Knowledge Base
enabling users to access a modern Help Center solution for a self-service
- ROLE
UX Researcher - TIMELINE
April 2024 - August 2024 - SKILLS
User Research
Interaction design
Information Design - TOOLS
Figma
UserZoom
PROJECT OVERVIEW
The existing knowledge base of Tradebyte contains all the necessary information for operating the software for 1,500 brand users, but it does not allow users to find and utilize help articles required to resolve cases quickly.
Improving the usability of the knowledge base was one of the company's strategic projects. Additionally, with the portal's redesign, many clients can solve their tasks independently, reducing response times and allowing the customer support service to focus on incidents that require their involvement.
As a UX Researcher, I worked with the Communication team, which produced help materials, and with the Product Managers and Marketing team. The project's business owner was the Head of CX, who approved the research plan and made decisions regarding the findings and recommendations.
Improving the usability of the knowledge base was one of the company's strategic projects. Additionally, with the portal's redesign, many clients can solve their tasks independently, reducing response times and allowing the customer support service to focus on incidents that require their involvement.
As a UX Researcher, I worked with the Communication team, which produced help materials, and with the Product Managers and Marketing team. The project's business owner was the Head of CX, who approved the research plan and made decisions regarding the findings and recommendations.
01
MOTIVATION
What’s the background?
Tradebyte is a software company that specialises in providing solutions for managing and distributing e-commerce product data.
Their flagship product, TB.One is a platform that enables brands, retailers, and distributors to efficiently manage and distribute their product catalogues across multiple online marketplaces and e-commerce platforms.
The current application design was created around 2017 and has only seen minor changes since then, even as the number of features has significantly increased.
Different naming, nonlinear user flow, and a back-end-oriented design led to users encountering issues that could not be resolved, resulting in revenue losses for each non-trading day.
Their flagship product, TB.One is a platform that enables brands, retailers, and distributors to efficiently manage and distribute their product catalogues across multiple online marketplaces and e-commerce platforms.
The current application design was created around 2017 and has only seen minor changes since then, even as the number of features has significantly increased.
Different naming, nonlinear user flow, and a back-end-oriented design led to users encountering issues that could not be resolved, resulting in revenue losses for each non-trading day.
Picture: corporate website
| But, TB.One users searching for help in the current Knowledge Base solution often failed to resolve their cases due to poor findability, limited accessibility, and inefficient information structuring. A brief analysis reveals that 80% of incoming cases to Customer Service could potentially be resolved independently. |
PROJECT RESTRICTION: the final Help Center portal must be integrated into the Salesforce solution due to business needs. These limitations included customisation and design flexibility constraints and compatibility issues with specific content formats.
02
UNDERSTAND
I audited five different competitors of Tradebyte to understand existing trends and Style Guidelines of Google, IBM.
It begun with UX Research
I recreated Customer Journey Map based on 1500+ messages in CRM (Salesforce) and Google Analytics current Help Center Solution (Wordpress)
2/ Competitors Analysis
1/ CJM Review
To understand the initial state of user experience from getting an issue to a final resolution, I reviewed and summarised 1500+ customer comments and visualised them on the Customer Journey Map. I used anonymised data in ChatGPT to measure emotional feedback from customers and the readability of instructions they receive.
The new portal format required meeting each user segment's needs without creating new issues. I carefully selected clients to gather information on each potential cohort to achieve this. The selection was based on the following metrics: problem resolution speed for similar issues, communication language (native English speakers, non-native English speakers, German-native speakers), and company size (small brands with a heavy workload on the commercial manager or large companies with their own IT department involved in problem-solving).
The new portal format required meeting each user segment's needs without creating new issues. I carefully selected clients to gather information on each potential cohort to achieve this. The selection was based on the following metrics: problem resolution speed for similar issues, communication language (native English speakers, non-native English speakers, German-native speakers), and company size (small brands with a heavy workload on the commercial manager or large companies with their own IT department involved in problem-solving).
The picture below shows the swimlanes and principles I used to analyse cases. Unfortunately, the whole report contains unsharable information.
As a result of the initial analysis, I identified three types of events that trigger users to start searching for help, along with the main distractive points.
- 01/ Technical Terms
Most help articles were written using special terms, when Service Managers translates them in a more simple language for users - 02/Confused Pathways
The access to Knowledge portal was hidden in menu, creating a user case was easier - 03/Articles Accessibility
As many of instructions were written in descriptive manner, users have to search for information searching different pages & documents - 04/Info
findability
All help materials were structured in several pdf files contains 50+ pages each, users have to open several of them to find an answer
To understand user patterns of looking for help articles, I visualised user flow using collected footprints in Google Analytics and Moto that were confirmed with product managers.
COMPETITIVE ANALYSIS
Next, I sought to better understand the current state of the Knowledge Base by conducting a competitive audit of five b2b integrators and one e-commerce application (Shopify). Despite the use of different approaches in application design, I identified key principles of content design and information taxonomy.
I mapped the Customer Journey Map (CJM) review results against the Knowledge Base principles to identify gaps and opportunities for improvement. This comparison allowed me to ensure that the user experience within the Knowledge Base aligns with the market content design and information taxonomy principles, ultimately enhancing the overall usability and effectiveness of the platform.
I mapped the Customer Journey Map (CJM) review results against the Knowledge Base principles to identify gaps and opportunities for improvement. This comparison allowed me to ensure that the user experience within the Knowledge Base aligns with the market content design and information taxonomy principles, ultimately enhancing the overall usability and effectiveness of the platform.
Clarity and Simplicity:
- Ensure content is clear, concise, and easy to understand.
- Use simple language and avoid jargon unless necessary, providing explanations when it is used.
- Write content in a way that resonates with the user's needs and vocabulary.
- Use the language your users are familiar with, reflecting their terminology and context.
- Focus on creating content that helps users take specific actions or solve problems.
- Include step-by-step instructions, checklists, or visual aids to guide users through processes.
- Maintain a consistent tone and style across all content to create a cohesive user experience.
- Establish guidelines for voice, tone, and terminology to ensure uniformity.
- Use headings, bullet points, and short paragraphs to make content scannable.
- Emphasize important information with bolding or highlighting, and use images or diagrams where applicable.
Logical Structure:
- Organize content into a clear, logical hierarchy that reflects how users think about and navigate information.
- Group related topics together and use categories and subcategories that make sense to users.
- Ensure that the structure of the knowledge base is intuitive and easy to navigate.
- Use breadcrumb trails, a clear menu, and search functionality to help users find information quickly.
- Design the taxonomy to accommodate future content growth without becoming cluttered or overwhelming.
- Allow for easy expansion of categories and subcategories as new content is added.
- Implement cross-referencing within the content to connect related topics and resources.
- Use hyperlinks or “related articles” sections to help users explore related information without leaving the current page.
- Ensure that the taxonomy and content are optimized for search, both within the knowledge base and via external search engines.
- Use keywords and tags that align with user search behavior to improve discoverability.
As the final stage of the initial analysis, I prepared a series of user interviews to validate observed issues and gather direct feedback on findability, accessibility, and information clarity, forming the basis for recommending targeted improvements.
Overall, I had several meetings with five users (Commercial Managers) from different types of segments (company size, language, history of interaction with the company).
The interview script was structured into three stages:
The interview script was structured into three stages:
- 10-minute warm-up to introduce the topic;
- 20-minute discussion focused on potential new entries for the Knowledge Base;
- 20-minute session to explore possible category naming for the information architecture (IA);
The research concludes that,
The pain points mainly focus on low accessibility, structuring information feature-based instead of process-based, and inconsistent content design.
It generated continued issues with user frustration, longer resolution times, and decreased overall efficiency.
The pain points mainly focus on low accessibility, structuring information feature-based instead of process-based, and inconsistent content design.
It generated continued issues with user frustration, longer resolution times, and decreased overall efficiency.
We decided to address all these pain points with the following improvements:
01/Enhanced Accessibility:
more intuitive navigation structure and search functionality
02/Process-Based Information Structuring:
reorganizing the content around user workflows and processes
03/Consistent Content Design:
standardizing content formats, tone, and style
01/Enhanced Accessibility:
more intuitive navigation structure and search functionality
02/Process-Based Information Structuring:
reorganizing the content around user workflows and processes
03/Consistent Content Design:
standardizing content formats, tone, and style
03
IDENTIFY & SYNTHESIZE
Identifying user entry points:
🚀 A brand-new button leads directly to our Knowledge Portal, making accessing valuable resources and information easier than ever
🔘 The new button on TB.One’s front page uses a symbol you'll recognise from many popular systems, ensuring a seamless and intuitive experience as you access the Knowledge Portal
Identifying the taxonomy principles:
🔍 Easy Access: The search button Knowledge Portal can now quickly reach articles
📂 The new information architecture allows users to reach articles in just tree clicks, making navigation smoother and more efficient
Identifying the content design principles:
🔗 Articles are structured in a process-based format, allowing users to follow instructions from start to finish effortlessly
👁️ The new scanning-friendly design ensures that key information is easily visible. With clear headings, bullet points, and concise content, users can quickly find what they need without getting lost in the details
🚀 A brand-new button leads directly to our Knowledge Portal, making accessing valuable resources and information easier than ever
🔘 The new button on TB.One’s front page uses a symbol you'll recognise from many popular systems, ensuring a seamless and intuitive experience as you access the Knowledge Portal
Identifying the taxonomy principles:
🔍 Easy Access: The search button Knowledge Portal can now quickly reach articles
📂 The new information architecture allows users to reach articles in just tree clicks, making navigation smoother and more efficient
Identifying the content design principles:
🔗 Articles are structured in a process-based format, allowing users to follow instructions from start to finish effortlessly
👁️ The new scanning-friendly design ensures that key information is easily visible. With clear headings, bullet points, and concise content, users can quickly find what they need without getting lost in the details
04
IDEATION
How I designed a work around it
1
New Taxonomy
Design new type of structuring content, tested it with internal and external users
2
Content Design
Design major articles applying the best practices, change narrative and decrease the complexity of the text
3
Usability Test
Combinig two pieces together to study the concept with real users
New taxonomy: Tree test
To improve the usability and efficiency of the Knowledge Base, we launched an in-depth UX research initiative focused on identifying a new taxonomy.
Tree test research: round #1
Based on previous research, we identified ten user cases encompassing processes such as onboarding and setup, daily operations, data analysis, and reporting. Leveraging this data and our internal expertise, we selected 50 customer questions representing the most common queries and potential articles. This selection formed the initial taxonomy for the tree test.
Before introducing the new taxonomy to customers, I conducted an internal pilot test with the following considerations: First, we decided to adopt a more user-centric language in the taxonomy, which would necessitate a re-learning period. To avoid premature external feedback and engage internal stakeholders, we chose to gather initial feedback internally. Second, as this was the first instance of involving company customers in user research, we aimed to minimize the novice effect by conducting a single study of this type to ensure focused and actionable insights.
Before introducing the new taxonomy to customers, I conducted an internal pilot test with the following considerations: First, we decided to adopt a more user-centric language in the taxonomy, which would necessitate a re-learning period. To avoid premature external feedback and engage internal stakeholders, we chose to gather initial feedback internally. Second, as this was the first instance of involving company customers in user research, we aimed to minimize the novice effect by conducting a single study of this type to ensure focused and actionable insights.
- Internal Users
Product teams, Customer Support, and Success Managers - 15 questions
formed from users requests and using their terms - 50 results
conducted for relevant results - 7 days
study duration
Results
Success Rates were different from 43% to 80%
Additionally, I conducted feedback from some users with reccomendations how to form questions better anf their suggestions. The concern regarding swifting from technical to user-centric language was confirmed with a moderate comments.
Tree test research: round #2
🧭 The test results revealed a significant challenge with the taxonomy's directness: users encountered looping behavior. They would begin their search in one category, and upon failing to find the information, would move to another category. For instance, when users were tasked with updating a pricing policy for Marketplace#1, 70% initially searched in the "Pricing" category, while 30% went directly to the "Marketplaces" category. This issue underscored the need for a more direct and streamlined taxonomy to enhance information retrieval.
🛠️ I recommended either eliminating the "Marketplace" category or establishing multiple pathways. However, analysis showed that due to the core principles of the software, it was impossible to create straightforward help instructions without retaining the Marketplace section.
✅ Considering all the risks, we decided to implement multiple pathways to improve information access and navigation. To validate this approach, we conducted a second tree test to assess how well these new pathways enhanced user navigation and to identify any remaining issues.
🛠️ I recommended either eliminating the "Marketplace" category or establishing multiple pathways. However, analysis showed that due to the core principles of the software, it was impossible to create straightforward help instructions without retaining the Marketplace section.
✅ Considering all the risks, we decided to implement multiple pathways to improve information access and navigation. To validate this approach, we conducted a second tree test to assess how well these new pathways enhanced user navigation and to identify any remaining issues.
- Users
Commercial Managers and IT teams involved in resolving cases - 15 questions
with some adjustments formulated in Round#1 - 50 results
conducted, minimum value to get representative results - 3 weeks
study duration, it was run with open date till collect minimum amount of answers
Results
Success Rates were increased in comparison with the first try with a minimum of 52%
Enhancing the cross-referencing showed positive feedback in test results.
The current tree was accepted as the first MVP version.
The current tree was accepted as the first MVP version.
Content Design
We identified three key challenges in our content design initiative: reducing the use of technical terms, providing comprehensive end-to-end instructions, and redesigning the content to be more scannable.
By addressing these issues, our goal was to make the content more accessible and user-friendly, ensuring that users can easily understand and navigate the information, follow clear and complete instructions, and quickly locate relevant details through an improved, scannable layout.
Example of new structure
The final usiablity tests showed,
process-based information structuring and refined content design significantly enhance user experience.
I tested the overall concept with five different users aking them to find information on a new portal
process-based information structuring and refined content design significantly enhance user experience.
I tested the overall concept with five different users aking them to find information on a new portal
40% reduction in time spent searching for information
users could now follow logical, step-by-step pathways that mirrored their workflows
30% decrease in user errors and a 75% increase in overall satisfaction
optimized content design, characterized by consistent layouts and clear language
users could now follow logical, step-by-step pathways that mirrored their workflows
30% decrease in user errors and a 75% increase in overall satisfaction
optimized content design, characterized by consistent layouts and clear language
05
Summary & Reflection
Execution of the overall concept schedule in a Salesforce team backlog, as they need to create a technical solution. Communication team will be resposible in translating article landscape using Concept Guidelines, that were created and pusblished as an artefact oh the overall project.
During my exploration of the knowledge base concept, I appreciated how modern knowledge portals have evolved to incorporate dedicated forms and established patterns. These structures make it easier for users to access and navigate information, ensuring a more efficient and streamlined experience.
However, while these standardized knowledge bases are effective, my research during the project also highlighted the importance of considering alternative support methods, such as self-service approaches, in-app help, and guided assistance. Each of these methods offers unique benefits, catering to different user needs and contexts. For instance, self-service options empower users to resolve issues independently, while in-app help provides immediate, context-specific guidance. Guided help can be particularly valuable in complex scenarios where users need more step-by-step support.
In summary, while I value the established patterns in knowledge portals, I also recognize the need to integrate a variety of support methods to create a more comprehensive and user-centric knowledge base. By considering these alternatives, we can better address diverse user requirements and enhance the overall user experience.
However, while these standardized knowledge bases are effective, my research during the project also highlighted the importance of considering alternative support methods, such as self-service approaches, in-app help, and guided assistance. Each of these methods offers unique benefits, catering to different user needs and contexts. For instance, self-service options empower users to resolve issues independently, while in-app help provides immediate, context-specific guidance. Guided help can be particularly valuable in complex scenarios where users need more step-by-step support.
In summary, while I value the established patterns in knowledge portals, I also recognize the need to integrate a variety of support methods to create a more comprehensive and user-centric knowledge base. By considering these alternatives, we can better address diverse user requirements and enhance the overall user experience.
Natalia Mitina
UX Researcher
UX Researcher
