“Data! Data! Data!” — an interview with Andy Wilson of Logik

This interview is part of our new series “Data! Data! Data!” — Cures for a General Counsel’s ESI Nightmares”.  For our introduction to the series  click here.

Logik is one of the more extraordinary companies to come onto the e-discovery scene.  A dynamic company, they derive their name “the formal systematic study of the principles of valid inference and correct reasoning, and; the interrelation or sequence of facts or events when seen as inevitable or predictable.” Or, as in today’s parlance: they’re a lean, mean e-discovery processing machine”.

Located in Washington, D.C., in beautiful loft space across from the National Portrait Gallery, the company counts AM Law 100 law firms and Fortune 500 corporations as its client base.

We had the opportunity to spend a few hours at Logik’s corporate headquarters talking with Andy Wilson.

TPL: Both you and your co-founder, Sheng Yang, were in school together at Virginia Tech, but you didn’t really know each other until after school, correct?

AW: Correct. I actually started college as an English major but moved to Computer Science, graduating with a degree in Business Information Technology in 2001. Sheng and I did meet briefly at Virginia Tech.  We both worked at a web design company for a few months, but we didn’t really know each other.

TPL: But straight after school you went home to Kentucky and tried “the entrepreneurial thing” and did independent web design.

AW: Yep. After a year in Kentucky, I realized I wasn’t really in the right place to launch my web career, so I headed to D.C. (proceeds from my tax refund check happily in hand) and began to interview with a multitude of government agencies. The government was in a hiring frenzy with respect to tech/web people. Thing was, as I cruised the various floors of endless cubicles (think “Dilbert” here), I said to myself, “Andy, do you really want to do this?” and the answer was a clear “No.”

TPL: And an opportunity popped up at Driven?

AW: That’s right. I was hired at Driven as a “techie,” which I did for several months before moving into sales. At the time, Driven specialized in various aspects of litigation support including digital reproduction, paper discovery, scanning, copying, printing and graphic design. It was not “e-discovery” per se, but it was fascinating because we sold the accounts and the projects, and we actually did the scanning, printing, blowbacks, etc.

It was incredibly long hours, but I was in my early twenties and didn’t mind. I was able to bring in a few top AM Law 100 law firms in my first year, generating about $1 million in sales over my first 12 months. And it was at Driven then I reconnected with Sheng.

TPL: And an idea was formed ….

AW: Yep! I wanted to go into application development, to create an easy and scalable software platform to do eDiscovery processing. And I found in Sheng a kindred spirit. But, at the time, Driven did not want to become a software company, so we parted ways and Sheng and I went off to start an eDiscovery processing company.

TPL: With an idea … but no clients.

AW: Yep, no clients – we had nothing to sell yet! (laughs) Sheng and I spent about a year writing code and our business plan, living off our credit cards, my wife’s salary… and working from my dining room. In the spirit of Steve Jobs, Apple and his garage!!

TPL: And then enter … Superior Glacier, your first client.

AW: Superior Glacier is an end-to-end litigation support provider focused on marketing in New York, Chicago and Washington, D.C. They first came to us through a friend of a friend. We were expecting the usual “tell us about Logik” – you know, a simple introductory meeting. Which we did on my couch. But then they whip out this DVD and say “Well, we are having issues with this data, what can you do with it?” So, we loaded it up onto our server (and quite frankly we were a little apprehensive thinking “Man, I hope this works!”) and we ran it through the software we developed: Gridlogik™.

TPL: And, of course, Superior Glacier was more or less expecting your response to be “We’ll get back to you in a few days…”

AW:  Exactly. Except our software went through its paces and produced the results they were looking for in about 30 minutes. All the files were accurately processed, converted, numbered and exported with ready to import load files.

TPL: And you blew them away.

AW: Pretty much, yes. I don’t think they were expecting two guys in a dining room to solve their eDiscovery problem, but we did.

They had been working on this data set for almost two weeks with no results. We pointed out the problems they were having and how our software identified and fixed them. After they thoroughly rummaged through the output to confirm the results, we got to talking about pricing. They were used to the industry standard, which at the time was to charge based on the number of gigabytes the data extracted to, post processing. (Note to readers: Data extraction is the process of breaking down structured and unstructured data into individual records or documents. For instance, saving attachments from emails as their own documents or extracting files from .zip files is considered data extraction. This process is time-consuming and can result in the original data set exploding in size, often doubling and sometimes tripling from the original size.)

We had then, and still do, a very simple pricing model. We built our technology in a way that allows us to price eDiscovery on the original, non-extracted data size. We engineered a data-mapping algorithm that quickly identifies all documents in a data set without actually extracting it. Basically, it’s our secret sauce. So, this attractive pricing model coupled with a new technology that was very capable got Superior Glacier thinking. They sent us our first “real” project the following week.

TPL: And they paid you $40,000?

AW:  Just about, yes. Sheng and I were singing and dancing. We thought “wow, we really are onto something here!” All of our start-up costs were covered and we had enough money left over to buy some more servers.

TPL: Ah, yes, the giddy feeling from the first paying client.

AW: In fact, they flew us up to New York to discuss a potential acquisition. Then we KNEW we had something! However, we took the road less travelled, so to speak, and decided to keep Logik between Sheng and I. Superior Glacier is still a client today and we help them on a few legacy cases.

TPL: I imagine you have had a number of companies who want to license your technology (which we think can usually end badly, resulting in “brand smashing,” among other issues).

AW:  There have been a few companies interested in licensing Gridlogik, but we have always turned them down. Right now we are a business-to-business services company. If we licensed our software, we lose control over technology and become only a software company. If someone else used Gridlogik and made mistakes, that would negatively affect our reputation and our brand. We always want to do it right.

TPL: So, we have a very happy Superior Glacier. And then Fried Frank comes on board.

AW: The DC litigation manager at Fried Frank had some complex processing problems that involved unified communications, specifically Bloomberg data. (Note to readers:  for some background information on unified communications click here).

While analog is somewhat easy to analyze and parse, unified communication offers one enormous text file. Meaning, you need to know how the software created the file and requiring you to break out the metadata, and so on. It’s much more complex. In the case of Fried Frank’s client, they had about 20 gigabytes of this stuff that needed to be reviewed. And of course, we were eager to do whatever we could to help our new client. So, we modified Gridlogik to quickly parse and piece all the data together into a reviewable format similar to what they were getting with Outlook email reviews. We finished the project in less than two days and the client was very, very happy.

TPL: And this led to Fried Frank referring Williams & Connolly, who referred you to Finnegan and Henderson, etc., etc.

AW: Exactly. We have done very little direct marketing. Almost all of our business, both for law firms and direct corporate clients, has come from referrals. Granted, it’s a somewhat slower customer acquisition process, but we find it beats cold-calling any day, and we’re fortunate to have very loyal, happy clients.

TPL: Tell us a bit about the first work for Finnegan — without mentioning the actual corporate client. You know, confidentiality and all that.

AW: Well, Finnegan had been working with a vendor (client’s choice) who had totally screwed up the data processing on a high-profile matter. The vendor had worked for months on it. There were missed deadlines, incorrect deliverables and poor communication throughout. Obviously, this would frustrate anyone, so Finnegan decided to look elsewhere for help. Logik was recommended to Finnegan by one of our clients and we ended up winning their confidence after processing some sample data. In under two months we re-processed all the data, matched up the already coded documents, and re-produced the data in a much cleaner and consistent manner. And with that, Finnegan become a happy client of ours, too.

TPL: You do a pretty large amount of work for a major top 3 accounting firm, yes?

AW: We do, yes. That work is mainly for rapid “banking-related” document productions to the government. They also work with us on more complicated Lotus Notes projects that they would rather outsource to us. It’s a great working relationship that we value highly.

TPL: Tell us a little bit about your work with predictive coding, that is, the capability to use a small set of (partially) coded documents to predict document coding over the complete corpus. I believe Recommind has done a lot of work in this area.

AW: Sure. Predictive coding is going to be big in the next few years. It makes sense, considering the volume of data lawyers have to review in a finite time frame. To get our feet wet in this space, we participated in the 2009 TREC legal study. It was fascinating and quite challenging, but we learned many useful methodologies to help our clients use advanced machine learning techniques to apply predictive coding to their documents. Like everyone, we are new to this area, but we are putting more resources into it in the coming year. We’re pretty excited about it.

TPL: As we discussed, the big “new new” thing all of last year — at every event we covered — was early case assessment and winnowing relevant data down to reduce the number of documents to review.  As the stats bear out, it is the most expensive part of the process.  But now we have predictive coding, plus the work being done in computer assisted review as evidenced by Patrick Oot and Anne Kershaw’s study “Document Categorization in Legal Electronic Discovery: Computer Classification vs. Manual Review “, plus the work being done by Google and Microsoft on auto-categorization or auto-coding.  Are we headed down the path to where machines can be statistically proven to be as accurate as human review?  Is the technology getting to the point where we can also winnow out the eyeballs — contract attorney reviewers?

AW: There’s too much value in humans to take us out of the equation. Technology is just a means to an end, and I don’t think we will see document review sans humans in our lifetime. I do see document review teams getting smaller, focused and more tech-savvy, like your “special ops” reference. With the right set of tools, a small number of tech-savvy attorneys can rip through massive amounts of data in a very short amount of time. The days of expensive linear review are numbered.

TPL: But the legal industry is… to put it mildly … risk adverse. Despite the lamentations of Richard Susskind and Jordon Furlong that the law profession needs to understand the tectonic changes that are occurring, and that change will only come slowly.

AW: “Risk adverse” is definitely putting it mildly, but smaller, younger and more agile firms are starting to sprout up, willing to take on traditional practices with new billing structures and a willingness to use technology in the best interest of the one paying the bills: the client. This follows a similar path as Logik – a small, fast-paced and lean company that can deliver results in a new way using new tools and methods. Things change, and to stay prime you have to stay on the wave of that change.

TPL: In Malcolm Gladwell’s book Outliers, he mentions Skadden, Arps as an example of a firm that has taken an opportune period of time, and some cultural advantages, to give it an edge. He also talks about the “10,000 Hour Rule.” Sort of reminds us of Logik, yes?

AW: I think so. The 10,000 hour rule is an interesting concept. Basically, it comes down to timing, perseverance and practice, practice, practice. We were fortunate to start our company at the right time, just when eDiscovery was starting to get hot in 2004. We persevered through very tough times trying to validate our market and our existence being a niche processing player. And we got a lot of practice. In the first few years, we focused just on eDiscovery processing, exposing us to many unique situations that our competitors haven’t even come across yet. Unfortunately, it meant I saw my wife for maybe only two hours a day for a year. But you do what it takes to succeed. It was a lot of hard work, but also very fun.

TPL: You now have a multimillion dollar business, all done with 14 employees. But you are expanding. Given the nature of your operation, I imagine you need to consider “culture matching” to a great degree.

AW: We do. We have been looking to add an additional eDiscovery project manager … and, in fact, we advertised via The Posse List and got a great response, thank you! Fitting into the culture is super important to us. Logik isn’t for everyone, so it’s tough to find the right person. And we don’t compromise based on someone’s resume. We encourage people to check the website at logik.com for openings.

TPL: Ok, now, a key question: your Logikbot mascot, symbol [shown above] … just what is that thing?

AW: Ha! Logikbot. Well, he’s both, really. Logikbot is a metaphor for who we are; a small team of smart and motivated people offering great technology and service, taking on the established big vendors, we call them BV2000’s. Unlike many companies in this space, we embrace the fact we are small and nimble. It’s a big advantage for us, so there is no reason to act all big and mighty. Our clients and our work speak for itself. Logikbot is akin to the main character in Rudy… who doesn’t want to root for the little guy?

TPL: And this new site you have created called eDDstuff.com. That’s charity-driven, yes?

AW: Absolutely. We’re really happy that eDDstuff.com is a fun, charity-driven destination. We figured the eDiscovery industry needed funny and witty t-shirts, so we created about nine different designs from “eDiscovery ninja” to “eDiscovery nerd.” Ten percent of every purchase goes to a local DC charity, with the rest going to the vendor who makes the actual products. It’s been a huge success since we launched it and the orders have already started coming in.

TPL: Andy, it was a pleasure chatting with you. We appreciate the time you’ve taken.

AW: This has been great, let’s do it again soon. We have some very interesting things coming out in 2010 that we think your “special ops” team will really like.