WHAT IS PEER INSIGHTS?
Gartner Peer Insights helps choose IT solutions with confidence by providing software reviews that you can trust. Every review is verified before publishing to ensure completely authentic insights from the peers
Who verifies these reviews? Peer Insights Verification and Moderation team (VNM)
MY ROLE
TEAM
TOOLS
DURATING
ARM
A smart Review Moderation tool designed from scratch with the goal to reduce user's efforts and increase efficiency & productivity.
UX Research
Market Research
UX/UI Design
Usability testing
Product
Design
Engineering
Review Moderation
Figma
Participatory design
Ongoing since May’21
Context
Reviews are rigorously vetted by Gartner to ensure there is no vendor bias, no hidden agendas, just the real voices of enterprise users.
The process includes verifying the reviewer’s identity, assessing whether a conflict of interest exists, and moderating of content. Gartner Peer Insights’ moderation team takes up to 3 business days to process a review and conclude its approval/rejection.
Problem
Review moderation is completely excel-driven and there has been a significant drop in the current review publishing cycle.
Goals & Success Metrics
Simplify process to save time
20% increase in Publishing cycle of reviews
Reduce moderator’s efforts, increase productivity
Task 1: Understand what & where are the process challenges
Now was the time to deep dive into details and identify and understand key problems within the current moderation process. I collaborated to perform:
1. Shadow Activities
2. Contextual inquiries
RESEARCH: SHADOW ACTIVITY
Breaking down moderation process
Manual processing & consistency check takes almost 6-7 hours of the day & thus that could be a good area to solve first.
The current moderation process is so complex, solving for entire journey seemed a long way, thus we considered taking baby steps and adding value for at least one section of the process that is highly impactful for our business
KEY TAKEAWAYS
-
Juggling between 8 tools and 9 files per review
-
Series of manual steps (chances of error)
-
Lot of offline data enrichments
-
Lack of flexibility
-
Redundant/extra steps
-
Extremely time consuming
-
Process developed internally within Moderation team
Overview of current Review moderation process
But wait...considering the above situation, If almost
7 hours goes into manual moderation, when do they attend other calls, have lunch, tea breaks...?
This question pushed us to have a closer look into how the users are actually processing the reviews. Are there any makeshifts?
Task 2: Understand User Pain points
While I was still struggling to understand each detail of this complex process, it was also important to understand the key challenges users face and their behaviors and attitude.
1. More Contextual inquiry
2. Use case Identification
KEY USER INSIGHTS
3 behavioral patterns were identified while talking to each Moderation team member (total 6 in numbers)
01/
Skim through spreadsheet data
This helps in providing cues and patterns to identify any potential error
02/
Batch processing for decision making
They prefer saving time by performing similar tasks together and batch process the final decision
03/
Seek help from team
Due to process complexity, they often seek help when there are doubts in the process
KEY USER PROBLEMS
5 problematic areas were identified
01/
Time inconsistency
Often it takes more time than estimation since some online details are difficult to find, for multiple reasons
"
"
04/
No collaboration
I couldn't find any previous notes so as per my initial checks, I rejected the review...
"
02/
No real time data updates
Why do I have to recheck the same details if somebody else has already done it?
"
"
05/
03/
No centralised data sources
Just to check one or two previous data, I have to open 4 files and take multiple actions to look for desired information
"
"
Constant thought of missing steps
I don’t want to commit mistakes thus I double check the details but this consumes more time
"
"
"
Task 3: How could I translate and scope out these pain points into design opportunities for our MVP?
Before moving to opportunity space, it was important for me to explore around Industry best practices for UGC Moderation.
1. Literature Study
2. Market Research
KEY USER PROBLEMS
MARKET RESEARCH & LITERATURE STUDY
I interacted with Linkedin Trust team members as well as Gartner’s other publishing team to understand their approach
KEY LEARNINGS
-
Focus on automation to enhance productivity
-
Design for quick decision making
-
A lightweight system thats easy to maintain
-
Eliminate biased decisions and differences
-
Eliminate extra efforts
-
Empathise with the moderator’s work
Also, as per "Use of AI in online content moderation", by Cambridge Consultants,
1.
Automated systems will be key for future online content moderation
2.
Human input will be required to augment AI systems for the foreseeable future
With that in mind, the next step was to define the Problem statement and scope opportunities
HMW provide easy access to all relevant information while reducing efforts & complexity of steps and provide best moderation quality?
IDEATION
Persona
To put my bias and assumptions aside, I created primary and secondary personas to remind myself who I essentially design for and what are their frustrations.
IDEATION
Design Opportunities
& Ideas
How can we reduce process complexity?
We tried generating ideas as per the pain points identified as well as the learnings from User and Market research
RESEARCH FINDINGS & PAIN POINTS
OPPORTUNITIES & IDEAS
IDEATION
Design Opportunities
& Ideas
How can we provide one point access to relevant information?
Once we identified what information would be relevant to users we started segregating information and clubbing similar items together,
It was time to design
As everyone says, don’t marry your first design idea... WELL...Not even my users did! Yay!
IDEATION
Initial Wireframes
(Low fidelity)
I made multiple versions of design starting with user flows to adding conceptual features so that we can ask for quick feedback
We wanted to get feedback on
-
Overall navigation of the process
-
Content validation
-
Usability of the design
-
Other requirements if needed from their side
KEY USER FEEDBACKS
Positive
-
Overall navigation is pretty clear
-
Happy to see rejection flags upfront
-
Good to see clear filters on top
-
More information and past notes on the same screen helps save clicks and time
Critical
-
Users wanted action items at closer proximity
-
They did not want to mix different review tiers together
-
They did not want modals for doing actions (will have to open and close it every time
-
Would want to have minimum clicks possible
-
Would want to see more red flags
It would have been great if I could just click on the name and see the entire history of the reviewer, so i can catch suspicion...
The “add comment” seems to be too far, every time i will have to drag my mouse from the details to add comments section individually...
Would want to see reviewer name and job title together to quickly copy and paste it on google for Profile search...
Final Design
Based on the key findings we synthesised through conducting affinity diagram, we iterated on the design and upgraded the prototype from mid-fidelity to high-fidelity
INTRODUCING...
Gartner Peer insights ARM (Automated Review moderation tool)
One point access tool designed to bring faster review moderation and increase moderator's productivity significantly
A R M
BREAKDOWN
1. Displaying a large volume of data on a single screen was a major challenge. My goal was to make the most of the available space to present information that can provide:
1.
Limited toggle between screens
2.
Quick online & offline search experience
3.
Instant decision making and updates
4.
Reduce errors, ensure consistency
-
Merged decisions and comments (lesser clicks, quick decision)
-
Flags for visual cue (reduce cognition)
-
Quick links (no more copy-pasting)
-
Batch level processing
-
Grouped similar checks
-
Navigation (flexibility and information segregation)
-
Binary filters with no's (faster slicing of information)
-
Soft nudge (avoid errors and provide next steps)
-
Quick real-time information to refer (lesser clicks)
-
Data capture and updates (consistent & informed decision)
2. Ensuring review moderation quality was another big challenge to solve. The idea was to
1.
Reduce decision bias among users
2.
Provide review context & sentiment
3.
Consumable & scannable content
-
Go through history without loosing context
-
Review Relevance score (easier to identify potential errors)
-
Matching keywords (easier to navigate through relevant content within review)
-
Sentiment analysis (to initiate what the review's context and sentiments are)
-
Identify any confidential data input (easy to find and mask it)
-
Consumable review format (skim through content and make edits wherever necessary)
5% increase in productivity
(quarterly basis)
POTENTIAL IMPACT
The shift from excel to a smart tool for manual moderation is expected to increase
"I wasn’t sure how all these Excel-driven actions could be handled in a single tool, but this looks really exciting and promising! I’m looking forward to trying it out. When does it launch?"
-Saumya sharma, Product Ops Associate
Wait... there's more!
Can this help us work on the review quality as well?
We can leverage some smart rejection reasons to potentially “nudge” reviewers about best practices of writing reviews to increase chances of approval
REFLECTION
Key learnings from the project
Wow! You get paid to watch videos and read posts?
I heard this from many people who connected me with those working in UGC moderation. And no—it’s not a “glamorous” job. Through market research, I got a deeper understanding of the emotional toll it takes. These moderators have to review disturbing content so the rest of us don’t have to see it.
Thinking small but impactful changes
I learned that taking small, steady steps helps me gradually understand the project end-to-end. Complex processes can feel overwhelming, but breaking them into smaller parts and focusing on the most impactful areas makes them manageable.
Ideas can come from anybody
As a designer, I’ve learned that aligning the team not just with the problem, but also with the thinking behind it, builds stronger engagement. I’ve had teammates reach out just to share their ideas, and when I was stuck, those inputs often sparked the next step forward. :)
Thanks for reading! :)