UX Project/Case Study 

The Jamba App

Problem Statement provided by Focus Brands

Our analytics show that the cart abandonment rate on our app is through the roof. How do we redesign the app to handle the problem?

My contribution in the team

contract.png

Project

Documentation

research-3.png

UX Research

answer.png

User Testing

ui.png

UX Design 

Context of the Problem

Jamba, formerly known as Jamba Juice, is a company that produces blended fruit and vegetable juices, smoothies and similar products. Jamba is owned by Focus Brands.

43%

Users complete an order

4.5%

App sessions lead to a purchase

jamba_logo.png

THE CHALLENGE

Screen Shot 2019-11-13 at 9.36.29 PM.png

We were required to understand the problem and identify the issues leading to cart abandonment. These problems could possibly range from confusing labels and poor information architecture - we were yet to find out!

Process for the Redesign

Jamba Process.jpg

Phase 1:

Researching Problem Space

Phase 2:

Testing with Wireframes 

Phase 4:

Testing Prototypes

Phase 3:

Testing with Mockups

Phase 1: Researching Problem Space

As one of the primary researchers on the project, I contributed in making the research plan and the phase distribution based on research on the existing design and the subsequent design iterations.

Research Methods

Screen Shot 2020-01-19 at 3.58.29 AM.png

Literature Review

(App Reviews and Digital Analytics*) 

App reviews.png

Why was this method used?

App reviews helped us to extrapolate the popular opinion around

the existing design and also identify the top areas that we could 

potentially work on.

What kind of data was collected?

Large amount of qualitative data by parsing through the user reviews & comments on Google Play & App Store. 

How was the data represented?

For app reviews, we generated insights based on each review and 

coded them according to broad themes to identify the top 10 pain points.

Top 10 pain points from app review analysis & the resulting insights

*For the analysis of the app's digital analytics report, we are bounded by the company's NDA to not disclose any analytical and statistical information.

Task Analysis

User Flow and Hierarchical Task Analysis (HTA)

Why were these methods used?

To visualize the architecture and the cognitive sequence of flow of information on the app. 

What kind of data was collected?

How the screens move, transition and the visual hierarchy of the system and an overview of sub tasks.  

In the form of user flow and HTA diagram designed using Adobe XD.

How was the data represented?

UserflowJamba.jpg

User flow chart for the Jamba application

HTA Jamba.jpg

HTA for Jamba App

Competitive Analysis

Desk Research, to understand features of other brands and identify gaps and opportunity in the research.

Why was this method used?

What kind of data was collected?

Additional features, ideas, design decisions and validation for our research findings.

How was the data represented?

In the form of features comparison tables presented below.

CA logos.png
CA - General.png

General Category Analysis

CA - Onboarding.png

Onboarding Category Analysis

CA - Ordering .png

Ordering Category Analysis

CA - Checkout.png

Checkout Category Analysis

Online Survey

To quickly gain quantitative data based on order-ahead

app usage.

Why was this method used?

Quantitative data from 53 responses.

What kind of data was collected?

DISTRIBUTION PLAN

FACEBOOK

WHATSAPP

In the form of pie charts, tables and graphs, etc. 

How was the data represented?

SLACK

REDDIT

SURVEY DESIGN

SURVEY DESIGN.png

Survey design 

Each question on the survey had response options that best represented the rationale behind the question. The table below lists down the response options and the rationale behind every question on the survey section wise. The full version of the table can be seen after clicking on it.

survey table.png

Click on the table above to read the full table about every question on the survey

Field Studies

Why was this method used?

To see the reaction of people and get an unbiased view of the user in the user’s own context.

What kind of data was collected?

Insights that helped us in validating our findings by observing the users in the context. 

How was the data represented?

In the form of an observation table with notes and interpretation for every note.

Location: Jamba Store in Tech Square, Atlanta.

I followed the Fly on the Wall technique, where I sat in one corner of the store and observed 4 people and took notes. 

Screen Shot 2019-11-17 at 3.51.55 AM.png
obs table.png

Observation session in

progress at Jamba Store

Contextual Inquiry 

3 circles.png

Why was this method used?

To acquire feedback on the existing application directly from the users in context.

What kind of data was collected?

Qualitative and quantitative data on the usability of the app by observing users perform a given task.

How was the data represented?

In the form of answers to interviews questions and SUS scores after the task. I interviewed 2 out of the 7 interviewees.

Application Walkthrough

We used the Think Aloud protocol to capture the behaviors, thoughts, attitudes, likes and dislikes of the users.

System Usability Scale 

We then used quantitative measure of the app's usability to get reliable and valid results.

Semi-Structured Interview

Finally, we followed up with questions to users about the walkthrough & SUS response. 

youtube.png
sus.png
Screen Shot 2020-01-03 at 5.29.51 PM.png

* System Usability Scale consists of 10 questions on industry standard for usability testing; 68 is the national US average.

SUS* Score Comparison

Personas

At the end of the research phase, to get overview and synthesize our research findings, we created 2 distinct personas.

Screen Shot 2020-01-19 at 3.25.06 AM.png
Screen Shot 2020-01-19 at 3.25.31 AM.png

Empathy Maps

Screen Shot 2020-01-19 at 3.25.58 AM.png
Screen Shot 2020-01-19 at 3.26.25 AM.png

Journey Maps

Screen Shot 2020-01-19 at 3.26.49 AM.png
Screen Shot 2020-01-19 at 3.27.10 AM.png

Research Findings from Phase 1

To consolidate our findings from all the research methods, I used my background in Computer Science to introduce the concept of "Clustering", which essentially involves visualizing the issues on a spatial area to figure out where the density is the maximum, based on the area that the sticky notes are most clustered in. 

System Usability Scale 

Quantitative measure of the app's usability to get reliable and valid results.

Jmaba miro.png
Screen Shot 2019-11-19 at 2.46.23 AM.png

Thematic Analysis of Data

issues.png
Screen Shot 2020-01-03 at 5.53.57 PM.png

Phase 2: Wireframing

To understand the division of the feedback from the contextual inquiry process, we divided the whole process into 3 parts: Sign up, Ordering and Checkout.

Screen Shot 2020-01-04 at 7.25.28 PM.png

Design | Sign up

Screen Shot 2020-01-04 at 6.46.52 PM.png

Jamba App

New Wireframe

Design | Menu View

Screen Shot 2020-01-04 at 6.48.48 PM.png

Jamba App

New Wireframe

Design | Menu View

Design | Order Status

Screen Shot 2020-01-04 at 7.04.58 PM.png
Screen Shot 2020-01-04 at 7.06.18 PM.png

Jamba App

New Wireframe

New Wireframe

Jamba App

SUS Score Comparison

(After testing the Wireframes with users by Contextual Inquiry)

Screen Shot 2020-01-04 at 7.19.52 PM.png

*We did not ask the users to fill SUS for the wireframes, since they are essentially skeletal designs, hence the score remains unavailable.

Phase 3: Mockups

We tested our wireframes with users using Contextual Inquiry method where we gave them 3 representative tasks in each category (Sign-up, Ordering and Checkout) and after the cognitive walkthrough and after the task, we conducted semi structured interview with the users related to the experience. 

Wireframes

3 circles.png

Contextual Inquiry

Screen Shot 2020-01-04 at 7.43.48 PM.png

Design | Sign up

Screen Shot 2020-01-05 at 4.58.06 PM.png

Design | Review Order

Screen Shot 2020-01-05 at 4.58.46 PM.png

Design | Customizations

Design | Track Order

Screen Shot 2020-01-05 at 4.58.27 PM.png
Screen Shot 2020-01-05 at 4.59.05 PM.png
Screen Shot 2020-01-05 at 8.08.25 PM.png

SUS Score Comparison

Phase 4: Prototypes and Final Evaluation

We tested our wireframes with users using Contextual Inquiry method where we gave them 3 representative tasks in each category (Sign-up, Ordering and Checkout) and after the cognitive walkthrough and after the task, we conducted semi structured interview with the users related to the experience. 

Mock ups

3 circles.png
Screen Shot 2020-01-05 at 10.05.43 PM.pn

Contextual Inquiry

Final SUS Score Comparison

Screen Shot 2020-01-06 at 12.21.18 AM.pn

Our final prototype shows a usability score increase of 53% from 60.71 to 92.92 

Evaluation of Final Prototype

Heuristic Evaluation with 3 experts

Screen Shot 2020-01-06 at 12.44.57 AM.pn
Screen Shot 2020-01-06 at 12.45.20 AM.pn
Screen Shot 2020-01-06 at 12.45.38 AM.pn

Dr. Bruce Walker

Assistant Director, Psychology Coordinator

MS-HCI program at Georgia Tech 

Dr. Richard Henneman

 Director, 

MS-HCI program at Georgia Tech 

Dr. Carrie Bruce

 Research Director, 

MS-HCI program at Georgia Tech 

Unmoderated Online Testing at UserTesting.com

We created elaborate tests for remote users to test our prototype. I took the lead in designing the tests for remote users.

Screen Shot 2020-01-06 at 1.03.19 AM.png

(Click on the video for a sample of the testing)

In-person

User Testing at a Jamba Store

We intercepted users at the store and made them do a comparative testing between our prototype and the original app.

Screen Shot 2020-01-06 at 1.02.43 AM.png

Final Takeaways from the Testing

homework.png

Users spent 29.7% less time finishing each task on our prototype 

save.png

5/6 users during the testing preferred the customization 

visibility on our prototype

To view this project in the form of a slide deck, please visit the link below for the detailed representation - 
 

Click here for the detailed presentation 

Problem Statement

How can we improve the Jamba application’s customer journey for new users in order to reduce cart abandonment and increase sales?