It all began with a whimsical idea: a WhatsApp-based therapy chatbot powered by OpenAI. The concept? You are your own best friend, and this AI companion would embody your ideal self, remembering your thoughts and helping you along the way. However, reality quickly set in:
- WhatsApp chatbots require dedicated numbers, and I couldn't migrate my personal number.
- While getting a new SIM was easy, maintaining an active phone and plan was not.
So, I pivoted. An email-based AI companion. Same concept, different medium.
The Concept: Your Future Self as Your Guide
What is Selftalk.ing? An AI companion that personifies your future, ideal self. It's designed to help you reach your goals, offer support when you're stuck, and keep you accountable—primarily through email.
Problem
Many of us want to discuss our life problems but feel shy or afraid. Or we struggle to stay on track because we're doing it alone.
Solution
Talk to an AI that learns about you, inquires about your life, reminds you of your goals, and acts more like your own 'future' self.
Reality: No Problem, No Solution
As I delved deeper, I realized there was no 'legitimate' problem being solved. I had fallen into the classic 'feature-making' trap, focusing more on building than on validating the idea with revenue or paid users.
This realization led to another pivot: something between an emailing service and a personal accountability tool.
A Glimmer of Hope: The First Paid Customer
On April 19, 2024, I hit a milestone: my first paid customer! $10 MRR (Monthly Recurring Revenue). But the celebration was short-lived. The user churned the following month due to inactivity.
The Lessons: A Retrospective (May 24-25, 2024)
- Distribution is key: Finding channels to acquire new users is crucial.
- Feedback bridges: Reduce friction in collecting user feedback.
- True validation: Find users who explicitly state your solution is valuable and solves their problem.
- Marketing hack: Present your product as if it's doing the impossible. I marketed Selftalk.ing as a way to talk to your future self. It piqued interest that led to signups. (I could always justify the details later)
- Network effects: If building a freemium product, try to incorporate network effects. Users should be inclined to promote it.
- Focus sprints: Two weeks of coding, two weeks of marketing. No multitasking—focus on one at a time.
- Pricing strategy: Start with one-time payments (e.g., Lifetime Deals for early adopters), then increase prices or introduce monthly plans.
- Usage clarity: Be clear if your product requires daily usage. If yes, pull all stops to ensure users return. If not, charge upfront for a "set-it-and-forget-it" model.
One More Pivot
Now, In a final attempt to salvage Selftalk.ing, I:
- Rebranded it as a challenge-based "Personal Accountability Companion" with a revamped landing page and dashboard (the fifth redesign!).
- Solidified the "why" and the problem being solved:
- Unmotivated to finish your life goal
- Inconsistent towards their goal
- Feeling lonely in finishing your life goal
- Stripped away non-value-adding features.
- Eliminated the free trial after an experiment showed that even when offered a free week, users lost interest after a day or two.
Leveraging Psychology
I also tapped into my knowledge of psychology:
- Public feed: Share your progress publicly for accountability.
- Time-bounding: No more free trials. Instead, commit to time-bound challenges:
- Starter (7 days): To try it out
- Habit (21 days): To explore building a habit
- Lifestyle (90 days): For the truly confident
- Risk elimination: Money-back guarantee if you complete your goal. This uses fear as a motivational trigger.
The Conclusion: Time to Move On
After three months of continuous effort, it's time to stop hand-holding Selftalk.ing. Despite showing initial traction, it faded due to the lack of a proper Ideal Customer Profile (ICP) for whom it solved a burning problem. Even with dedicated marketing efforts, active results were elusive.
My goal remains clear: achieve financial independence through building products. While I could persist in hopes of a breakthrough, my six-month timeline and the current stagnation suggest it's time to reprioritize.
The Wins
- Hit my first $10 MRR
- Reached 100 users in record time
- Successfully time-blocked for building and marketing
The Losses
- Not having a proper onboarding in the first place, led to users not fully utilising the product.
- Not having enough ways to collect feedback from users, thus ended up with less feedback and more assumptions only.
- Not being clear with ICP and marketing channels
The Silver Lining
The product is in good shape to sustain itself with minimal maintenance. Current active expenses are manageable:
- Render.com for email sending (cost can potentially be reduced)
- ChatGPT usage (unlikely to exceed $5 at the current pace)
Parting Thoughts
This post compiles my efforts over the past 3.5 months building Selftalk.ing. It's been a rollercoaster, but rich in lessons.
I hope you've gleaned a thing or two from my journey. Thank you for reading!