I rode along with a full day Relativity analytics training session yesterday. No. It was not billable, but I will be working with that client to develop their workflows and protocols to leverage the analytics in their discovery requests. I consider it a good investment of my time in this time. Besides, I love an excuse to get hands on with tech. Overall, I was impressed at the usability and functional improvements achieved by the expanded PM team. There are sexier GUI’s that support very specific workflows, but few that combine functionality and flexibility. Relativity is still an ‘expert platform’ requiring substantial skills investment to manage your ESI lifecycle, but it is slowly becoming more accessible to my corporate clients.

Relativity’s analytics fall into three buckets and specific corporate pain points:

  • Structured analytics – threading, near dup, name/alias normalization, filtering out repeated content (footers)
    • A corporation needs to have basic search, processing and culling capabilities to support internal investigations and their steady state discovery demand to control costs. Reassembling conversation strings and spotting gaps is a key function in reconstructing who said what and when. You can open exported PSTs in Outlook to group conversation messages, but this does not scale when tracing contract issues over years or spotting side conversations.
    • In this age of M&A activity and cloud migrations, most longterm custodians have multiple SMTP email address, username or display name variations. It can take hours for staff to run down all the variations and build out the identity profile. This can be critical to place holds and collect from archives, legacy repositories and content migrated into O365.
    • Who does not hate removing the search hits found in email footers, standard disclaimers and security notices? The repeated content analysis builds out a list that counsel can review/approve for suppression from the index. A straight forward workflow that can be documented for defensibility and periodically updated.
  • Conceptual clustering (LSI) – keyword expansion, similarity, concept navigation, categorization
    • I cringe when reviewing proposed search term lists generated by associates from opposing discovery requests. Relevance criteria should come from the key players and data. Companies like H5 made their fortunes tackling the high art of linguistic relevance analysis. Concept clustering and keyword expansion provide more accessible tools for a practitioner to navigate from a known seed term to develop data based relevance criteria. Scope definition should be a defined, documented process. Frankly I would prefer to see it performed on the in situ data. One of these days Microsoft’s O365 features and scalability will enable this. Until then, corporate discovery teams can export key custodian data directly to RelativityOne for ECA and investigation analysis.
    • ROI calculations drive almost every one of consulting projects. Corporations must have the tools to retrieve rich, relevant data sets rather than exporting raw custodial sets for every request. With unlimited O365 storage, I am seeing custodial hold metrics growing at astronomical rates. Legal teams need tools like concept dials and bubbles to visualize conceptual and term relationships.
  • Active learning (SVM) – training the SVM to predict relevance in large sets and validating the model
    • Although I enjoyed seeing Relativity’s machine learning strategies, I do not see many corporations eager to ‘own’ the actual review process. My in-house team grew out large scale contract review capabilities during the Enron investigations back in 2000. We dissolved those teams after the company went back to ‘normal’ litigation discovery. So while I can see how an aggressive IG program could leverage active learning engines to create retention rules, compliance monitoring and other neat functions, I do not see a clear and compelling ROI for these use cases yet. Could a sharp paralegal use the active learning to create very rich data sets for counsel to resolve many matters? Yes. Is the effort and review effort of active learning justified proactively in most matters. No. I believe that the structured and concept cluster features will get you there without the review batches.

So are Relativity analytics adequate and worthwhile for corporate legal departments with a reasonable litigation profile? I think so. There are cheaper alternatives who can arguably claim better usability or unique functionality. But a smart client remarked, “Everyone speaks Relativity.” His point was that he did not have to pay to have reviewers or associates trained in Relativity. It was expected. Doing the upstream scoping, culling, categorization and assements makes sense if the review will be done in Relativity by firm/contractors anyway.

So have you been using analytics for corporate ECA and ROI? Love to hear about it.

Greg Buckles wants your feedback, questions or project inquiries at Greg@eDJGroupInc.com. Contact him directly for a free 15 minute ‘Good Karma’ call. He solves problems and creates eDiscovery solutions for enterprise and law firm clients.
Greg’s blog perspectives are personal opinions and should not be interpreted as a professional judgment or advice. Greg is no longer a journalist and all perspectives are based on best public information. Blog content is neither approved nor reviewed by any providers prior to being posted. Do you want to share your own perspective? Greg is looking for practical, professional informative perspectives free of marketing fluff, hidden agendas or personal/product bias. Outside blogs will clearly indicate the author, company and any relevant affiliations.

4 2 votes
Article Rating