The NZ E-Discovery Blog  Facilitating proportionate and efficient e-discovery

Improving the application of keyword search terms

September 16th, 2015

Like it or not, keyword searching is still a very common method used in litigation or investigations to target potentially relevant information.

For many lawyers the use of search terms is an approach that they are familiar and comfortable with. It is not hard to see why keywords are so comfortable, as we all use keywords every day through the likes ‘google‘, but many are unaware that search terms in eDiscovery might not bring exactly what they are after.

The review exercise usually equates for about 70% of the cost of a discovery exercise. To reduce these costs and to maintain the emphasis upon proportionality, it is important to reduce the volume of documents for lawyers to review. Achieving this will require the development of a strategy that allows lawyers to get to the most important information in a way that is quick, cost effective and accurate.

Limitations of keyword searches

keyword searches frequently do not produce the results expected

Traditionally lawyers have turned to keyword searches in an attempt to reduce the document volumes. The reality is that keyword searches frequently do not produce the results expected, and more importantly parties are often unaware of the limitations of keyword searches.

Keywords create a black or white scenario – a document is either in or out – a document either contains the selected word or it doesn’t. Running search terms across a set of documents may reduce the volume of documents to review, but it could mean that many relevant documents are missed if the keyword does not result in a ‘hit’ in a document. Alternatively the search terms may produce many irrelevant documents that may have different meanings to that intended.

The problem with how search terms are applied

Anyone that has read any of my material or heard me speak will recognise my scepticism in how keyword searches are often applied.

The problem is not using a search term approach, but more the method in which they are applied. It is not uncommon for lawyers from opposing parties to agree the search terms which then become entrenched in the discovery order. These search terms are developed far too early in the piece by lawyers without any testing or sampling of the search terms. The legal team usually compile the search terms based solely on their knowledge of the case – often without sighting documents.

This approach can still be salvaged if the process has the ability for the parties to refine the search terms as required. Often this never happens.

The result of this is the legal team then commence their review, with much of the information being irrelevant with documents having different meanings or be used in a different context for what may be relevant for the matter, or worse still having missed potentially relevant material – information that lawyers may never see to review.

Anyone that has been on the end of this type of keyword centric approach as a reviewer will appreciate the frustration this causes. Frustrations that are compounded by the knowledge that lawyers are needlessly manually reviewing documents – that only add to the time and cost of the discovery exercise.

Steps to improve the effectiveness of keyword searches

Search terms can be an effective method of isolating important information, but it is essential to invest the time devising appropriate terms. Any approach (at the very least), should involve testing, sampling and then potentially refining terms as required.

My 2012 NZ Lawyer article highlighted steps that may assist in making a keyword approach more effective –

  • Construct initial keyword search terms: Use the knowledge of individuals involved in the matter to identify any specific language, terminology, or abbreviations that may be specific to the matter. Too often, search terms are devised only by lawyers, without direct knowledge of potential words and usually devised too early before any documentation has been looked at.
  • Test the initial search results: Test and sample searches to find if they have produced the results you were seeking. Confirm the results are not over inclusive or under inclusive. Ensure the keywords are comprehensively searching all file types. If one party has scanned paper documents that are not 100 per cent searchable, their search terms will not capture the documents you are seeking. Confirm the tool(s) being used to search for documents are capable of running the searches you require.
  • Refine search terms: Keep refining search terms, until all parties can agree on the keywords that will be effective in locating the information you require.
  • Consider if alternative search and review strategies may be more appropriate: Spending sufficient time devising effective keyword searches may reduce the volume of documents to review, but often these documents are then just reviewed in the same way as paper – essentially conducting a page-turn exercise, but on screen. This can still lead to a time-consuming and expensive process. There are now more effective methods to carry out the document review, using the full benefits of technology to reduce the cost and improve the accuracy of the review.


Keyword search terms are still a common method used in eDiscovery exercises, and they can be conducted effectively. However, keyword searches do not find what you are looking for unless they are carefully crafted and refined if necessary.

Keyword searching is one of many approaches to isolate the most important information. Like all other methods, it is important to test and sample your approach and the refine if necessary. The ability to substantiate the results of your proposed search terms will be an advantage when looking to agree the search terms with opposing parties.

Spending time upfront in devising an effective search and review strategy for your matter will help to provide benefits later in proceeding.


No comments yet.

RSS feed for comments on this post.

Sorry, the comment form is closed at this time.