Adventures In Email Threading
By Brian J. Kapatkin (bio)
Any attorney who has reviewed documents collected from a client could tell you that email is often the primary way in which people communicate matters of importance.
Unfortunately – both for the attorney who has to review the emails and the client who has to pay the bill for that review – for every email thread containing back and forth communication regarding potentially relevant material, each individual email within that thread will traditionally be reviewed, and potentially produced, as separate and distinct documents. This obviously means that a significant amount of time will be spent going over and over the same communications that make up the earlier parts of an email thread. Luckily, there are now tools available to help alleviate this issue by recognizing the thread structure of emails, and allowing reviewers to potentially eliminate or lessen the number of documents they actually need to put eyes on. In practice, even when using these tools conservatively, I have seen a 20-30% reduction in document counts of the emails that need to be reviewed. So while the topic of email threading may seem pedantic or technical, it can have a major impact on the time and money spent during discovery.
So that’s great, right? End of story. Unfortunately not. While this sort of a tool is very appealing, there are some pitfalls to consider, and precautions that need to be taken before jumping in headlong.
Over the past year, I have used email threading software in a number of matters to reduce the volume of emails reviewed, and ultimately produced. In the context of a commercial contract arbitration, where discovery rules were less rigid, and where there was a primary value put on efficiency, using email threading to reduce the volume of documents reviewed and produced was a no-brainer. However, when dealing with document productions to investigating government agencies, I have encountered push-back that warrants consideration. To the extent using email threading technology will result in producing significantly fewer documents, I would argue that it is almost certainly still worth pursuing. That being said, it is also likely required that a party get clearance before employing the technology to reduce the production volume. Based on my experience, agencies have expressed reasonable concerns about the following issues.
How does the threading tool treat email threads that branch off in different directions among diverging groups of recipients?
In terms of dealing with multi-branch email threads, it is important that the final emails in all of the branches get reviewed. While emails from different branches of a thread will surely have some overlap, there is also new content in each of the branches that obviously has to be accounted for and reviewed. Any email threading tool worth using should be able to account for this issue.
How does the threading tool deal with attachments that are part of the thread at one point, but drop off or change at a later point?
It is important that the workflow employed captures emails in a thread where attachments are added, removed, or altered, in addition to the final emails of each branch. There are certainly tools and workflows being used right now that handle these first two concerns well, and that allow all branches/attachments to be captured, segregated, and independently reviewed/produced. I cannot emphasize enough, though, how important it is that the attorneys and technical staff working with the documents account for these concerns when planning a review.
How does the threading tool deals with the metadata of the earlier emails in the thread?
This is quite possibly the biggest single limitation with using email threading to reduce production volume. If emails are threaded, and only the final email (as modified by the fixes for the other concerns described above) is reviewed and produced, it is only the metadata for that specific document – the final produced email – that is available to the receiving party. Thus, the metadata, including the sender, recipient, and date fields related to the earlier constituent emails of that thread is no longer available to be searched as separate fielded entries. This, at least in theory, limits the effectiveness of certain searches based on metadata field criteria. However, given that the entire textual content is still present, though, this concern may not – and for the most part should not – present a major issue, especially when weighed against the likelihood for significantly increased efficiency on both sides. Parties should be prepared to deal with possible criticism on this point, though, and based on my experiences, it can be overcome. And while I am not presently familiar with a tool that handles this limitation extraordinarily well, it is possible that through the clever use of technology, the metadata from the filtered-out threaded emails could be preserved and produced in some form.
So in conclusion, email threading can greatly increase efficiency and is definitely a tool that should be explored for many cases. That being said, it is important that the workflow processes be thoroughly vetted, and that the anticipated pushback be weighed when making an ultimate determination.