Crafting collaborative searches
Shifting the paradigm of how users create, edit, and share
search queries on Everlaw
Released May 2024
Overview
Creating precise search queries is critical for legal teams to curtail terabytes of data. Although powerful, the Everlaw search experience was so siloed and out of sync that when it came to reviewing results with teammates, users preferred collaborating via binders on the platform instead.
Working closely with a product lead and 2 engineers, we completely shifted the paradigm of how users collaborate on building searches.
PROBLEM
It started with a cry for “dynamic binders”
On Everlaw, searches and binders share similar properties. Both help users filter, organize, and review key documents.
Often moving search results into binders, users perceive searches as a vehicle 🚗 for filtering documents and binders the final destination📍.
On Everlaw’s homepage, Search and Binder cards both open table of documents. Users can add documents from a search into a binder.
Binders were convenient for bucketing, tracking, and sharing search results to conduct document review on.
However, this workflow was too manual as cases progressed. Teams needed to auto-incorporate new, relevant search results into binders over time—teams wanted dynamic binders.
“Your competitors can do this!” they claimed.
USER RESEARCH
Realizing it’s about the journey
We interviewed various document reviewers, paralegals, and litigation support specialists to uncover the following pain points and core needs:
New data constantly rolls in
Project admins need an up-to-date understanding and fine-grained control of what goes into their document sets.
Culling documents takes trial and error
The current iterating experience is messy, embarrassing, and unforgiving.
People make mistakes
Teams need to be able to quickly backtrack, fix issues, and hold teams accountable.
It was unclear how searches and binders differ
Given the similarities and limitations between each one, users resort to hacky workarounds for both.
We learned that creating and reviewing document sets is a non-linear, adaptive process. For users to collaborate and refine document sets more effectively, our design goals included:
increase transparency and automation,
leave room for human error when collaborating,
and enable clear mental models for the most efficient workflow.
THE CHALLENGE
Challenging users’ mental models
Technically, binders and searches both open a table of documents for users to review. How can users be so sure dynamic binders were the answer to their pain points, when in theory, both binders (📒) and searches (🔍) can serve as document destinations?
Thus, we had to examine the features’ differences and limitations to determine the real ask.
Document traceability
🔍: There is a clear criteria of which documents are included in the document set.
📒: Arbitrary collection with no ability to trace where documents came from.
Sharing and editing
🔍: Searches are personal, but a new manual copy of the search can be shared with teams.
📒: Collaborators on the binder can automatically see document updates.
Dynamic vs. static document sets
🔍: Search results automatically update as the project’s database changes.
📒: Binders require manual additions and removal of documents.
To address users’ pain points, it became clear they needed the best of both worlds:
The traceability and dynamic inclusion of new documents from searches
The ability to easily share, refine, and auto-sync document sets with collaborators from binders
The flexibility of having dynamic or static document sets when needed
Users expressed tweaking binder behaviors would solve this, but we wanted to challenge that assumption, leading us to ask…
Whether via binders or searches, how might we help users
retain the full context and control on their document sets while
ensuring teams are up-to-date as they collaborate and iterate over time?
SETTING THE SCOPE
It boils down to branding
We ideated two potential approaches to give our users the best of both worlds:
Have our already collaborative binders adopt the dynamic property of searches (as requested)
Have our already dynamic searches inherit the collaborative aspects of binders
The key decision came from stakeholders asking us how we want to brand binders and searches. It became obvious once we examined each’s conceptual model respectively.
Binders and folders are generally an arbitrary collection—its contents are dictated by users, naturally feeling more permanent.
Search results are transient in nature.
Would users expect content to change automatically in their personal organization systems?
We believed no. Applying dynamic properties to a binder defied our users’ basic mental model of what a binder is.
By preserving the distinction of binders being static and searches dynamic, we also help users develop a stronger intuition for each tool’s purpose.
It was clear—upgrading the search experience it was!
USER TESTING
Getting the details right
As we optimized the collaborative search experience, we grappled with a crucial question: should all edits be instant or require a manual save and a page refresh?
Live changes risked encountering jarring updates for teammates if a user was only intending to experiment with the query.
However, manual saves posed version conflicts during simultaneous edits.
Testing a clickable prototype with 6 users resulted in a 50/50 split in preferences. To navigate the mixed feedback, the product lead and I revisited the most common use cases, considering factors like simultaneous editing, user roles, and frequency of changes.
We ultimately opted for a mixed approach—live changes for small tweaks like applying filters and manual save and page refreshes for larger changes that would alter the document set. This aligned best with our design goals of automation while preserving room for trial and error.
SOLUTION
Decimating friction in searches
Although there were minimal UI changes, the new behaviors we introduced fundamentally changed how our users create and modify searches on Everlaw.
Collaborate in real-time in a shared workspace
Users no longer need to constantly re-share new versions of searches to ensure everyone has the latest individual copy. Toasts keep users updated on each others’ changes.
Iterate with ease with flexible editing and save options
Tweaking a typo in your search query no longer creates a new search card that clutters the homepage and requires re-sharing.
Trace changes and easily recover from mistakes via version history
Users can now see how document sets evolved over time for reporting purposes, and can easily restore and create copies of prior versions of the search.
RESULTS
Users were absolutely ecstatic
Since this feature was released in May 2024, we have gotten extremely positive user feedback, including some users expressing that this update was “hugely impactful” and “opens a significant level of differentiation between Everlaw and your main competitor”. One even claimed that search update, alongside a few other long-awaited features, made our release day “better than Christmas!” 🥳 🎁
Given that this was a recent release, there is no quantitative data available at this time. In the following quarters, I would like to compare user satisfaction scores from Everlaw’s annual user surveys and continue to conduct user interviews about the new search experience.
TAKEAWAYS
Remembering to reason from first principles☝️
The product lead and I came into this feature with a user’s proposed solution at hand. Constantly going back to first principles allowed us to extract the real problem and provide users with a simpler, yet opinionated solution that aligned more with our users’ mental models and Everlaw’s goals.
I also learned that it is always helpful to understand the context around past decisions to inform current ones. Learning why search lacked more developed collaboration features in the first place helped us determine how to close the gap between its current state and the ultimate product vision.