Cut Review Costs and Time With Insight Predict
Predictive Ranking is a process that ranks documents by likely relevance to a particular issue. Once ranked, they are reviewed by ranked order, with likely relevant ones reviewed first. Because the ranking groups documents roughly by topic and text similarity, the review team becomes more efficient. Savings come from increased review speeds and the ability to cut off review once you have found enough relevant documents.
Lower Review Costs
Review costs make up the greatest part of your discovery spend (73% by one estimate). By focusing review on more important documents, you save on review costs. Documents below a certain threshold can be tested and, in appropriate cases, removed from the review population.
This is just like keyword culling in most respects. Rather than craft keyword searches to determine what is “in” and what is “out,” you use the Predict engine, backed by statistical sampling, to determine your review population.
If you use Predict to remove documents from the review process, the savings are immediate and obvious. Simply multiply the number of documents removed from review by the cost per document for review. Then subtract the cost to run Predict. In almost every case, you will come out ahead.
More Efficient Review
Predict allows you to place the most relevant documents before your review team first, leaving less relevant documents to be considered later. Even if you don’t choose to remove documents below a certain ranking, you still benefit. Review teams can move more quickly when documents are ordered by relevance.
That applies at the beginning, when there are a higher number of relevant documents. It also applies at the end, when most of the documents are irrelevant. Review teams can move quickly through this segment with less fear of missing something important.
More Effective Review of Third Party Productions
Predict is highly effective for ordering the review of documents received from other parties. Often, these productions come in just before depositions. They typically involve a lot of irrelevant materials with a smaller amount of important or “Hot” documents. The goal before the deposition is to find the Hot documents to present them to the witness.
When document numbers are high, reviewing them on a short schedule is either impossible or cost prohibitive. Most counsel use keyword searches to find what they need. They then review the results set which likely contains relevant documents but often contains lots of irrelevant ones as well. The team is still bogged down in review with deadlines looming.
With Predict, the team can still use keyword search to find Hot documents. Send the good ones to Predict to start the ranking. Let the review team work from the top of the order, feeding their judgments back to Predict to make it smarter. The percentage of Hot documents quickly goes up, making it far more likely they find what they need for the depositions.
Earlier and More Thorough Case Assessment
Predictive Ranking can be especially useful for early case assessment. Find key documents and give them to Predict. It will find more like your samples for early analysis and review. You get an earlier understanding of your case and your potential exposure.
At their core, Predictive Ranking algorithms are simply smarter searches than humans can produce. Insight Predict empowers your review team to build a more effective and comprehensive search, resulting in more relevant documents found and irrelevant documents ignored.
Ultimately, Insight Predict does not replace human review, it makes it faster and more efficient. You benefit from lower review costs, shorter review times and a more defensible production.