Encoding the expert’s mind.

• Sample.
  Zoom selects samples of documents from the collection.
• Train.
  A subject-matter-expert tags sample documents as relevant or not to the target category.  This is an iterative, self-correcting process, where the SME is essentially “training” Zoom to evaluate a document’s relevance.
• Monitor.
  Zoom uses a patented statistical model to continually monitor training progress and quality.
• Calculate.
  When the training process has optimized, Zoom calculates the relevance score for each document.
• Verify.
  Prior to finalizing results, verify output quality using Zoom’s built-in predictive coding QA mechanism. 

Chosen by winners.

Since the beginning of 2013, Zoom has won each of a series of intensive, head-to-head evaluations for predictive coding technology.  The evaluations have been conducted by law firms, government agencies and e-discovery service providers. Customers report three key criteria for choosing Zoom:  

• Better results.
  Zoom finds more of the relevant documents (better recall), while retrieving less of the junk (better precision). That’s due to a combination of the right classification technology, quality monitoring and statistical verification.  
• Better usability. 
  Zoom’s decision-support environment provides all the data the user needs to make key business decisions:
  • In e-discovery -- the culling cut-off to ensure proportionate review volumes
  • In information governance -- the records retention cut-off to manage risk
• Less effort.
  Zoom minimizes the training effort required in predictive coding. Via efficient sampling, rapid generation of samples, and a fully automated workflow.