Solopreneur's Product Iteration Toolkit: Practical Steps and Examples from Llanai
Explore a comprehensive toolkit for designing and iterating products from the perspective of a solopreneur, focusing on practical, actionable steps and real-life examples from building Llanai.
Posted by
Ari NakosRelated reading
Introducing the TOEFL iBT Practice Test: Your Free Path to Success!
Take a free practice test for the TOEFL iBT and boost your confidence before the big day!
Llanai Global Career Catalyst Scholarship Fall 2024
Jumpstart your international career with Llanai's Global Career Catalyst Scholarship. Understand a new language, immerse in culture, and open doors to global opportunities.
Impulsa tu Carrera: Beca Catalizadora de Carrera Global Llanai Otoño 2024
Inicia tu carrera internacional con la exclusiva Beca Catalizadora de Carrera Global de Llanai. Domina un nuevo idioma, sumérgete en la cultura y abre las puertas a oportunidades globales.
Introduction to User-Driven Product Iteration
Solo building a product is an amazingly rich experience. If you do not mind working on a Saturday night, 6 Saturday nights straight, you can do this. In this piece, I will dive deeper into the framework I used to build Llanai, a chatbot that helps you learn languages through oral fluency.
There are always questions that swirl in your mind, as a first time solopreneur, such as:
What is "enough" for an MVP?
How to iterate on a product?
How do you gather feedback on the product?
What are the minimal tech stack requirements for an MVP Chatbot?
I have been there, and now that I am iterating on the 3rd version of Llanai, I want to share the answers.
However, I did not answer these questions alone. I scoured the Twitter and YouTube for blueprints. I leveraged ChatGPT as my technical co-founder and teacher. I took copious notes. Let's dive deeper into the recipe for a successful product.
Role Models
While I decide alone, I do not build from scratch. I leverage a curated wealth of information from brilliant entrepreneurs out there, which I wholeheartedly recommend. They help me think clearly, as they lead by example. One thing they all have in common is that they share succinct advice to become a better operator of my brand.
Here are 5 people I recommend:
- Marc Lou: A husband, surfer, and genuinely hardworking indie hacker, who has amassed wisdom and shares it in crisp tidbits. He has the mindset of a designer and the skills of a full-stack web developer. I also bought Llanai's template from him.
- Justin Welsh: A husband and former corporate employee with major success. In 2019 he pivoted and leveraged his expertise and network to build his own brand.
- Sean "Swyx" Wang: A pure genius with a background in finance, then pivoted to front end development. Now he's a VC and hacker in the field of AI engineering. I wholeheartedly recommend the "Latent Space" podcast.
- Jordan O Connor: A former electrical engineer, husband, and father of 5 running his brand as a solopreneur. I took his course Rank To Sell to dive deeper into SEO, which gave me direction on how to position my brand.
- Li: A mother of 2 boys specializing in User Design for Startups. She owns her craft with palpable examples on how to create a positive UX.
Subject Matter Expertise
I need to highlight the importance of subject matter expertise in designing products. They provide you with context on user experiences. An entrepreneur attends to poor user experiences. Be it a lack of on-demand ride-hailing or consumables delivered to one's home, we have Lyft and Amazon as solutions nowadays. Of course, these are examples for the market in the USA, as Lyft and Amazon are not the exclusive providers of such experiences.
In the case of Llanai, the core offering is an educational experience. It aims to make oral fluency in languages easier to achieve.
Surely one can summarize an educational experience as a set of clearly communicated information, which left the learner satisfied. However, it does injustice to the process needed to arrive that endpoint. For example, language learners have various learning goals from knowing basic phrases to giving speeches in technical conferences. To keep the learner engaged, we must serve their needs.
How does one craft this experience that may last days to months, depending on the complexity of the learning goal and the learner's background?
Fortunately, I have experience in education on both ends of the transaction. From 2010-2017, I was an educator in STEM to students aged 8-24. Ever since, I can remember, I have never stopped learning languages or taking courses.
Solopreneur's Product Design and Marketing Toolkit
To design and scale from MVP to a desirable product, I focus on user friction. I use the following steps, in order.
- Design the Vibe
- Craft Real User Personas
- Map the User Journey
- Map the User Flow
- Get Feedback from the User
- Submit product in Directories
Design the Vibe
Strive for a positive user experience. While profit is the ultimate goal, a positive user experience is the means to that end, leading to measurable organic growth over time, trackable by key metrics such as page rank. For instance, searching "language learning" or "speak Spanish" consistently brings up Duolingo in the top 3 on Google worldwide, thanks to its consistent service quality.
To achieve such results, implement a vibe that retains users by exuding trust and user-friendliness.
Building Trust: Key Steps
- Create helpful content: Free and valuable content builds trust.
- Establish social proof: Gain recognition through reviews, media mentions, hackathon rewards, or funding.
- Clear product policies: Terms & Conditions, Privacy Policy, and Cookie Policy are essential for all modern SaaS, also boosting your Google ranking.
Ensuring User-Friendliness
- A soothing user interface: Use a suitable color palette and proper font, ensuring accessibility.
- Easy navigation: Ensure seamless flow across mobile and desktop platforms.
- Fast load times: Crucial for user retention and search engine ranking.
- Easy transactions: Simplify commercial or informational transactions. For Llanai, this means easy subscription purchases, lesson scheduling, or authentication.
- Review past transactions: Allow users to easily find receipts, past threads, or progress reports.
Category-Specific Features for a Digital Language Teacher
As we build a digital replicate of a language teacher, consider the following features for a chatbot. These will be examined in greater depth in the next blog, "Prompt Engineering a Language Teacher"
- Personalization: Remembers the user is a student with specific learning goals.
- Targeted Feedback: Focuses on the student's weaknesses and strengths.
- Adaptability: A great teacher is fluid and can tailor their knowledge to their target audience.
Understand your User
User research informs us of various User Journeys. A user journey describes the entire experience of a prospective or current user, before and after using your product. Consequently, there is no single user journey, as there are multiple experiences to consider, which we call user personas.
Be simple and efficient. Visualize this info using pen and paper for the sake of shipping fast. If you prefer digital tools, I recommend an intuitive and minimalistic whiteboard environment such as tldraw.com. There's even an AI-powered version for creating animated graphics. Here's a short tutorial I made with it.
Finally, I'll dive into user feedback, as that's the fuel for the aforementioned workflows!
Craft Real User Personas: Young Spanish Professional
Let us take a young professional in the service sector, based on research I did on HelloTalk.
Name: Verónica
Native Language: Spanish
Practice Language: English
Proficiency Level: Intermediate
Learning Goals: Improve speaking & listening skills to collaborate to pass TOEFL.
Topics of Interest: Explaining work to customers
Map the User Journey: Discovering Llanai
This is a fictional map, but simulates one of the most plausible ways of interacting with Llanai based on website analytics.
- Awareness: Discovers Llanai through English language learning groups, such as Facebook, to address the need for improving the speaking & listening scores for the TOEFL exam.
- Consideration: Evaluates app features, user testimonials, and cost-effectiveness compared to other language learning tools.
- Initial Use: Tries the app's free version; begins personalized, AI-customized sessions.
- Habit Formation: Uses the Scheduler tool to establish a routine. Engages regularly through AI feedback and reports, building confidence.
- Progress Evaluation: Uses the app’s progress tracking features to assess improvements in speaking and grammar.
- Advocacy and Expansion: Upgrades to Serious-tier and recommends the app to peers, and requests new app features.
- Objection Handling: App addresses concerns about AI effectiveness and budget with clear results, flexible pricing, and constant support by the maker.
Map the User Flow: Landing Page
User flows are properly depicted via flow charts of steps. These techniques feed into the technical requirements of user interface design.
Let's take an example of Llanai's Landing Page. Goal is to chat with Llanai. I have to use signals that indicate trust, because I do not have a brand that speaks for itself! So include social proof along with a clear description of the product's use case. As a result, they help increase the activation rate.
This is a simple user flow for Llanai's landing page. The user is greeted with a clear title, social proof, and a call-to-action (CTA). The user is then directed to immediately start their journey with Llanai on WhatsApp, no sign-up needed. If they don't click on the CTA, they can scroll down to learn more about Llanai.
User Feedback
As a solopreneur, starting small is crucial. Talk to friends and family, especially those who would be qualified users, to gather quick and meaningful observations. Use niche directories, which brings us to the final tool of this toolkit.
Directories
Visibility is fundamental. Social media is essentially an interactive directory. You need to be seen before you can get feedback. Additionally, you need to keep an eye on the competition. A few directories I have first-hand experience with are the following.
- ProductHunt: A platform frequented by beta-testers, early adopters, and big enterprises. Most visitors of Llanai are from here, because of my constant presence. Because of my activity, people have tested my product. Here's my Llanai Launch, which was mildly successful.
- HackerNews: YCombinator's channel for tech news. Be succinct and honest with your product pitch. Here's my post on Llanai.
- Facebook: Great for reaching out organically to potential users. Llanai on Facebook.
- LinkedIn: Generates traffic, with about 30 visits per post and growing! I have become more active in the last 2 months. Llanai on LinkedIn.
Niche-Specific Media Channels
- Reddit: Find a suitable subreddit. Respect the channel's rules and avoid spamming to prevent being blocked. Here's the Spanish sub-reddit, which has helped me further my scope of user research.
- HelloTalk: I leveraged HelloTalk, where I talk about Llanai and other tools I build. Read more about it in my blog on making user market research as a solopreneur.
In Llanai's case, the users lacked an accessible speaking partner that provides valuable feedback. This is my MVP. An application that can reliably chat and provide corrections and suggestions to improve the learner's fluency.
Solopreneur's Technical Prototype Toolkit
To build an MVP SaaS, this is all you need. Given that chatbots are becoming mainstream, including a chatbot interface is essential.
- Website Template
- Chatbot Interface
- Chatbot Logic
- Data Storage
- Chatbot Hosting
- Code Version Management
- Website Hosting
- Logo Maker
Website Template
A storefront is crucial to establish a trustworthy outlet for your product. A paginated User Interface using HTML, CSS, and JavaScript is the baseline. For Llanai, additional features include authentication and subscription management, requiring an OAuth portal for logging in via Google, Facebook, LinkedIn, etc., and a Payment Portal for selling products via Stripe or Paddle.
This is a significant undertaking, so it's important to systematize it for future iterations. I used the Shipfast template by Marc Lou, which was a valuable purchase (179 USD) and taught me a key lesson: productize your workflows!
Website Hosting
Your website needs to be live on a server. For simplicity, use Vercel, which integrates with GitHub. Other solutions like Cloudflare are also available and may be cheaper. Vercel also offers basic website analytics, such as referral domains to your website.
Chatbot Interface
A reliable and trustworthy interface for the chatbot is essential. Bots are ubiquitous on apps such as WhatsApp, Discord, Telegram, and Messenger. Using an existing interface can avoid friction from app downloads, and reduce time to development.
For Llanai, I chose WhatsApp due to its 3B user base. You'll need to purchase a phone number; I recommend Twilio. You also need to set up a Meta Verified business account.
Chatbot Logic
This is the chatbot's backend. There are few no-code platforms for building a WhatsApp chatbot for interactions that require additional logic control, such as time-based interactions. I built this from scratch and deployed it on AWS. Will talk about it further in a later piece.
Data Storage
We need to create and store user profiles and chat history to perform NLP on, making it essential for a digital teacher. I recommend Supabase, which leverages PostgreSQL, a robust database. Supabase also offers a vector database solution, which is a data storage solution designed for LLM.
Chatbot Hosting
You need to host your bot, and I recommend using AWS. More details will be provided in a future blog.
Code Version Management
This foundational step involves signing up for a version management platform like GitHub, GitLab, or Bitbucket. It is necessary for managing the versions of the chatbot and website.
Logo Maker
A logo is key for branding. Keep your desired vibe in mind and choose colors that reflect your brand. For Llanai, I chose a friendly animal that connotes fast learning—a rabbit.
I used Leonardo.AI to create my logo by prompting it with text. Another option is Marc's logofa.st.
Iterations: Using Feedback for Llanai
I engaged early on with users and the two most pressing features to work on were:
- Make Llanai's responses faster
- Create a lesson plan
I addressed Llanai's speed by optimizing prompts and by switching LLM models and providers for key steps. For example, Groq is an LLM provider that leverages their TPU inference engine for serving high quality frontier models, such as LLAMA-3-70B.
Lesson plan creation was another layer of chatbot logic, which proved to be necessary. It certainly differentiates the app from others, which look at each learning session as an exclusive experience. However, a journey to perfecting oral skills is quite dynamic and requires tracking metrics longitudinally, as well.
This is where subject matter expertise comes in handy. I used literature and scientific expertise to construct a useful dashboard for language exchange. More details will be shared in the coming piece.
Conclusion: Iterate, Improve, Succeed
To summarize, when you set up shop, you need to have a solid grasp on the user's needs. Establish a framework that intertwines user experience design with a technical stack. Master your tech stack.
As a solopreneur, you are both the architect and the construction worker.
In the following issue, we will dive into what made Llanai possible, namely prompt engineering.