Precision Searching with dtSearch: A Practical Guide for Boutique Law Firms
In eDiscovery, boutique firms often compete with larger firms for matters. One of the most effective ways a smaller firm can punch above its weight is by mastering search syntax, specifically the dtSearch syntax in RelativityOne. By writing smart queries, you find the right docs faster, reduce review costs, and produce work that’s both rigorous and defensible.
Here’s a walk-through of the key dtSearch syntax options in RelativityOne, how they work, and how your team can apply them in practice.
Why Syntax Matters
Simple keyword searches (“contract”, “change order”) are easy but they often produce too many irrelevant hits or miss subtle variations of key terms. dtSearch syntax gives you greater control, you can search variations, proximity, numeric patterns, fuzzy matches, and more. For a firm with limited reviewer hours, that precision makes a tangible difference.
Key Syntax Options in RelativityOne & How to Use Them
Below I outline the major syntax constructions, what they do, and how a boutique litigation team might use them.
1. Boolean Operators: AND, OR, NOT
Use AND to require multiple terms. E.g., apple AND pear returns docs that contain both “apple” and “pear”.
Use OR for alternate terms. E.g., apple OR banana returns docs with either.
Use NOT to exclude terms. E.g., applesauce AND NOT pear returns docs with “applesauce” but not “pear”.
Note operator precedence: dtSearch treats OR before AND unless parentheses are used.
Practical tip: When you’re trying to capture all communications about “change requests” but exclude routine “change acknowledgments”, you might do:
("change request" OR "change order") AND NOT change acknowledgements
2. Proximity Searching: W/N and PRE
W/N: “within N words”. Example: apple W/5 pear returns docs where “apple” appears within five words of “pear.”
PRE/N: specifies the first term before the second term. Example: apple PRE/5 pear finds “apple” within five words before “pear”.
Practical tip: If you want emails where “termination” appears within 10 words of “contract”, but you’re not sure of ordering, use:
termination W/10 contract
If you want only where “termination” occurs before “contract”, use:
termination PRE/10 contract
3. Wildcards and Pattern Matching
* : matches any number of characters (including zero). Example: appl* matches apple, apples, application.
? : matches exactly one unknown character. Example: folder? matches folders but not folder.
= : matches numerical digits in patterns. Example: === == ==== can be used to match certain numeric patterns like social security numbers.
Practical tip: You can use the asterisk before and after a word such as below to return Smith.
4. Stemming (~) and Fuzzy Searching (%)
Stemming (~) at the end of a root returns grammatical variations. Example: apply~ finds apply, applied, applies, applying.
Fuzzy (%): matches terms with minor spelling variations. Example: app%ly means terms starting with “app” and differing from apply by one character.
Practical tip: If you anticipate OCR errors or misspellings (e.g., “authourize” vs “authorize”), you might use authoriz% to catch the variations.
5. Exact Phrases and Quotes
Without quotes: pear orange is treated as exact phrase “pear orange”.
With quotes: When a phrase includes connector words or dtSearch operators (AND, OR, NOT), you must use quotes. Example: "clear and present danger" versus clear and present danger.
Practical tip: If the phrase you are searching contains “and” or “or” and you want the exact phrase, use quotes to avoid mis-interpretation.
6. Noise Words, Alphabet File & Other Considerations
dtSearch ignores a default list of “noise words” like “and”, “the”, “to” unless they’re removed from the index’s noise word list.
The “alphabet file” determines what characters are treated as word separators or ignored. For example, if underscores ("_") aren’t recognized as spaces by default, Project_Log may not be indexed as two words unless configured.
Indexes are case-insensitive by default. If you need case sensitive (for a trademark, acronym), you must build an index accordingly.
Practical tip: Before launch of review, ensure that your index recognizes characters like “_” or “-” if the client’s data uses them heavily in filenames or custom identifiers.
Putting It into a Boutique Firm Workflow
Here’s how your team can apply dtSearch syntax in a matter with constrained resources:
Define your issue codes and facts: e.g., “Change Order”, “Termination”, “Bid Rigging”.
Develop search seed queries using dtSearch syntax: write a few targeted queries using wildcards, stemming, proximity to test hit counts.
Refine iteratively: use hit count results to refine — if you have 10,000 hits for “change order”, narrow using proximity or Boolean operators:
("change order" OR "change request") W/10 (approve OR deny)
Document your search strategy: Save each query, record hit counts, and reasoning (for defensibility).
Use results to drive review volume: With more precise retrieval, fewer irrelevant docs to review → cost savings for client and your firm.
Produce and report: Generate search term reports (STRs) or dashboards to show progress, review volumes, and refined hits — all speaking to your firm’s professionalism and transparency.
Why This Matters for Boutique Firms
Efficiency: Better queries = fewer documents for review => faster turnaround + lower cost.
Defensibility: A structured search syntax strategy is easier to explain to opposing counsel, the court, or a client.
Competitive Advantage: Even without large eDiscovery teams, your firm can operate like one by leveraging robust search syntax.
Client Trust: When you provide clear reporting (“Here’s how we searched, here’s how many docs we returned”) you build credibility.
Conclusion
Mastering dtSearch syntax in RelativityOne isn’t just a technical exercise, it’s a strategic investment for boutique law firms. It transforms search from scattershot to surgical, review from costly to lean, and output from “good enough” to best-in-class.
If your firm is ready to elevate its eDiscovery game, start by practicing these syntax patterns and build your internal playbook. The payoff? Faster case insight, lower costs, stronger client relationships, and a defensible discovery process.
Next steps, leverage Generative AI such as aiR for Review to quickly analyze and find relevant data.
