Skip to content

Building an MVP

A Minimum Viable Product (MVP) is to me, building the easiest implementation of your idea.
It is the easiest and best way to validate that the idea is not crazy and will help prove 3 things:

1. Someone wants this (at least a handful)
2. Core concept works (to an extent)
3. You want to build this (how much time am I willing to spend)

Someone wants this

With an MVP, there is little to no marketing and coverage of your solution. If you get users, even a handful, this means that they were actively searching for this solution and was not shepherd in by advertisements. These users are also generally quite interested in helping your solution work so do focus on them.

The demographic of these users are also quite interesting as they represent the users who can reap the most value out of your product. If you think about it, they are actively using a potentially buggy product with little to no features and confidence that it will work.
This could come from curiosity or pain but both gives value.

Core concept works

The product at this stage will only include a few features but embodies your core concept. Can you even code this out? Does this concept work and how do you prove it? Having some key metrics and feedback from users is crucial. An example is for AI companies, theoretically almost all things can be predicted with models as long as you have enough data points. For instance, if I had a large dataset of peoples height, weight, muscle density and distribution, it would be not be a hard strength to predict their 100 metre sprint times to a good accuracy. Translating that into a real environment is much more complex. How do you collect this data? How do you action on the predictions?

Working on the MVP will help immensely in digging deeper into these issues. Everything works in the mind but may fall apart as you start building it.

You want to build this

Its essentially to show that you actually want to build this. Brainstorming of an idea is easy but doing the work is hard. Building the first MVP will give you an idea of what your idea actually means and what it will take to build it. There will be parts where you hate ( I despise frontend ) and many things that you will need to learn to overcome. Are you confident and passionate about your idea to go over these obstacles?

Passion is key here. Do you wake up only wanting to build the MVP?

My experience building The Silly Home’s MVP

The concept of The Silly Home is to make Home automation seamless. To do this, we want to change the paradigm of a manual static rules-based approach (if this then that) to an dynamic predictive approach where an AI controls all your devices (trained on past data).

Adhering to YC’s advice on MVP and to build things that don’t scale, I have started to build the MVP on the Homeassistant platform. The MVP is an addon that learns on past device data, make predictions and action on predictions.

So far, there have been stirring interest on the product and I have been really enjoying building it out and making new features. I can’t conclude yet that the concept works yet as there are still holes such as model accuracy, model transparency to fill but it is looking quite hopeful.

Look out for more updates!

Leave a Reply

Your email address will not be published.