2019 MAR - 2020 MAR
End to end UX/UI design
The team wanted to redesign the existing search experience on Indeed for Japanese job seekers. Also, we wanted to refactor our infrastructure for the future scale.
In this project, I worked closely with PM and four engineers. I was responsible for the end to end UX and UI design, from defining the problem to deliver the final visual artifact.
This project had quite a few technical dependencies with other teams & products, we also needed to deal with many prioritizations to make sure we can deliver MVP on time.
We launched the MVP in six months that increased our overall metrics significantly, especially it increased the revenue by 12%. We also received a lot of positive feedback from users.
Indeed is one of the largest job search sites in the world, and it has launched its products to over 60 countries, Japan market is one of the high priority markets as the company strategy.
However, Indeed Japan site shared the codebase with the global Indeed site. It hasn't been optimized for the Japan market in terms of UI and UX since it has built.
Additionally, the global product teams kept rolling out new content and features worldwide, which may not be an optimal experience to Japan market, it also affected the Japanese sites in many aspects
Back in 2015, the company started investing in promoting Indeed Japan site to the local market, which brought a lot of traffics and brand awareness growing rapidly by the commercials. However, our no.1 user complaint in NPS reports is about the fact that our UI is generally outdated and not familiar/usable for Japanese users. This data is reiterated across other user feedback and competitive research.
The legacy Indeed Japan site looks too simple and unstructured which doesn't feel right to the Japanese job seeker, most of them have no idea how to start a search by using the legacy UI.
In the early phase of this project, we conducted user research to identify the core problems and get a sense of how job seekers perceive our product. The followings are what we learned:
Job seekers express that it feels like something is missing when they use Indeed. Compared to other sites it feels too plain & simple, and there’s a lack of information.
Job seekers don’t necessarily know the answer to “what(keyword)” or “where(location)”.
In general, job seekers don't feel our site is trustworthy because they don't see our site is supportive and legit in terms of what we presented.
We also worked with the vendor to conduct user interviews, then we consolidated the insights and came up with the Jobseeker journey map to help the team better understand the problems.
Job boards in Japan usually look rich, vibrant and informative which provides job seekers different ways to get start a search easily.
Indeed site is a keyword-based search since it has built, it helps job seekers to find relevant if they know exactly what are they looking for.
However, Japanese users are much more used to browse and filter experience as opposed to keyword search, and many of them have a hard time thinking about what they should put in the keyword boxes.
The information we displayed on the job card was unstructured and lack of visual hierarchy. It caused a lot of cognitive load to users and it was not scannable. They couldn't easily skim through the jobs and find out the one that interests them.
There were tons of essential elements in our site were missing. For example, photos and job-related filters are crucial for the job seeker, but they were not available on our site. Additionally, the overall look and feel of our site were not familiar to Japanese job seekers in comparison to other local job sites.
We spent entire two days sat together with the team(UX, PM, Dev, and QA) to talk about the goals, hypotheses, and constraints that helped to align everybody on the same page about what we want to achieve.
Since we want to build something we can fully-launch and scalable in the future, we dug into a lot of implementation details and how can we gradually roll out more features after we launch the MVP.
At the end of the design jam, we came up with a bunch of concrete ideas/wireframes that can help me go straight to the feasible design solution.
Launch the new search experience(MVP) that meets Japanese job seekers' expectations.
Improving the UI/visual design and show more structured information will help job seekers easier find jobs that fit them.
Introducing the tapping experience will help job seekers better express themselves, narrow down their search results easier, eventually they will see more relevant jobs and apply more.
Given that the goal is to launch the MVP in 6 months, they are a lot of works that need to be done. The team has to work collaboratively and prioritize the most important things to make sure we can deliver what we want on time.
As mentioned in the beginning, Indeed Japan shared the codebase with Indeed global, which makes it hard to make change drastically in terms of functionality. Apart from that, jobs in Japan need more metadata and filters which is not supported by our global back-end system.
After a few rounds of discussions as a team, we scoped out what's the feasible MVP to deliver on time. This launch is called “Rich Design”, which is about creating a modern UI with richer content that is much more usable and familiar to Japanese users, it includes:
This new experience starts with location and filter boxes which are easier to fill in for Japanese users and puts the keyword box at the very bottom for users who still like to use it, which will make the UI much more familiar and trustable for Japanese users. We also will have a location picker UI that allows the user to enter their desired location with a few taps without typing.
The MVP mostly focuses on upgrading filters from mere "refinements" on top of the existing search to a full-fledged filter box. The next steps will make sure we have more usable filters such as job tags, salary and job categories included in the filter box.
Our current job cards don't include much information, they are unstructured and very hard to parse. The new cards proposed here will provide more information on the card (specially job tags) as well as better structure, photos. Next steps will provide better labeling of Indeed specific information.
💡Why did we start with mobile web?
Given that 80% of our traffic came from mobile devices, and 82% of them came from the mobile web(Indeed app is around 18% ), and a lot of job seekers use smartphones as their first option to browse and search the job.
So the team decided to focus on mobile web redesign first, then bring those new changes to other platforms such as desktop and app once we see the positive signal from the experiments.
Even though we're designing for the Japanese market which is very unique, we still need to align with the global brand at the foundation level to make sure we have a consistent experience across devices, countries, and products.
Japanese job seeker used to see a lot of content on one page, we want to make sure the navigation is intuitive, and present the important function as the way they used to see.
Japanese job seekers read every piece of content carefully, the copies we provide need to be concise and clear, the wordings we're using should be consistent and understandable.
These are our components in our shared design library that allow designers in the team to use to keep the consistency.
Introduce the tapping experience for location search
For a very long time, people have to type in the location keyword manually if they want to search job on Indeed. This is not a common way to start a search for a Japanese job seeker. As one of the primary changes in the redesign, we'd like to introduce a more effortless way to start a search, which led us to select their desired location from the new location picker.
Users can click on the location picker to trigger the location pane, we created a linear flow to drill down their preferred location from region to district.
Due to the technical constraint, we were not able to include the station in the location picker for the MVP, but users still have the ability to type in the desired station as needed.
We made existing radius function more visible and intuitive to help users narrowing down their selected areas, after MVP launched, we will work on the station/train route search to help part-time job seekers find the relevant jobs easier.
Beyond search by keywords - Filtering experience
The filter is one of the most frequent requests from our users, it's also commonly used on other local job boards. Because keyword-search(where & what) is something unfamiliar to Japanese job seekers, they don't know how to express themselves by entering the keywords.
We used to have pretty limited filter options when started building MVP, later on, we added more and more options as the team developed a more robust data structure, that allows us to map those filters to the right job.
Filters help job seekers easily select the criteria that fit their preferences and conditions without draining their brains to come up with any terms to input the keyword box. Our goal is to provide various filters that help job seekers narrowing down the search result quickly.
Since jobs at Indeed cover all kinds of employment types, we wanted to make sure every job seeker who comes to Indeed can easily find the filters that relevant & useful for them.
We introduced dynamic filters which allow the specific type of job seeker (i.g. part-time job seeker) to find the customized options as soon as they choose the corresponding employment type.
Rich & structured information on job cards
We redesigned the job card structure and visuals to make it more readable and neat, we created a clearer hierarchy and more spaces to differentiate content based on their priority. Also, we added essential information such as photos, job tags, application-related labels to help job seekers have better sense to easily assess if the job suits them or not.
Since we have compact information on a tiny card, we wanted to let users anticipate what will be shown before the page loaded, we designed a skeleton card that will be displayed while the page is loading.
There are various jobs on Indeed, meaning we have to deal with jobs with a variety of information. We need to display that information in a meaningful and consistent format whether richness of card, I provided a unified design structure that can accommodate dynamic content. Here's an example:
A scannable job description
The job description is a crucial part of job searching, Japanese job seekers tend to walk through the job description carefully in order to judge if they're a good fit for that job.
Job descriptions at Indeed used to be unstructured since we aggregated jobs from a variety of sources, which made us really hard to make all of the job descriptions unified and consistent.
The team has been working on parsing jobs that we aggregated to our site, we wanted to extract the key information from jobs, then display them on top of the description in a meaningful and structured way to help job seekers easily skim through the job.
Since the jobs on Indeed are varied the team tried to parse the content and label the section content into the relevant category.
This semantic label approach will empower us to rearrange the sections of the job description and make it unified across all jobs at Indeed.
We created a "title & content pair" enhancer and apply this new rule to all jobs, which makes the jobs at Indeed look more consistent no matter where are the jobs coming from.
Clearer & inviting content
I worked closely with content strategist to ensure we use the consistent wording, voice and tone in every piece of product. Also, we tested variety of copies to validate which one can drive more job seekers to apply.
Since we have some Indeed-specific features/terms that sometimes might confuse new users, we want to explain our value proposition in a more clear way, so we reviewed all the critical and commonly used copies to be aligned across the products.
Unified saving job experience
Saving jobs is especially important for Japanese jobseekers, they save almost 2x more jobs than other markets, we've learned from previous user researches that people prefer to save job first when they found the interesting jobs, and they are actually more likely to apply if they save a particular job.
However, the currently save job experience at Indeed is kind of broken, the navigation is not clear and we inherited some features from Indeed global which might not very useful for Indeed Japan at this point.
We used to have many statuses of jobs which is pretty complicated and confusing.
The interaction of change job status was not intuitive, the visual hierarchy was unclear.
The UI between logged-in and logged-out is inconsistent.
The new entry point of the saved job become more visible and interactive, it's invisible by default and it shows up when user saves the job or scroll up the page.
Simplified saved job page helps job seekers to only focus on the most important status on the job searching phase.
Consistent UI between logged-in and logged-out to ensure users have a seamless experience.
We A/B tested each new feature to make sure we see a positive impact on business and user side, then merge them back to our core MVP foundation.
As a result, this new search experience redesign was very positive in terms of many aspects, it massively increased the critical business metrics as well as the user engagement, we've heard a lot of good feedback from users on App Store as well.
Overall, users who came to our site stayed longer, viewed more jobs, saved more jobs and eventually applied for more jobs. These numbers are a shred of compelling evidence that our users became more engaged happier with their job search experience at Indeed.
After we launched the rich design across the core page in Indeed Japan site, the foundation of the product became stronger that empower us to unblock more ideas we'd like to build & test, we kept working on optimizing the end to end user flow, conducting user research to know our users more, building more powerful features that help job seeker searching more efficient.
The following items are something on our plate, and we're ready to bring users more delightful experience in their job searching journey!
Left: Saved search experience | Middle: New apply flow | Right: Saved job comparison
It was a pretty intensive year that everybody in the team was working so hard for launching the MVP, we delivered a lot of good stuff, but we also struggled with many things.
For example, before the project started, the team kept discussing what's the best way to approach this big redesign? How can we plan a decent A/B testing that we can validate what we want to learn? How can we get buy-in from the leadership teams? What would be the next step if MVP failed? There's no absolute right answer but we need to embrace the expertise from each team member.
Although we've conducted user research and learned people are not satisfied with current experience, we still not 100% sure people will appreciate the new experience since we made drastic changes to our product. But our team had strong hypotheses that job seekers will be benefited from those changes as long as it helps them find jobs easier.
We talked everything in detail such as concerns, user experience, product strategy, because we really wanted to make this experiment successful, that faith pushed us to break through a lot of limitations.
Throughout this process, I understood that a huge leap might take high risks but come with a high impact, the gradual experiments might slow us down but relatively safer. How to find the balance among time, resources, constraints, and goal is really important, I learned that prioritization is very crucial for the team to speed up and focus on the right thing to do.
Lastly, everyone in the team should clearly understand the goal, plan, and problem to build a more meaningful product as a team. This project helped me broaden my thinking to be more micro-macro balanced and built a stronger connections with the team.
KAPI - Find your desired Cafe in a most efficient way.
READ CASE STUDY