When considering the use of TAR (technology assisted review or predictive coding), too often the question centres on the volumes involved in the matter.
Even though the volumes will be a significant factor, other considerations that may also impact are the time frames that you have, together with the resources that you have at your disposal.
Substitute resourcing
TAR can be a great option for smaller to medium size firms or barristers that may not have the same resources or budget at their disposal as some larger firms. They can leverage the expertise of their legal team (which may only be one or two lawyers), with the aid of technology to get to the information that matters most, doing so quickly, cost effectively and accurately.
One of the most successful TAR exercises that I advised on was when a small specialist litigation firm only had the one senior solicitor to work on the matter. You can read the TAR case study here.
Even if you can call upon junior resources, it may be that TAR is a more effective option to increase the speed and accuracy of the review.
OK then – but, how many documents?
One question that I am constantly asked is what size matter should we consider TAR? From my experience, any matter with in excess of 100,000 initial documents then the use of TAR should at least be considered.
There are arguments that the technology can be used on volumes less than this, but every matter is different. I have heard examples of TAR being used on matters with as little 30,000 documents, to others requiring a one million document threshold to consider its use.
Inevitably as the volumes increase, then options like TAR should at least be considered.
When TAR might not be the right option
If you have high volumes of paper or numbers, then TAR might not be the best option for you.
TAR can struggle with large volumes of paper documents (especially if the scanning quality and subsequent OCR is poor), together with documents that have considerable numbers in them.
To be fair, any search centric approach would come across similar issues.
One of the problems with TAR has been the increased iterations and training required if the scope or issues change – and we all know these never change !! If the scope does change, then earlier calls may need to be changed. The same would apply with a linear review when information may have earlier been classified as ‘not relevant’, but later becomes ‘relevant’ – again are we holding TAR to a higher account?
To help alleviate these issues new TAR 2.0 technologies have developed that enable continuous learning throughout the review process.
TAR alone will not reduce the costs of eDiscovery, as there are many other options that can be used in conjunction with TAR, or simply as an alternative too. Options like keyword searching, concept searching, near duplicate and email threading may be more suitable than TAR. Even a linear approach can still have a place !
To evaluate if TAR is right for you, you should invest the time at the outset devising the most effective plan for the matter.
I would suggest the advice you get is independent and not solely from a TAR provider !
