COLAB26 - Web App

MyPDP

Get your DREAM JOB with AI-Powered Career Roadmaps: a web-app where you can easily create a customized Professional Development Plan (PDP)

Product Experience

Introduction

Have you encountered the latest surge of online courses? Just about everyone is now either offering a course, or writing blog posts on which courses to take – That’s why we created MyPDP to help aspiring individuals sort through the myriad of options and curate a customized professional development plan (PDP).

By leveraging Open AI’s Assistants API, MyPDP equips users with an AI-powered career roadmap that’s unique to their goals and learning styles, ultimately saving them countless hours of research and giving them back time to focus on getting what they want: their dream job.

As the supply of online courses skyrocketed, so did the number of institutions/websites that offered these courses, which made accurately quantifying the growth of the supply impossible; however, the impact is reflected in the demand.

  • {Coursera, a leading online learning platform, saw} enrolment numbers more than doubled in 2020 and increased by 32% the following year, peaking at 189 million {globally in 2021}” - World Economic Forum citing Coursera 2021 Impact Report
  • Nearly 8.5 million U.S. students take online classes at public higher education institutions {and a total of 11.2M from all institution-types, during 2021}” - Forbes citing the National Center for Education Statistics (NCES)
  • Online Education Market will reach US$ 350 Billion by 2025, globally" - Forbes citing Research and Markets report
  • The US E-Learning Market size is forecast to increase by USD 56.44 billion, at a CAGR (Compound annual growth rate) of 16.48% between 2023 and 2028" - Technavio (report)

Based on these industry statistics, paired with our personal experience and intuition, we hypothesized that there is an unmet market need with pain points that we could solve for and help make better, by applying advancement in AI to this problem area that has seen minimal technological enhancement.

This led us to the following problem statement:

How might we make curating a professional development plan better* for aspiring/ambitious career professionals so that they can have clarity on how to grow their career?
*better being: smarter (tech/AI-assisted), easier (time saving), fair and clear

Research & Validation

Primary Research, Findings, and Insights

To validate our hypothesis, we conducted an initial market research survey using Google Forms and received 39 responses from a variety of age groups and professional backgrounds. 

Based on the responses, we found that:

  • An overwhelming majority (95%) of participants had taken some form of professional development course in the past
  • Almost half of participants had sought/taken a course every couple months
  • 85% of participants stated that they have taken online learning courses before, such as Linkedin Learning, Udemy, and Coursera
  • 59% of respondents would like a way to track their development courses, of which 64% of these were not already tracking their courses:

Common pain points from survey participants were:

  • It’s hard to know where to begin the search and feel confident about the results that come up and knowing whether it would actually be worth the time and investment"
  • I think there are too many courses nowadays and everyone is now a mentor/specialist
  • Sometimes there are many courses out there and it's hard to know what I should be focusing on to break into a certain industry

A prevalent trend became clear to us: Many respondents expressed concern over the vast number of online courses available to them and not having a reliable way to filter out which one to pick.

After conducting our survey and validating our hypothesis, we invited 6 survey participants back to talk more in-depth about their pain points, frustrations, and potential solutions. We then translated their responses into an affinity map, and gathered the following insights on how they chose a course.

Important factors when choosing a course:

  • Price
  • Difficulty
  • Credibility
  • Learning Type

Affinity Map

Ideal Customer Profile (Persona)

Although we envisioned that MyPDP can serve anyone interested in curating their own development plan, we decided to start with helping Driven Daisy - these are her main goals and concerns:

Driven Daisy - Persona Profile

Market Segmentation

To better understand the market size and financial potential for our product solution, we employed a TAM/SAM/SOM segmentation model. Our latest assessment* of the opportunity highlighted:

  • 30K North Americans (US & Canada) is the number of users that we could expect to help with our product
  • $441K (USD) is the estimated revenue via affiliate partnership, assuming all features related to monetization are functional

*This is not a fixed assessment of the market - our intent is to continuously adjust this assessment as we build, pivot, and learn more about our users and competitive landscape.

Market Segmentation Assessment

Competitive Analysis

To round out our view of the problem space, we conducted competitive analysis. We found that there is no dominant player leveraging AI in the same approach as we are, signaling an opportunity for first-mover advantage.

We classified “direct or similar competitors” as either Innovative or Traditional. Innovative direct/similar competitors are also exploring AI to innovate the space but are either targeting a different audience or offering a different design experience; examples are: Explainlikeimfive.io, CareersPro, Audo. Traditional direct/similar competitors are those offering static roadmaps or development guides; examples are: Roadmap.sh, Career Path feature from Glassdoor.

Alternatively, the most dominant players in this space are “indirect or tangential competitors”. These are primarily the online learning platforms (such as Coursera, Udemy, LinkedIn Learning, and similars) that are top of mind solutions when thinking about professional development.

Lastly, our survey data also strongly signaled the lack of dominant player(s) and unmet market need. Their current solutions for career planning were: self-made plans (usually on Excel or Word documents), templates/documents provided by their employers, bookmarking blog posts or guides. Their most tech-assisted solution was leveraging filtering features in online learning platforms.

Competitive Analysis - Select Top Competitors

Solution & Execution

Solution Explanation

So, how does MyPDP solve the needs and pain points of our users?

MyPDP takes user input via a simple 7-question questionnaire (capturing their unique career goals, budget, preferred learning style, etc.) to determine the best courses for them to take. It will then visualize a list of professional development courses for the user via a roadmap-style layout.

This solution addresses the user’s pain points of being exposed to too many online course options and not knowing which ones to choose - whilst providing them clarity on their next step with a visual roadmap that takes into account their career goal and budget.

Feature Prioritization & Defining our MVP

One of our biggest constraints was the timeframe for our software development lifecycle. We had 8-weeks to not only go from planning, to designing, to ultimately building, testing and launching the MVP, but also had to go through team formation, and find a viable product-market-fit.

To help achieve our goal to launch an MVP, we had to align on a core vision (our MVP). By utilizing insights from our survey, interviews, and continuous discoveries, paired with our individual expertise and experiences, we created a Feature Prioritization Matrix.

This matrix takes into account the impact, risk, and effort of each feature we had ideated, and highlights which were: (1) Must Have, (2) Nice to Have, (3) Could be Done, and (4) Not Worth Doing.

Four features out of 21 landed on Must Have, of which 2 were prioritized as MVP features:

  • Capture User Input - prioritized as MVP feature
  • Visualize a Roadmap - prioritized as MVP feature
  • Broad Course Overview
  • Clickable links to the actual courses

Feature Priority Matrix

Pivot 1: Roadmap Loading Time

One of our major pivots was due to the loading time for lengthy roadmaps (i.e. 12 months roadmaps). This was a complex challenge that brought various risks and implications. In short, our solution faced two technical challenges that resulted in high loading time: (1) the greater the length of the roadmap, the longer it took to load, and (2) our cost-free server approach limited our speed to generate the roadmap.

Through arduous evaluation and problem solving as a team, we emerged with a pivot that we determined would minimize impact to user value, whilst also reducing risk of not meeting a key product success metric (i.e. Roadmap creation rate) and only required incremental design and tech effort that we were comfortable with - see ‘Pivot 1 Design Change’ under the Designs & Mockups for visuals on pivot 1.

Pivot 2: Preparing to launch MVP (Descoping & quick pivots)

As we approached the launch of the MVP, we were faced with a series of descoping and quick pivots. Below are a few examples:

  • Descoped: Capturing Desired Skill as user input (removed from MVP)
  • Descoped: Search bar for Job Title (removed from MVP)
  • Quick pivot: Design adjusted to include more pre-defined Job Title options
  • Quick pivot: Progress bar added per user feedback
  • Quick pivot: Copy tweaks from Functional QA testing

Future Pivots (from user feedback and continuous discovery)

Our goal is to perform continuous discovery to identify future pivots and features we should consider implementing. In fact, our affinity map from our in-depth interviews has already identified roadmap features that our users would want:

  • Ability to edit the roadmap details (i.e. criterias the user had input)
  • Gamified experience
  • Detailed breakdown of courses
  • Course credibility: Reviews / Ratings

Product Design

In terms of our product design, we strategically identified that our strategy to leverage AI could require a great portion of discovery and learning effort from our developers. Paired with having to build a web-app from scratch in just a few weeks, we focused on a design that emphasized functionality over applying complex design principles that would put at-risk launching an MVP.

With that said, we still ensured that fundamental experience design principles and best practices on building a web-app were in place.

Designs & Mockups

Figma was our main tool to create and share designs and mockups within our team. We even were able to create prototypes to help us understand the user flow.

Lo-Fi Mockup

Hi-Fi Mockup

Pivot 1 Design Change

Before Pivot

After Pivot

Tech Implementation

Hosting:

  • Backend: Render
  • Frontend: Netlify

Tech Stack:

  • Backend: Python, FastAPI
  • Frontend: TypeScript, React
  • API: OpenAI - Assistants API

Technical Challenges:

The hardest part so far has been adjusting the assistant instructions and regex to enable us to display the data in the desired format while keeping the response time as low as possible.The more instructions we give the chatbot, the more time it takes to “think”.

Additionally, the length of the roadmap affects the response (loading) time. We found that hardcoding the timeframe to three months greatly increased the speed of the response - and built our pivot around this constant. With that in mind, the beauty of OpenAI’s Assistants API is that we are able to hold running conversations with the chatbot, so we hope to build the functionality of adding more courses to the roadmap in the near future.

Scaling and Future Growth:

Our current setup with free tier services on Render and Netlify works well for testing and demos. However, for future growth, we anticipate needing to upgrade to paid plans or migrate to a cloud-based solution. This will allow us to control load balancing and auto-scaling ensuring our app can handle increased traffic and user activity without compromising performance.

Tech Adaptability:

No matter how much you plan, things can still pop up along the way that require us to adjust our process. For example, when we found that a 12 month roadmap could take up to 2 minutes to generate, we pivoted to generate a 3 month roadmap instead and plan to add additional months in the future.  It's important to maintain a “bend but don't break” attitude during times like these so that we are flexible enough to handle anything that comes our way.

Future Steps

Now, what is next for MyPDP and the team?

We will continue building our product! We are passionate about helping our users with the problem they are facing.

Don’t hesitate to reach out to us with any query or feedback on the product or on anything else (our LinkedIn contacts are at the top of the page).

In terms of MyPDP’s product roadmap, we have planned for a V2 build and V3 build after officially launching the MVP in early March 2024. The V2 build will focus on rounding up the product with key features that have become industry standards for a web-app. And, V3 build is intended to focus on unlocking monetization features and opportunities.

MyPDP Future Product Roadmap

Learnings

Product Manager Learnings:

Juan Carlos Yao

Tactical Planning: Coming into the program, I knew that planning ahead was gonna be a key requirement to be successful. However, what turned out to be even more critical was my ability to ‘tactically plan’ what each individual weeks’ goals would be, as each week greatly required different objectives due to the hyperfast pace that we were going through the Software Development Lifecycle.

Product Vision & Prioritization: Ensuring we had a clear product vision was a critical perspective to have when prioritizing which features and pivots we should focus on to drive the most value to our users. Adaptive & Empathetic

Communication: Understanding not only how my team members communicate, but also seeking to understand their perspectives (or concerns) was critical for us to quickly align on last minute pivots.

Designer Learnings:

Jonathan Chiang

Open Communication: Provide clear design guidance to developers to reduce design iterations

Understand Technical Constraints: Collaborate with developers to understand technical constraints to implement the best design that is also technically feasible for our timeframe

Cross-Functional Collaboration: Being the sole designer doesn’t mean that you are alone in designing the product. Other team members can provide valuable insights into the designs and user flows.

Developer Learnings:

John Clapper

Cross-functional Collaboration: This program taught me a lot about working on a cross-functional team; how it draws on expertise from across different disciplines and points us all toward a common goal

Agile Methodology: One of my favorite learnings was gaining a deeper understanding of the Software Development Lifecycle as we practiced Agile methodologies, such as: Stand-ups, Sprint Planning, Grooming, Sprint Reviews, and Retros

Having fun while making group decisions: “Bend but don’t break”, finding the joy and fun throughout the journey is what makes it worth it

Developers Learnings:

Juan Hernandez

&

Cross-functional Collaboration: I learned cross-functional collaboration in relation to the software development lifecycle. It provided perspective on the value that each discipline brings in bringing a product to market.

Software Development Lifecycle: Learning more in depth about the role that each discipline plays. It’s not just brainstorming and coming up with an idea for an app or product. There’s user surveys, problem and solution statements, use cases, etc.

Full Team Learning

Overall, we learned critical and practical skills through our Software Development Lifecycle journey. Building an MVP in just 8-weeks let us see first-hand the value and expertise that each role and individual brought to the table. We not only learned about and from each other, but we have also become a family.