from the computerized system that controls the streetlights which light was facing which direction at what time prior to the accident, assuming those clocks are in line with your GPS clocks and your vehicle clocks. So, we have different sources of information, some of which are in control of third parties. Some are very technical to access. Some we have tools to make access easier. Taking all of that into consideration, we can work out a data discovery plan that highlights the sources that are most affordably accessible, most reliable, and least subject to misinterpretation or inadvertent modification. We can come up with an overall discovery plan.
That’s how discovery has changed in a world with multiple sources of data. Some of this is in our control, some is in the control of third parties. Some we’re aware of, some we’re not aware of. Some of it persists, some of it lasts only milliseconds. This can create a data rich environment. But you’ll never get access to some of it. We like to joke that NSA may be monitoring cities and may capture, through satellites the actual accident that happened, but you’re probably never going to get that video.
What is the optimal order of operations or procedures for dealing with these varying types of data?
That’s where crashing comes in. Historically, discovery has been very expensive, and it’s become even more expensive because we’ve created more and more documents and documents are expensive to work with. We can use AI, TAR, clustering, threading, offshore coding, onshore coding. All these are great techniques that are very economical, but at the end of the day, the volume of documents continues to increase at a rate that seems to outpace the efficiencies gained by technology. So discovery is and remains expensive and will become more expensive going forward.
As a result, most parties conduct exhaustive motions practice before they move into discovery. We haven’t conducted any actual discovery depositions yet because of the expense. But if we live in a world where there’s a different type of evidence, a transactional digital record that does not require humans to read it in the classic sense because it doesn’t have the ambiguity of language, it has objective hard numbers and values like timestamps and date stamps, we don’t need teams of people to review these documents. They’re not subject to privilege.
That evidence becomes much more economically accessible, and we have found it can become so economically accessible that there’s an argument for dealing with it before one engages in an extensive motions practice. That helps reframe how one litigates a challenge. Do you go through motions practice for months or even years until you have exhausted every remedy, or can you look into the facts and come up with a quick assessment? If there’s not a problem, you may have a shortcut to dispute resolution. If there is a problem, you still have the option to pursue your motions practice with a revised sequence. And changing the sequence can have a dramatic effect on the overall cost and time involved in resolving disputes. That’s the crashing part.
How do you identify actionable data, and how do you work with verbal data versus other forms of structured or unstructured data?
Well, again, there are some techniques for identifying what we’ve referred to as actionable data or action data – data that is available, abundant, mechanically captured, oftentimes instrumental (captured by instruments). It’s not dependent upon human data entry. It’s more methodical, more consistent, more accurate. That type of data can exist in a multitude of locations. We’ve identified techniques for reviewing a given environment for sources of data. We do drive-alongs and ride with employees to see the systems with which they engage. For example, you may use your badge to swipe into your building, but you may not think about your smartphone’s automatic connection to your office Wi-Fi.
We don’t realize the data trails we leave as we go about our lives. Most litigators, either from experience or through training, stick to documents, emails, text messages and social media postings. As a result, they are not looking at structured data when they get into discovery, they’re postponing it to the end of discovery, maybe with an unspoken desire that they’ll resolve the case before they have to touch any databases. If they don’t know databases and are uncomfortable with them, it’s a risk. So they postpone that risk. What suggest is not only that they should not put it at the end, but that there’s also a good argument, in many cases, for making it the very first thing you deal with. Obviously, that requires different skill sets. It requires techniques and tools to deal with data and it requires a rethinking of civil procedure.
Please discuss frame case issues as they relate to data.
Someone asks a question such as, “Was reasonable care given to the patients at a senior living center?” Reasonable care sounds like an issue that can only be answered by the jury or the judge in a bench trial. That’s correct, but there are precursor issues that can be identified by data if you frame your issues in terms of data. For example, what was the ratio of staff to residents? You can look at time and attendance records to determine how many people were on-site at any given time versus how many residents were in the center. That could give you an idea of staffing ratios, which can be compared to staffing ratios at other facilities to come up with an opinion of whether it mattered, whether it was sufficient, whether it met industry standards, whether it was best practices.
Likewise, you could analyze medication distribution. It turns out some of these extended care centers and hospitals use rolling carts that track the distribution of medication by patient and by pill. They barcode and track everything We could look at water consumption to see if people had sufficient access to bathrooms. We could look at food distribution to see how they were fed. We could look at temperature controls to see what the climate was. We could look at HVAC settings to see if air was being pushed into the center or air was being extracted from the center, which has a dramatic effect on communicable diseases. We have a lot of ways we can look at reasonable care or standard of care through data once we put on those data lenses.
What are the most pressing issues your clients expect to face in the next two to three years? Technology is evolving, there are new data sets, new ways to quantify or qualify. Where will they put their attention?
The biggest challenge my clients have spoken about is employees unknowingly engaging systems that create data trails that may be informative, responsive and relevant to litigation. It’s the emergence of a multitude of systems that can generate and even preserve or aggregate and maintain digital records creating a situation where you don’t know what you don’t know. That’s an area of both risk and exposure, as well as lost opportunity to the extent the data would be helpful for a given litigant’s position.
The other big challenge is data privacy. I work with several groups that deal with data privacy, domestically and internationally. I’m very aware of the sensitivities. In our current discovery profile, there is no objection to discovery in the United States because the information is “private.” In fact, one might argue the reason that you go to discovery and litigation to get documents is because the information would otherwise not be available because it is private. We have situations where companies are collecting data for parties who are not part of a litigation and it can be very personal to those parties. We have either explicit or implicit duties to protect that information, and we’re looking for ways within litigation to make the information accessible and protected at the same time. That’s going to continue to be a challenge.
Is there anything you want to add?
At iDS for the last 14 years, as well as in other groups that I’ve had the opportunity to work with over the last 25 years, data has always been a part of litigation. Today we’re seeing more and more litigations with structured data and more and more structured data per litigation. These techniques are not new to us and we’re proud to share them with other people. We think it will help raise the skill level across the bar – raise the bar across the bar, so to speak. This is the new frontier of skills needed for a well-rounded litigation team. These techniques for identifying, qualifying and prioritizing data sets today have to be done explicitly. But as we get more comfortable, they’ll become, as with so many other aspects of discovery, part of the intrinsic and implicit way that we approach dispute resolution. It’s going to be an exciting time over the next 10 years as we get accustomed to these new demands.