Over the past decade, the demand for eDiscovery software has grown dramatically, and there is now a multi-billion dollar market for it. As such, a lot of legal industry people are familiar with eDiscovery software. Contract analysis software has been in the midst of its own little boomlet. New companies keep entering the market, and at least a couple businesses in our space are making money. If you are more familiar with eDiscovery than contract analysis software, this post is for you.
What Is Contract Analysis Software?
Trying to get away with not reading the rest of the Contract Review Software Buyer’s Guide? You slacker! :) This post goes into a lot of detail on what contract analysis systems do and why they exist. But if you are coming from an eDiscovery background, feel free to start here:
Differences Between eDiscovery and Contract Analysis Software
There are three principal differences between eDiscovery and contract analysis software.
Document Classification vs Word/Sentence/Paragraph Classification
eDiscovery projects are often about classifying information at the document level. E.g., determining whether a given document is relevant to the issue at hand, a smoking gun, or privileged. In contract review, on the other hand, users sometimes care about document classifications (e.g., identify all leases to pass them on to the real estate team, or find all non-English documents to give them to people who can read them). But the meat of contract review projects tends to be identifying important information—clauses, sentences, words—within the reviewed documents. What happens if the target company changes control. Are there restrictive covenants (e.g., exclusivity, non-competes, non-solicits) in any of the agreements. What law are these agreements governed under. Which documents have anti-bribery or FCPA compliance clauses. Which contain a price increase provision, and what percentage increase is allowed.
Needle-in-Haystack vs Every Document Matters
In a traditional law firm contract analysis project, pretty much every document being reviewed needs to be reviewed in some level of depth. For example, if doing a merger-related contract review project, every document that is reviewed will be examined for change of control, assignment, exclusivity, and more. A law firm contract review project typically covers from tens of documents up to perhaps a few thousand; a typical Biglaw mid-market deal review often covers 150–500 agreements.* In contrast, a typical eDiscovery-use-case involves culling a large dataset down to a smaller number of relevant (non-priviliged) documents, and a far smaller group of documents that will be reviewed in detail for their potential to be introduced as evidence. As discussed, the eDiscovery documents will typically be reviewed for far fewer data points.
This difference in review detail is reflected in the per document review cost. An old rule of thumb for eDiscovery review is $1/document for human review. Law-firm-done M&A due diligence contract review runs $200–500/document, and most M&A legal review continues to be done by law firms (though we are seeing law firm low cost centers and alternative legal service providers do increasing amounts of diligence).**
Because nearly all documents are reviewed in detail in law firm contract analysis, traditionally, law firms have done review in a linear fashion, reviewing document after document. In contrast, eDiscovery projects may use more advanced techniques like sampling and having software help guide reviews.
A related distinction to the project size issue is that format normalization—which can be a big deal in eDiscovery—tends not to be a significant issue in contract review projects. While our system, for example, will process over 60 filetypes, a high portion of all documents it processes are in PDF (including scans) or an Office format.
Interestingly, we are seeing increased use of our technology to help select which documents to review in more detail in contract review projects. We expect more and more contract review projects to use AI functionalities like ours to help drive the decision of what to review, so this distinction may somewhat decline over time, though eDiscovery projects are likely to remain far larger.
New Case Every Time vs Comes Pre-Trained to Work
What is relevant in an eDiscovery review will change from case to case. Even if you are broadly doing an antitrust review, what makes for antitrust will not be the same across projects. As such, technology assisted review systems tend to be able to be taught anew what is relevant for each situation. In contrast, the relevant points in contract analysis tend to be more steady across projects. Doing a M&A deal? You’re going to be looking for roughly the same provisions every time. These often include change of control, assignment, restrictive covenants, indemnification, governing law, and maybe more. This means that contract analysis systems can come pre-programed to find relevant information out of the box. Load a contract into our system, and it will almost immediately identify a large number of data points in it. While some contract analysis software vendors target provision extraction models built specifically for vertical areas (e.g., confidentiality in employment contracts, as opposed to a confidentiality model that works across most or all contract types), or models custom built for a particular company’s contracts, our system is set up to find metadata any (English-language) contracts you put in it.
Here too we are seeing some blurring of this line, as users have used our Quick Study feature to train our system to identify thousands new data extraction models.
* Note that contract management and spinoff projects can be much bigger: tens of thousands of agreements or more. Here, all contracts will have to be reviewed for the agreed list of data points. Data points tracked on these projects can range from the tens to over 100.
** Note that contract management data extraction work tends to be done for a LOT lower per document costs, but still $10–50/document. Also note that lease review can be much more expensive. Like $2,000 a lease, if done by a large firm.