DataRobot | Logo | White | Full
Wawe
Docs portal reduces DataRobot team's workload
December 2022  •  17 min read
Problem
Non-technical users often struggle to find clear, accessible resources to understand and effectively use DataRobot's workflows, features, and AI capabilities. This lack of centralized, user-friendly documentation leads to confusion, increased reliance on the support team, and slower adoption of the product.
Business goals
Reduce the workload on the support team by encouraging self-service, especially for non-technical users.
Centralize all knowledge assets into an easily navigable, public-facing hub.
Build trust and commitment by simplifying complex concepts for all users.
Enhance product adoption by making resources accessible to a wider audience.
Increase customer satisfaction through seamless access to information.
User Goals
Access clear and concise resources to confidently navigate the platform.
Quickly find answers without needing extensive technical knowledge or support intervention.
Learn how to solve business problems using AI with step-by-step guidance.
Gain an understanding of DataRobot workflows and features in a way that’s approachable and actionable.
Result
As a result, at least 70% of tasks are now being completed by users with non-technical backgrounds.
Company
DataRobot
ROLE
Product Designer
Product Strategy, Design Vision, UX Research, User Flow, Information Architecture, Wireframing, Prototyping & Testing, Visual Design
Team
Product Manager, Product Designer, Dev-team (Frontend & Backend), Data Manager, Customer Representative Manager
Project DURATION
March 2021 - August 2021
One step closer to self-service
DataRobot is the AI Cloud leader, committed to delivering a unified platform for all users, data types, and environments, aiming to accelerate AI deployment for every organization.


Embracing the self-service approach, it was crucial to ensure seamless access for all users, allowing them to easily learn how the DataRobot platform can solve their business challenges.
Learn more about DataRobot.
DataRobot is the AI Cloud leader, enabling users to:
• Build practical machine learning models
• Deploy to a multi-cloud environment
• Monitor performance

3.42B+

total models built

2.5M

built models daily

6.3B

valuation

# 26

the world’s best cloud company by Forbes

Augmented Intelligence scheme
Dynamic methodology
My process at DataRobot is based on the Double Diamond Model and Lean UX process. I incorporate the key phases into the project.
DR_DD
Understanding the problem
To create a documentation portal that effectively serves users, I needed a clear understanding of who would benefit from it.
My research focused on:
• Identifying user goals and needs.
• Uncovering pain points in the existing user journey.
• Measuring task success.
20 surveys
I distributed over 20 surveys to better understand our customers, using SurveyMonkey to collect quantitative data efficiently.
12 users
12 Users participated in user interviews through the process to help me dig deeper into motivations, needs, goals and pain points
Gathering insights
After collecting recordings from the user interviews, we completed affinity mapping with my teammate to synthesize the pains identified. We grouped these problems under common themes and features in the platform.
Feedback
Analyzing preferences
Based on interviews conducted with users, we found additional insights.
71%
of users feel that the explanations of the features and functions usage could help a lot.
49%
of users consider DataRobot as an advanced app, so for achieving goals they would need additional information and theoretical knowledge.
86%
of users want to use search to find results relevant to specific topic.
+1 persona
Based on insights from customer support requests, user interviews, and surveys, I identified persona(key roles) through affinity mapping and combined with JTBD framework.
The Insight Seeker
An Entrepreneur, Consultant, Manager, or Analyst who is focused on growing their business and relies on actionable insights. She/he values tools that are intuitive and help them easily find the right information to achieve desired results without needing to build complex models.
JOBS-TO-BE-DONE:
When I use the app, I want it to guide me to the most relevant insights quickly, so I can confidently make data-driven decisions and focus on my business priorities.
DataRobot | Persona 1
Summarize requirements
With all gathered information, I defined the documentation needs to address audience requirements and align with the company's long-term strategy.

Business

Scope. Define the project scope and business objectives.
Personalization. Customize content based on user roles and behavior.
Feedback & Analytics. Collect feedback and analytics, adhering to DataRobot policies to enhance user experience and trust.

User

Goals. Ensure content meets user goals and expectations.
Inputs & Outputs. Provide clear, concise content with practical examples.
Search Functionality. Implement a convenient search feature that aligns with user queries.
Consistency. Maintain a recognizable, consistent style that is easy to read and navigate.

Technical

Functional. Technical documentation is kept close to source code, docs-as-code philosophy.
Non-Functional. Generation of technical documentation.
Documentation Philosophy. Adhere to a docs-as-code approach, keeping technical documentation close to the source code and facilitating its maintenance.
Design principles
Provide Guidance
Provide guidance for complex process and ideas
Give Flexibility
Every models varies in user’s needs and knowledges
Build Confidence
Build confidence by showing value and power of the platform
Act as an Advocate
Be a fair partner and advocate for customer
Be Transparent
Communicate clearly and truthfully to owners.
Keep it Simple
Optimize for easy understanding

Unified platform for all assets

Advantages:

Increase Customer Satisfaction:
Enhance trust and commitment among users.
Advance Self-Service:
Support the transition to self-service tools.
Enhance Product Offering:
Strengthen the DataRobot product package.
Expand Community:
Grow and promote the DataRobot community.
DataRobot Assets

Image: Depicts the transition from three separate development tools to a unified documentation portal.

Early designs
The Docs DataRobot Portal wireframes outline the structure and functionality of key pages, ensuring an intuitive experience for users seeking guidance and technical resources.
Entry point / Main page
The main page serves as the starting point for users navigating the portal. It provides quick access to essential resources and guidance.
DR | Wireframe | Main page
Article page
The article page presents documentation in an easy-to-digest format, helping users find and understand information efficiently.
DR | Wireframe | Page
Code snippets
Designed to enhance the developer experience, the side modal allows users to interact with and adapt code examples seamlessly.
DR | Wireframe | Code Snippet
Wireframing the solution

Scope:

Easy Access:
Ensure necessary information is readily accessible from the app.
Step-by-Step Learning:
Provide a clear learning path for users.
Feature-Centric Search:
Implement a search function focused on features.
Simple Navigation:
Design straightforward, structured navigation across sections.
Task-Based Grouping:
Organize documentation by workflow and task.
Docs-as-Code:
Maintain documentation closely aligned with source code.
Tech Glossary:
Include a glossary for technical terms.
Wireframe with Prototype
Collecting feedback for iteration
To validate the prototype, I engaged with various stakeholders, proposing solutions tailored to different personas. The feedback gathered was invaluable and thoroughly analyzed.
Insights
Navigation Complexity
The sidebar required too many clicks to access desired information.
Location Clarity
Users needed better clarity on their current location within the portal.
Content Structure
To prevent confusion, a maximum of 3 levels of hierarchy was recommended.
Content Categorization
More "useful" content with improved categorization was necessary.
Mobile Experience
The mobile interface required a complete overhaul.
Based on this feedback, I created a list of improvements and new functionalities to plan the next iteration, ensuring a more refined and user-friendly experience.

Validating the design

To ensure the new designs addressed the needs of our primary users, I conducted usability tests.
During these sessions, I observed how participants interacted with the prototype and assessed how effectively it answered their questions about platform usage. The sessions provided clear direction on what actions to take next and how to achieve better results on the DataRobot app.

Key findings:

Users found it easier to understand how features work, locate information, and access helpful suggestions.

92.5 SUS score

Strongly Agree - 100 Yes

“I thought the portal is easy to use”
“I think that I would like to use this portal”

Strongly Disagree - 100 No

“I think that I would need the support of a technical person to be able to use this website”

Overall, the Hi-fi prototype consistently scored above 90 on usability, significantly above the benchmark average of 68, indicating an A grade of user satisfaction.

Wawe

User Interface

Based on insights from research, customer support requests, and gathered data, I implemented the following key features to improve the DataRobot platform's documentation portal.

Easy to find in the app

The navigation menu includes a documentation link that assists in finding materials when needed.

DataRobot | Opportunities | 2
Specific case in the right place

All new features will have documentation support in addition to deep knowledge features.

DataRobot | Opportunities | 1
Simple entry point

The homepage is designed to be both functional and visually appealing, offering four main pathways to guide users. A powerful search function also allows for even quicker access to relevant content.

Doc Portal _ Entry
Advanced search functionality

The cutting-edge search engine offers advanced filtering by sections and categories. A result preview feature highlights relevant text, enabling users to find the most pertinent information without opening multiple articles.

Doc Portal _ Search
Rational navigation

Each article includes a table of contents, allowing users to easily navigate between sections, a feature highly appreciated during usability tests.

Doc Portal _ Article
Interactive code samples

Code snippets are linked and grouped by tabs, making it easy for users to copy and paste the necessary code.

Doc Portal _ Code Sample

Design Handoff

I created high-fidelity mockups in Figma, allowing developers to easily navigate through pages, launch prototypes, and access all exportable resources, including objects, visuals, and styles. Figma’s Code panel supports CSS, Swift, and XML, providing a straightforward way to extract code information from specific page elements.

To ensure the design vision was fully realized, I worked closely with the Front-End team, specifying any missing interactions that weren't covered in the mockups. I also reviewed each development ticket to verify alignment with the designs before going live.

DataRobot | Image | Explanation

Business Impact

1 source

that our CFDS share with our customers

78%

customer satisfaction shows that users found what they need

81%

task completion rate for non-technical background persona

64%

is the percentage of reducing the support team workload

480 DAU

We are seeing a positive lift in using doc portal

Results and takeaways

Since the launch of the Documentation Portal, we've seen a significant reduction in user questions about the application, indicating improved user understanding.

Additionally, users have shared that the portal’s use cases have saved them considerable time, highlighting the effectiveness of the resource.

Key Lessons

Strategic MVP launch

This helps deal with out-of-scope requests that could potentially derail the project and help deliver a quality product in time.

Continuous user testing

Design is a constant iteration of improving the experience for the end user. Always find ways to collect and listen to your user's feedback.

Focus on essentials

Iteration is the basic timeframe that is needed to develop a digital product. Iteration always includes new features. Thus, it helps you to identify the most efficient tactics in a short period of time.

Benefits
The Docs Portal has become a central hub for all knowledge assets, significantly enhancing the overall DataRobot product package.
The portal also serves as a tool to help grow and promote the DataRobot community.

Explore even more