Enhancing Business Insights: A Focus on Dashboard Analytics
Botco.ai I 2022 -2023

OVERVIEW
The dashboard is the heartbeat of a chatbot's interactions, showcasing high-level metrics that provide an overview of its overall performance. As Botco acquired more clients, it became essential to streamline the bot and track its performance effectively.
The goal was to offer a quick and clear overview of essential metrics and data, helping the CX team monitor performance and make informed decisions effortlessly.
Learn more about Botco.ai
PROBLEM
-
Managing multiple tasks can be challenging, especially with numerous clients. CX needed a way to understand, track, and share the bot's performance directly from the dashboard without additional calculations.
​
-
Limited analysis and reporting hindered gaining valuable insights for strategic planning and growth. Additionally, the graphical cards lacked a clear information hierarchy and failed to highlight key data points.
​
-
The dashboard lacked a clear visual hierarchy
SOLUTION
After multiple iterations and observing live data tracking, I independently designed a dashboard that allows users to understand metrics intuitively, without having to spend too much time analyzing the data, all while ensuring a visually appealing design.
​
+70%
User Satisfaction
+20%
Frequency of Use
RESEARCH
The Customer Success team has a weekly reporting call with customers where they currently share a table of the bot’s key performance metrics. The goal is to have an updated dashboard with key metrics so that CX can share snippets of the same with customers, instead of preparing a table.
.jpg)
How CX initially updated the bot’s performance to customers
😣 Common problems CX has with the current dashboard:
Complexity and Overwhelming Information: Feeling overwhelmed by the amount of data and information presented on the dashboard, leading to confusion and difficulty in finding relevant insights.
Poor Data Visualization: The dashboard fails to present data clearly and isn’t visually appealing, CX struggles to interpret the information effectively to customers, impacting decision-making processes.
​
Lack of Customization: The dashboard is too rigid or lacks customization options, preventing a tailored display to specific needs and preferences for each client.
Performance Issues: Slow loading times, lags, or technical glitches within the dashboard can frustrate users and hinder their ability to access real-time data efficiently.
​
Inadequate User Guidance: Challenges in navigating the dashboard, utilizing its features optimally, or interpreting the displayed metrics without sufficient guidance from engineering.
​
Limited Integration Capabilities: Dashboard lacks seamless integration with other tools or platforms that users rely on, disrupting workflow efficiency and hindering cross-platform data analysis.
Question I asked myself before the redesign
-
What needs their attention?
-
What do they need to report on?
-
What deserves a spot on the dashboard?
-
What are the most global metrics that deserve higher visibility?
-
What needs to be visualized?
DEFINE
I created initial designs using existing metrics and consulted the CX team to understand their issues with the current dashboard. I didn't anticipate that solving these problems would require system-wide changes (for this case study, I'll focus on the dashboard).
We developed various cards to integrate into the authoring process, which would then be reflected on the dashboard. Previously, the CX team manually reviewed transcripts and sourced leads from user-submitted forms, but our goal was to display everything directly on the dashboard.
We introduced a Data Submit card to capture relevant data from specific conversation points and a Goal card for setting financial or numerical objectives in chatbot flows.
To evaluate these cards' effectiveness, we tracked their performance on the dashboard.
Design: Learn > Iterate > Learn > Iterate
​
The dashboard is extensive, but for this case study, I'm focusing only on the top section (key metrics). The screens went through multiple iterations, and due to the nature of the project, I couldn't conduct user testing. Instead, I tested the features using data to evaluate functionality. This process was slow but provided ample time for me to understand and refine the design.
​
Version 1
While this new dashboard marked a significant shift, I realized that in dashboard design, visuals often prioritize form over function. but in reality graphs should fundamentally be easy to understand.
As a rookie, I didn't realize that the graphs on each card were not only hard to build but also impractical.
It turned out complicated to have toggle options in the graphs. If we were only allowing one metric view at a time, why have the toggle?
Version 2
I designed this version to show key metrics on top, with a toggle to display additional metrics and graphs. However, this approach didn't work well due to the large scale differences—some metrics were in single digits while others were in thousands, making the graphs hard to interpret.
Version 3
Though this dashboard was the longest-lasting and most effective. We added more key metrics and discovered that users frequently compared current metrics with older ones. It was clear that we needed a comparative date filter.
Version 4
During testing, I noticed that users frequently compared current metrics with older ones. It was clear that a comparative date filter wasn't enough, we had to show them on each key metric. This new dashboard would always show the bot's performance relative to the past, which is crucial for our ongoing efforts to improve its functionality.
THE BEFORE VS AFTER
What didn't work
The lack of real-time updates hinders decision-making and operational efficiency.
Limited analysis and reporting impede gaining valuable insights for strategic planning and growth.
A lack of clear visual hierarchy and the use of unnecessary graphs make it difficult for people to understand the data.
Graphical cards lack a clear information hierarchy and fail to highlight key data points.
What changed
Data-driven insights to analyze KPIs.
The addition of numerous metrics reduces the manual workload of the CX team.
Global Filters.
Display additional data with fewer clicks and avoid using tabs within cards.
Visual hierarchy through the page.
Date comparison - for tracking trends and understanding performance changes over different time periods.
Clear visuals highlight new features or unused features, drawing user attention and encouraging action.
Ability to see transcripts from the conversation details section.
KEY ELEMENTS
Filters - Customizable filters let you focus on specific chatbots, time frames, or sources, allowing deeper insights into key metrics for your business.
Date Comparison - Being able to modify date ranges allows you to concentrate on data within specific timeframes, which is particularly valuable for analyzing campaigns or tracking post-update performance.
What I learnt
What about customers who never want to configure this feature? We should have an option like "Don't show me this again," which would dynamically reconfigure their dashboard layout. However, this would be a significant undertaking, and Botco decided to leave it as is, since the product is used in-house.
Visually indicate that users can connect to a live agent, and note that this feature hasn't been configured in their dashboard yet.
Takeaways
User Focus: Ensuring the dashboard is intuitive and meets the needs of the end-users, you can have a all the information but its not easily accessible, there is no point
Visual Clarity: Organizing information clearly to highlight the most important data. If the key data isn't easy to find, users won't benefit from it
User Feedback: Using feedback to continually improve the dashboard. Without user input, the dashboard may not address their real needs effectively.
Scalability: Designing the dashboard to handle future growth and new features. A dashboard that can't scale will become obsolete as needs evolve.
Simplicity: Keeping the design clean and uncluttered to enhance the user experience. A cluttered dashboard overwhelms users and diminishes its effectiveness.