Making Review More Effective
Merlin brings over two decades of experience making review more efficient and effective.
We start with Sherlock® a smart search algorithm which finds a higher percentage of relevant documents that was ever possible with keyword search. Bringing relevant documents to the front of the review queue is the single most important step to making review more effective.
We then employ a number of thoughtful techniques to make the reviewer more effective and to reduce the number of required keystrokes. Every small improvement in user efficiency pays large dividends over the course of the review. All of it helps save time and reduce review costs.
The Merlin platform is being updated by the week but here are a few of the ways we help to make review more effective:
Smarter Search + Efficient UI = More Effective Review
Most review systems provide what we call a grid view of search results. The grid is similar to that presented in a spreadsheet, with columns displaying different elements of document metadata (e.g. BegControl, Date, From, To, Subject). Each record returned from the search is then displayed in a row.
This view can be handy but it rarely provides helpful information about the nature of your search results.
To address this issue, we developed a Google-like view that includes snippets of key search results with search terms highlighted. Users can choose to display document metadata and tags across the page while showing extracted text below. This view provides immediate insight into search results saving the reviewer time that would otherwise be spent going from document to document.
Clean Customizable Review Page
Unlike many systems, Merlin IS offers a clean, functional review page without a lot of distracting but ultimately not useful options. We present document metadata in a left pane that can be collapsed if not needed. This information can be helpful but does not involve review judgments. The review administrator can determine which fields to present and the presentation order. The review can see other fields with a click.
In the center, we present either a native view of the email or other document along with text, image and produced copy views. We highlight search hits to focus the review on key areas of the document.
Reviewers focus on the right pane, which presents tags chosen by the review administrator. Tags save automatically, saving an extra click. We will be adding macros for multi-tagging in the coming months.
Family Review and Tagging
We keep email and attachment families together for consistent and uniform review. You can see that a document is part of a family at the top of the screen, moving from one to another member with a click on the document icon.
You can also view the entire family by dropping down the family view panel. This allows you to see metadata and tags about each family member and to move from one to another with a click.
Ultimately you can tag each family member individually or tag all of them with a click to select family tagging. Administrators can make family tagging mandatory or allow the reviewer to decide whether group tagging is appropriate.
When you send documents to Sherlock, he will return additional ones for individual treatment in what we call Focus View. The review can view the document in native or text formats, record tag judgments, perform redactions and then click “Thumbs Up” or “Thumbs Down” to record a judgment. Positive documents are automatically copied in a designated folder.
We are building additional macro options for Sherlock review sessions. This will allow a review administrator to set tags to be included when a review clicks Thumbs Up and other tags to be set for a negative judgment.
Cluster Batching View
Reviewing by individual documents may be the standard but it is not always the most effective way to review documents. As an alternative, we developed our unique cluster batching capability to increase review output, without sacrificing accuracy.
Instead of asking Sherlock to bring back the next relevant document, a reviewer can ask for the next fifty ranked documents grouped by similarity of content. In milliseconds, Sherlock brings back the requested documents after running a clustering algorithm to group them by similar content.
Reviewers can see a cluster label, a summary and the individual documents which make up the cluster, all at a glance. Cluster documents can be viewed and tagged individually or as a cluster. This allows a reviewer to classify documents rapidly but safely because they have already been ranked and grouped by a powerful machine learning algorithm. An experienced review can tag fifty documents in minutes rather than the typical hour it takes for reviewers proceeding on a document by document basis. And, at any time, the review can return to Focus View and move through the documents one by one.
Ultimately, building your review around a smart, machine learning algorithm like Sherlock is the most effective way to make review more efficient and effective. Our goal in designing Merlin’s review interface was to develop ways to make Sherlock even more effective to bring relevant documents to the forefront as quickly as possible, saving our clients on both review time and review costs.