Project Overview

This project focuses on the UX improvement of the key task flow of updating goals on the BetterWorks platform. It is a designer-led project without a PM. I collaborated with another designer throughout the entire  design process including user research, problem definition, design exploration, prototype testing, design iterations and more. Moreover, we engaged stakeholders and the executive team in the discussion process.

Duration

2 months

Role

UX/UI designer

Tool

Paper sketch, Sketch 3, InVision

Deliverables

Paper prototypes, Interactive prototype, Hi-Fi Mocks, SPEC, Project summary report

Project background

BetterWorks is a cloud-based human capital management software that allows employees to create Objectives and Key Results (OKRs) and link them to top company goals. Managers and employees establish goals in BetterWorks and then track them over a period to make an organization more open and transparent. Everyone can see what’s happening with goals and leave comments for communication across the organization.

Goal progress update and commenting is the key task in the Goals module. Most users check-in on progress on a weekly or monthly basis. Today, the ambiguous definition of the update concept and the inconsistent UIs can confuse users.

The goal of this project is to improve the experience of the key task flow to be more clear, intuitive, efficient, and consistent. Thus, increasing the usage and reducing the number of relevant support tickets and the complaints in NPS.

Research for understanding the problems

We have multiple channels to receive user feedback. In our Slack channel, we gather insights from the cross-functional teams at BetterWorks, who share their comments on their own experience as well as transfer the feedback from the customers to us. In addition, we also label and summarize complaints via Zendesk and NPS vendors as the direct feedback channel. For this project, we collected all the mentions related to the progress updating and commenting experience. In addition, I conducted 1:1 interview sessions to understand our users' current workflows and pain points.

BW_1 Research@2xBW_1 Research@2x

Aside from qualitative research, we also did audited and analyzed data on web and mobile clients. The main takeaway is that there were multiple inconsistent ways for user to complete the progress update task, as shown below:

BW_2 Research@2xBW_2 Research@2x

Problem analysis and definition

According to the research, we categorized the problems into 3 types:

BW_4 Problem@2xBW_4 Problem@2x
BW_3 Problem@2BW_3 Problem@2

Problem-solving design process

Cornerstone Problem: Related to our core goal terminology and philosophy

For “Update” definition, previously, users only can submit metric number update with the contextual comment or only submit metric number update. They can not only submit a contextual comment without metric number update. But they can do that on the right panel.

We redefined the concept of the “Update”, which is following the OKRs best practice. Users can update in three ways in one place.

BEFORE

BW_5 Defination Copy@2xBW_5 Defination Copy@2x

New Definition

BW_5 Defination@2xBW_5 Defination@2x

Quick fix problems: Can be explored independently

For navigation, it is showing the shadow for row selected state now. It is not noticeable and the shadow on the list since it implies a nested or parent-children structure. So we used consistent visual and comment pattern for making selected state obvious.

For consistency, it shows different UIs for different types of milestone and goals. Today, users can’t submit contextual comment for the type of binary or integrated milestone. So we created the component of updating UIs.

For digesting info, it shows information pretty busy on comment input field and wastes many places of comment history on the right panel currently. We cleaned up suggestion hashtags and comment feeds for digesting info more efficient.

BEFORE

BW_12 Problem 3@2xBW_12 Problem 3@2x

AFTER

BW_12 Problem 2@2xBW_12 Problem 2@2x

Dependent problems: Need to be explored holistically

Design explorations

Based on the problems in the analysis, we explored the proposals respectively. Moreover, since some problems are highly dependent, we also consolidated the ideas into the integrated design directions. Here are the four paper prototypes for the direction validation.

BEFORE

BW_6 Direction 3@2xBW_6 Direction 3@2x

PROPOSALS: 4 Directions

BW_6 Direction@2xBW_6 Direction@2x
BW_7 Direction@2xBW_7 Direction@2x

Round 1 paper prototype user testing

We conducted the user testing with the paper prototypes both internally and externally including an onsite session at Evernote. Seven Users went through the same tasks of the goal progress updating and commenting with each prototypes. During the process, we moderated the conversations to help them think aloud.  At the end, they provided their preference order of the four directions. In term of results, most users preferred direction A and C.

BW_8 Round1 testing@2xBW_8 Round1 testing@2x

Iteration: 2 directions

As the learning from this round of user testing, users preferred A and C because of having only one place to do one thing. However, we also recognized two types of users. One type is that the user preferred inline lightweight experience, and only want to see comment history when necessary. The another type is that the user like update one thing with comment history at a time. The reason why they like A or C.

Initially, I planned to reduce the entry points by removing the existing bulk “Update progress” page. Since only 10% of users visited the  “Update Progress” page according to the data I foundTo my surprise, users really like this flat list with lightweight experience and declared the expectation of not taking it away. It services users who want to easily update multiple milestones at once. So we combined the bulk update page into direction C. Meanwhile, we iterated A and C to Hi-Fi prototypes for round 2 user testing.

BW_9 Round 2@2xBW_9 Round 2@2x

Round 2 Hi-Fi prototype user testing

We conducted 50 minutes 1:1 onsite session with 6 participants from INTUIT in one day. They have different roles, such as an individual contributor, a manager, or an HR. They have different experiences and use flows  using BetterWorks platform as well.

The session was split into 2 parts: 1. Key task flow walkthrough 2. Comparison analysis on four specific areas of the design:

        Entry point: Inline  field VS “Update” button

        Progress update: Inline update VS Panel update

        Comment: Action icon VS Expose comment column

        Bulk update: Same page experience VS New page experience

 

BW_10 Round2 Testing@2xBW_10 Round2 Testing@2x
BW_11 Round2 testing@2xBW_11 Round2 testing@2x

In the Round 2 user testing, most users preferred the direction C. They consider it more intuitive and discoverable than today so that they easily find the starting point and complete the task efficiently.

 

Finalize the holistic proposal and deliver

Based on direction C, we did some finishing touches and evolved it into the final proposal. Additionally, we summarized the design process with the design rationale to share with the CS and the ENG team for gathering the final feedback. Along the way, we also discussed with PM and engineer on the estimation of the cost and adjusted the scope of MVP accordingly.

BW_13 final@2xBW_13 final@2x

What I’ve learned

The final proposal looks simple while the design process is not simple. We have faced different kinds of constraints and the intertwined dependency. Here is the summary of what I’ve learned:

  • Brainstorm more ideas in different directions first instead of generating only various layouts of the same idea.

  • Think holistically.  An integrated proposal can possibly satisfy multiple use cases well.

  • Make a comparison of the specific areas across different proposals. It helps dive deep into users’ rationale behind their overall choice.