Sift browser plugin

Empowering people to feel confident when travel planning

Project info

I collaborated with a team of 4 students at Carnegie Mellon University (CMU) to design Sift, a browser plugin that cross-validates information from various online sources, including travel websites, blogs, and reviews. The Sift plug-in displays comparison results, streamlining the workflow for travel planners by eliminating the need to switch between different browser tabs and windows. This project was for the course User-Centered Research and Evaluation at CMU.

lead product designer
ux researcher
project manager

September - December 2022
4 months

2 product designers
2 ux researchers
1 ux technologist

My Contributions

As the lead designer, I advocated for a research-driven design process and collaborated with the team to refine the user flow and prototype through user testing. I led the interaction and visual design, culminating in a high-fidelity prototype utilizing multiple interaction types (highlighting, clicking). As the project manager, I supported the team to navigate ambiguity by breaking the design problem into actionable steps to achieve sprint goals.

I led the high-level business strategy of Sift to ensure the design would deliver business value. The strategy aims to support business goals by gathering information that users highlight and sharing those statistics with the host websites. Providing detailed insights into how users interact with website information helps host websites to better curate their content.



How can we help people make travel decisions when navigating the overwhelming abundance of internet resources?


Sift is a browser plugin that cross-validates information across various online sources to help travel planners make efficient and informed travel decisions with confidence. Sift helps users make faster decisions, whether traveling solo or in a group, allowing for more time to collaborate, helping group members feel included, heard, and informed during trip planning.

Final solution

Sift is a browser plugin that cross-validates information across various online sources to help travel planners make efficient and informed travel decisions with confidence, whether traveling solo or in a group.

Highlighting travel review text seamlessly fits into user's workflows without tab or window switching. Summary results, utilizing text sentiment analysis, help users make decisions faster to streamline the travel research process. The plugin format frees up valuable screen space, reducing cognitive load and visual clutter.

Final solution design decisions

I iterated upon user feedback to create the final design features.

Results page

you've highlighted page

minimized plugin window


80% of travelers will spend about 4 weeks pre-travel researching and planning before booking, according to TripAdvisor.

Travelers often face the challenge of making informed travel decisions due to the overwhelming abundance of online resources. This becomes even more pronounced as many travelers have limited time to plan their trips, juggling their responsibilities as students or working adults. Navigating this overload of data can be daunting, leading to decision paralysis and potentially making poor or uninformed choices, leading to a less than ideal trip.

Background research

I consulted white paper and informal resources (social media, blogs, etc.) to understand how people make travel decisions through research and planning.

Conducting research using a wide variety of sources helped to gain diverse perspectives on the travel landscape.

Finding 01

There is no single universal resource that people use for their end-to-end travel planning.

Finding 02

The process of making travel decisions depends on the nature of the trip, whether it's solo or with a group.

User interviews

I interviewed 5 participants to learn firsthand how travel planning differs for solo and group trips.

As a team, we interviewed a total of 25 participants using artifact analysis and directed storytelling methods. Participants shared their processes, tools and itineraries for researching and planning trips.

Finding 01

When planning a group trip, often one person becomes the designated researcher/planner and will ask for others' opinions.

Finding 02

Research is the most tedious stage of travel planning, whether it's for a group or solo trip.

Finding 03

Travel planners have go-to tools that they use for documenting and communicating travel plans, but do not have defined resources for research.

User journey

I created a user journey of how a trip is planned when one person assumes the role of 'primary planner' to reveal pain points about the current workflow.

From our interviews, we found that the beginning of planning a group trip is collaborative because group members need to decide on important travel milestones, such as the destination and dates. Other travel aspects, such as booking flights, finding accommodations, and planning activities, fall onto the primary planner. They will conduct research, communicate to the group the different options they found, wait for approval, and then book. If the group does not approve of the primary planner's options, then they have to go back to research until they find more suitable options.

Key research findings


Researching is the most crucial yet tedious step of travel planning. People feel they’ve made an informed decision only after consulting multiple sources.

"The most tedious part of the planning process is referencing different information and weighing what all the sources are saying to make a decision."

"I use a mix of reviews from different travel websites so that I can view the place or activity objectively."

"How do I know if the source is reputable? I try to cross-reference across different sources."


Decision-making becomes less collaborative as the trip date approaches because there's less time to deliberate and make informed decisions.

“Lots of back and forth on the dates. Date preference was important, it also influenced how many things I can realistically plan to do on the trip.”

“For activities, I look at reviews and book them. It’s not very necessary that everyone is a definite yes for whether they want to do it."

“As for activities I knew what my friends liked and just went with it, and didn't spend a lot of time on getting individual approvals.” 


Trip initiators assume the role of the primary planner in order to simplify the complicated process of trip planning.

“He planned it because he was the one who initiated the trip. It was just easier if he did it because it would take too long if we planned everything together."

”She likes to take charge of the planning. And since we were all busy it was just easier for her to do it.”

“He proposed a broad range of activities to his group."


I used crazy 8's to come up with ideas based on "how might we statements". This helped to reframe our insights for ideation.

My ideas were inspired by the statement, "How might we help users identify and compile web resources most relevant to what they're looking for?". In my sketches, I was interested in utilizing natural language processing (NLP) and AI technology to provide users with relevant resources.

Storyboarding and concept validation

As a team, we narrowed down our ideas to create storyboards for concept validation sessions to understand what potential solutions might best address user needs.

The goal of concept testing using storyboard scenarios was to get initial user feedback on whether a product that cross-references information from multiple online sources would be a valuable and trusted tool when travel planning. Participants responded positively to storyboards that I drew for a plugin that cross-references information from the web, indicating that the user pain point of a tedious research process resonated with participants.

Determining the plug-in format

Utilizing Sift within a browser window is compatible with the investigative nature of travel research, as opposed to research conducted during casual browsing, which might take place on a mobile device.

The Sift plugin format optimizes screen real estate and minimizes context-switching. I sketched how Sift might be expanded from a plugin to a full browser screen. A plugin format required the fewest steps to achieve the user goal of a decision determination.

Paper prototyping

We created a paper prototype and conducted 3 user testing sessions to identify user flow gaps.

Transforming the project concept into a tangible reality was one of my favorite parts of the design process. I was surprised at how well the highlighting interaction translated to the paper prototype, as users found the interactions to be intuitive. However, the low visual and content fidelity did not yield very accurate results. So we decided to increase the fidelity for the next round of user testing.

Refining the user flow

I created a user flow diagram to map out the user steps and identify any points of friction.

Low-fidelity prototyping

I created low-fidelity wireframes to increase the visual and content fidelity for 5 user testing sessions. We learned that users would want to highlight both overly positive and overly negative reviews to be analyzed.

This helped to obtain user feedback that more accurately depicted the experience of using the plugin. Participants liked the small size of the plugin and were curious if it would be able to work on any website and if it could be used for analyzing subjects other than travel.

High-fidelity design iterations

I conducted 3 user testing sessions to refine the clarity of the highlighting instructions and interactions.

User testing sessions played a crucial role in identifying and resolving user friction points and enhancing interactions. Observing participant responses and behaviors within a screen-based environment proved invaluable, given the unique nature of the plugin. Our primary focus was enhancing the onboarding process and instructions, as users found the highlighting interactions to be intuitive once they grasped the scope of what they could highlight and comprehended Sift's functionality.

Activation of highlighting mode was not clear. There was only one visual indication that the mode was activated.

I added a colored border around the webpage as an additional visual indicator of the highlighting mode without obstructing the webpage content. I added a toggle to make it easier to switch between modes.

Users were surprised that highlighting immediately added the text to the plugin to be searched.

I added options that appear after text has been highlighted as a confirmation of the user's selection.

Users were not clear on what kind of text would be valuable to highlight.

I added two screens to the onboarding process that provide an example of how to highlight and what kinds of text could be highlighted for cross-validation.


I learned how to guide design decisions using research insights.

When we faced a difficult decision between 2 project directions, we leveraged research methods (survey, storyboarding, speed dating) to make an informed decision.

Developing users' trust in an algorithm takes time.

Moving forward with this project, I would conduct further research and user testing to develop strategies to increase users' confidence in Sift's cross-validation algorithm. I would be interested in exploring the extent that greater transparency into algorithm workings may be necessary, as well as the length of time required for users to obtain meaningful results and build trust in Sift.