Pondr | Intelligent Product Analytics Platform
June 2021
This is a culmination of my work over the past year working as the Co-Founder & CMO of Pondr, a product analytics platform which aims to help companies build better products through unlocking the potential of their online customer reviews. I led the design and implementation of our B2B enterprise solution, which is currently deployed live and is being used by companies around the world. I am currently working on our B2C solution which involves an AI personal shopping assistant powered by GPT-3, along with a modular browser extension to help consumers make smarter purchasing decisions. Pondr is a SaaS analytics startup backed by Microsoft, and one of the few startups in the world to leverage GPT-3 for consumer product analytics.
Project Overview
Project Deliverables
-
- User Interface
- Coded front end application
My role
-
-
User Interface Design
- Personas
-
Wireframing + prototyping
-
Visual design
- Front-End Engineering
-
Project Context
-
-
Product Design for Pondr
- Timeframe: 6 Months
- Startup Team: Akshay Murthy, Graham Sabin, Thomas Stahura, Abhi Balijepalli, Zyad Elgohary
- Design Lead: Akshay Murthy
- Tools Used: Adobe XD, React, Tailwind CSS
-
Background
Pondr offers a suite of enterprise market research and product analytics tools to help customer experience/product development departments do their job more effectively, and pinpoint specific areas of their brand and product that require the most improvement to be successful. Pondr uses NLP (Natural Language Processing) and GPT-3 (Open-AI’s new powerful language model) to analyze customer reviews effectively. Our notable feature, alongside our interactive report, is our powerful AI Q & A system, which allows companies to ask any questions about their reviews, and get detailed and actionable insights based on our GPT-3 model which we have trained on a user’s dataset of reviews.
Pondr’s consumer-facing side is a shopping assistant built for the average shopper. Our goal is to provide shoppers with an informative product analytics marketplace where users can view detailed insights about every product to help them become smarter shoppers. From our product insights, users can compare products side by side, save money, and get the most value from their budget.
As the Co-Founder / CMO & Product Design Lead, I had to wear multiple hats to bring our product to fruition, due to our small founding team of four people. In the span of 6-8 months, my job was to design, prototype, and code a functional web application that would cater to the needs of our customers, and be highly intuitive and visual to make our product easily usable.
Alongside the design, I coded the front-end for our website and web application using React and Tailwind CSS, but also worked with Python’s Flask framework.
The Design Process
Before starting the development of our application, we conducted extensive research to understand our various stakeholders and end-users. After interviewing a large number of Amazon sellers about how they utilized their existing customer feedback and some of the problems they encountered with existing feedback platforms, we came up with four main design requirements. These requirements addressed the major pain points our audience faced and would help us build a streamlined solution.
While we had a general idea of who we wanted to build our product for, we needed to understand our ideal customers to pinpoint their specific pains, needs, and job responsibilities. Customer experience managers and market researchers within companies were the ones who were willing to talk us through their daily duties and technologies they used daily to complete their work. From the research we conducted and the responses we received, we developed three ideal customer personas for our product – Alexia, Guart, and Mike.
Ideation + Brainstorming
The scope of the problem we were tackling was pretty big, so we need to focus our thoughts and efforts on tackling a specific part of the customer feedback lifecycle in order to provide the most value to our users. We wanted to focus on leveraging existing customer feedback for market research purposes, to help market researchers and customer experience managers perform their duties more efficiently and effectively. As a team, we came up with four main categories and grouped our ideas within them using an Affinity Diagram. After voting as a team on which features and categories would be the most useful, we came up with our main application features. Here are some of our conclusions:
-
- Creating a web-based application with no external database connectivity required from user
- Leveraging interactive charts, tables, and graphs to convey product insights
- Harnessing the unlimited potential of our AI model through built-in Product Q & A
To develop a truly unique and tailored solution for our audience, it was important to understand the various competitors out there that were centered around the idea of customer feedback. Our initial adoption strategy was to attract Amazon sellers, as we realized the Amazon platform is very surface-level when it comes to customer feedback, which proves to be unhelpful for the companies themselves. We tested 6-8 competitor products, detailing their user experience, how they organized and presented product feedback and the affordability of each solution. Some notable competitors include UserTesting and Qualtrics, but these companies do not offer the same type of customer insights as we do. Additionally, they are very expensive for the average small to mid-size company to utilize. Understanding our competitors allowed us to note down important differentiators to make our solution user-centered and see what features users liked and disliked.
Once we narrowed our ideas to focus on building a web-based experience, we explored our ideas from our brainstorming stage to extract key product features that would directly address the needs of our users.
Our focus was to display the product insights we gather visually through interactive graphs, charts, and tables that can be filtered by specific time periods. Comparing products side by side with overlapped trend graphs was another core feature that would allow for in-depth market research and to clearly pinpoint differences in customer feedback and product behaviors. We also wanted users to ask questions about their product using our Q & A feature, and get responses based on their customer feedback.
The architecture of our application was pretty straightforward, with the top of the hierarchy being our main website, and the main application being accessed through a sign-up portal. Within each tab that represents a product feature, we have many different interactions and tasks that can be performed that all serve a unique purpose.
Sketching + Protoyping
Wireframes
Using our information architecture as a guideline, I created a series of wireframes to act as a framework for our enterprise application, as well as our consumer-facing analytics marketplace. The wireframes were then turned into a low-fidelity prototype to model the various user interactions and overall flow of our application.
Enterprise Platform Wireframes
Product Analytics Marketplace Wireframes
Below is the low-fidelity prototype of the application which outlines the key features and overall layout.
Prototype Evaluation + Metrics
After launching the first version of our product beta, over 80% of our customers reported an increase in customer understanding and a better understanding of their overall product performance online.
Based on our feedback from the companies we talked to, most of the features were well perceived. We asked them to rank each feature based on usefulness, and most companies stated that the “Analytics Report” was the most helpful in pinpointing product insights. Our “AI Q & A” feature was also highly ranked to the potential it has to answer any type of question. However, companies wanted to see the following changes:
- Add product grouping features to the main product dashboard
- Categorize reviews based on date ranges within categories, same with graphs
- Gather more specific features of the product rather than the general category
- Switch pie chart graph to a bar graph for distribution of total sentiment
When defining the visual identity of our application, I wanted to emphasize the creativity, potential, and power our solution has to make companies think about their products differently or, as our name suggests, ponder about them. From interacting with our application and engaging with our features, I wanted users to feel like they are learning something new every step of the way and understand their customers’ behaviors, emotions, and motives on a deeper level.
I continued my design style of emphasizing rounded shapes and smooth typography to make the UI cater to our data-heavy interface. The color palette I created follows a range of deep purples and magentas, which can be combined to form striking gradients to emphasize trends, as seen by our graph components. The app itself follows a lighter color scheme to create further contrast with the visualized data, and help the user focus on details. I incorporated iconography within our application to indicate to users product likes, dislikes, review volume, star rating, etc. I used drop shadows to make components stand out, such as search boxes and card designs with the use of icons, to help extract key insights in a snapshot to differentiate from the main product graphs.
Final Design
Project Reflection
Pondr is a SaaS analytics startup backed by the Microsoft for Startups program and was founded in Seattle, WA in 2020.