Amazon Seller News
A web experience redesign of the Amazon news delivery system.
Jan 2018 - Jun 2018
User Research, Ideation, Prototyping,
Marketplace sellers are an important part of the Amazon experience, and China-based sellers count for 30% of total sellers on Amazon. Amazon wants to make sure they are communicating effectively with them, but they have really limited data about Chinese e-commerce markets and how communication strategies and services are conducted in China to China-based Amazon Sellers. Amazon News communications that are not designed for a China-based audience may erode trust in news sources. In this six-months capstone project, my team aimed to help Amazon envision a future communication system between China-based Sellers selling in U.S. and Amazon to improve mutual trust.
TODAY: Amazon Newsletter
TOMORROW: What could the future of Amazon News communications look like?
PROBLEM & PROCESS
Design solves a problem. In order to provide a solution, we first need to understand the problem. The initial design research question that Amazon gave us was very broad:
How to help Amazon craft trust-building communications with Chinese-speaking Sellers selling in U.S?
Through team discussion, we found it was hard to “craft” trust-building communications, since relationship between specific behaviors and improvement of trust is difficult to define with our limited project time and resources. However, finding out the factors that led to a decrease of trust is much more feasible. So we scoped down our design research question to:
How does Amazon current news delivery decrease Seller’s trust in Amazon?
By avoiding behaviors that may decrease Sellers’ trust in Amazon, we can therefore help Amazon envision a future communication system to improve trust between local Chinese Sellers and Amazon.
To achieve this goal, my team investigated the relationship between trust levels and news content through a comprehensive user centered research and design process.
1. Domain Research
- Analyzed delivered news
- Gathered metrics
- Researched social medias
- Paper sketches
- 3 hi-fi mockup iterations
- Interactive prototypes
11 interview sessions:
- 9 with China-based Sellers
- 2 with Seller Associates
- 3 Usability Testings
- 2 Amazon Co-design Sessions
2 workshops help define:
- Design direction
- Prototyping plan
Research Process Highlight: Interview
We conducted a total of eleven one-on-one semi-structured interview sessions, nine with China-based Sellers and two with Amazon Seller associates, to actively looking for the goals, information needs and pain points for China-based Amazon Sellers have using current Seller News system.
Amazon had limited data on China-based sellers because of low response rate from surveys and limited research resources. We used interviews as our main research method to understand China-based Sellers’ individual experiences as we knew that Chinese Sellers experience varies much from U.S. Sellers.
And talking to Amazon Seller associates allowed us more opportunities to learn about Sellers in another perspective. Associates behave as a hub of Chinese Sellers and we could obtain a lot more information from them.
With the rich data gained from our research and constant communication with Amazon Seller Team, I conducted affinity diagram to analyze the data and concluded the following three major challenges China-based Sellers have with current news delivery:
Current news fails to deliver scannable information.
Notifications are not sufficient and not on time.
Newsletter is not an efficient and effective way to deliver information.
Through our research, we also noticed that Sellers who are less experienced, new to Amazon or small Sellers tend to have more complaints about Amazon’s current news delivery. The previous research findings that Amazon provided to us also indicated that small Sellers had relatively lower trust rate to Amazon. So we scoped down our target users who fall in the following categories as our primary users.
Less experienced Sellers
New to Amazon
Less than // years overall selling experience
Tenure bands: D & E; // months - // months (Amazon Dataset)
Revenue bands: C, D & E; /// - /// dollars per year (Amazon Dataset)
After conducting a design ideation session with the team, I made prototypes with Sketch first and then used Marvel to make an interactive prototype for easier presentation and usability testing.
China based Amazon Sellers were not fond of the format of newsletter or email. I explored several ways to deliver Amazon News that is not based on email. Several solutions I came up with are:
1) Card on Seller Central Page.
3) Post news through Wechat platform.
4) Web/Mobile Application.
I sketched out the prototypes, evaluated the pros and cons with the team to see which one to move forward with, and finally decided to design a component on Amazon Seller Central for news delivery.
Below is a brief walk-through of the design:
I categorized news into four categories -- “policy change”, “Seller tips”, “marketing/recommend” and “others”. I made five tabs for the categories as filters so that Sellers can easily find the information they need.
Under each tabs, one article is shown on the page at a time, with a slider at the bottom of the page to let Sellers switch among the articles.
There is a label at the bottom right corner of each news, which gives Sellers an idea of what the news is about. And that addresses the problem of current news delivery is not scannable.
There is a refresh button on the upper right corner of the page. When Sellers are not interested in all the news on the page or they have read all the news already, they can read more articles by clicking the refresh button.
There is a “not interested” button at the bottom right of the page. If Sellers are not interested in the news, they can click on the button. Some mystic Amazon algorithm can help Sellers screen out the information they are not interested more accurately if Sellers use this function more frequently.
I made a notification icon on the upper right of the page. By clicking this icon, Sellers can see the notifications they received from Amazon. Compared to the existing notification function on the Seller Central page, this notification is not for specific Sellers. When something happens in the industry, Amazon can notify all the Sellers to reduce Sellers’ fear of unknown so that Sellers’ trust to Amazon will not be affected.
We organized and visualized our research findings through outlining:
Characteristics of each particular user type
Information that sellers need or desire for communicating with Amazon
How and why the sellers want the information
Usability testings and Findings
We recruited three China-based Sellers and conducted three remote usability testing sessions at Amazon Usability Lab. Our goal of usability testing session is to test on how our redesigned news delivery system solve the problem indicated from our research. Our study was mainly consisted of four tasks and was guided by the following three questions:
How efficient are sellers finding information?
Can users successfully find news that matches their criteria?
How effective is "Notification" in delivering notifications?
How the Amazon news section work in the big picture of Amazon Seller Central?
Users are not able to retrieve past news.
Two out of three participants indicated their needs of findings and reading past news sent from Amazon. The information category they mainly wanted to find from the past is policy updates since a lot of policy changes are time based. Policies sometimes can be effective several months later than the policy is announced to Sellers.
Also, users can only use the slide bar at the bottom to slide left and right for pieces of news once at a time. It is difficult for users to find a specific piece of article they want.
“Not Interested” is hard to understand.
“Not Interested” button was hard for participants to find. Also, all of the participants were confused with what will happen if they click on it. They did not understand the purpose of “Not Interested” that Amazon Seller Central will contextualize based on their interests to news.
After explanation, all of the participants indicated that they will not click on the button because they believe that the information from Amazon is useful. They are afraid that once they click on it, they will have a risk of losing important information. On the other hand, two out of three sellers stated that they will click on “interested” to have Amazon push more related articles to them instead.
Newly released news is not noticeable.
Another problem we found is not being able to notify Sellers of newly released news. Our news section is initially designed to present Sellers with updated news on the first several pages before the other old ones. However, we are having troubles of letting users know that there are important and updated information for them. There is currently not a noticeable approach to inform users that there are important information being released.
Prototype 2.0 -- Final Design
Takeaways & Next Steps
What was the biggest challenge and what did I learn from it?
The biggest challenge for me was to keep track of the main goal of the project which was "trust". It was a big scope and it was hard for me to validate how my design decisions hit this goal. For this project, I learned to break trust apart into things that I can test with such as click rate on the news section and frequency of contacting seller support. For future projects, when I am given a broad goal, it's important to break goals into specific, measurable, actionable, relevant and trackable pieces before making other progress so that I don't get lost through the project.
What is the next steps of this project?
Content is important.
During the co-design session, one of the Amazon designers indicated that they valued the content the most. In the end, it was solid information who delivered to the Seller's that interests them the most. The format of delivering the content before, which was newsletters, was basic. Newsletters was simply consisted of title, subtitles, and text. However, we had limited ability to respond to this feedback at that time because we knew little about content strategy and we could not write articles for Amazon to send out. We kept in mind that the content matters the most to Amazon and if there is future research, it will be worth digging deeper about pain points and needs of Sellers about content.
Validating solutions with older, more experienced and bigger sellers.
While scoping down the user group, we focused on sellers who are new to Amazon, less experienced and with small business. I would like to gather more data from other user groups to see if they have different needs, and if the proposed solution still apply to them.