Magic Studio : State of the Machine : Prologue
The first chapter in our story.
Hello Folks!
I hope that each of you is reading this in the safety of your homes, that you are hale and hearty and continue to stay that way.
In the middle of all the chaos in the world, we started building our little company.
Here's the 4 of us founders: 
On the 10th of March, we kicked things off at an AirBnB in Banashankari, Bangalore. We were also joined by Akvil, our first team-member, who brings his design chops to the venture. If want to know more about us, I have put a few links at the very end.
What have we been up to?
The first few days are cryptically demonstrated by this chart: 
It started with picking apart the jobs we knew how to do, and trying to find the biggest broken things we could see. Engineers are like that sometimes. 🤹🏾
Sometime in the evening on the 14th, we decided that we wanted to build a "prototype"! And since we were applying for the YCombinator summer cohort: we decided to do it in ~24 hours.
What's going on here?
We are taking a (somewhat) human question and using an AI model to convert it into a machine-readable query (SQL), and then delivering the answer. For our prototype, understanding the human question is a bit like an adult trying to understand baby-speak. You can't really make out much. But you know the baby is hungry. Point is it doesn't work most of the time. 🤪
But the prototype is still chugging along at abracadabra (deprecated). Give it a spin.
In order to actually solve the problem, we had to severely constrain the information that humans could be asking questions about. For the prototype we created synthetic data, representing a fictional cosmetics store.
But is there a real world case for something like this? We were looking for a potential store of data, that:
- Was impenetrable to the people who needed it
- Was organized in a somewhat specific and standard way
- Had lots of such instances
Turns out there is this wave of such folks who are coming online; who can't be bothered about code or data, and are keenly occupied with putting out and promoting their wares.
Who?🕵
Shopify Merchants did $61B in GMV in 2019, 50% growth over 2018. For the same period Amazon's online retail business was $141B, growing at 15% Y-o-Y. And Shopify is just one of many platforms that enable "small" retail businesses to get online, leaping over the wall of technology that keeps them out. Overall it's more or less even-stevens for platforms (like Shopify) vs marketplaces (like Amazon). Amazon and all the other big e-commerce players corner about 50% of the online retail spend, in North America.
Crucially Shopify (+WooCommerce, Magento) is rapidly powering more merchants (+40% Y-o-Y) and more business per merchant (~+20% Y-o-Y). The number of merchants doing >$1MM/year in GMV on the Shopify platform grew 44% in 2019.
Cutting to the chase, ever more small retail businesses are coming online. They are using standardized setups that work like blackboxes, to help them display, market and sell their wares. Platforms like Shopify, but, blackboxes still.
We think that these merchants feels a pain that we have felt: understanding how customers find products (marketing) and how customers buy those products (sales). Most large businesses or tech-startup types, have armies of analysts to help them deal with this. For the merchants however, marketing channels (eg. Facebook) and transactional platforms (eg. Shopify) present an impenetrable veil over their businesses. Some merchants we have spoken to, tell us that they spend ~80% of their time dealing with how & what to promote.
Oh and just to check the boxes:
- Their data is "standardized", same database schema across Shopify stores
- There are lots of such instances, 600K+ store-fronts on Shopify
What are we building at aarzoo then? 🎅
A product:
- for ~100K (our estimate) Shopify Merchants (Filter: ~500 SKUs, doing ~$20K+ GMV/month)
- that helps them get insights into their products, promotions and customers
- by building human-being compatible access to data from Shopify and the promotional channels (we might start with Facebook)
Help Shopify Merchants see: 
Roger roger?
Our current challenge is figuring out the entry point. Merchants (mostly) don't know what questions to ask. We will probably struggle to answer them anyway. We are working on a basic product that does enough things right out-of-the-box. It should offer merchants a way to interact further to discover opportunities and threats.
With every edition of this letter, I am going to ask you, the reader, to help in some way. Get used to it. 🧞
Here's what you can do for us now:
- If you know someone that runs a retail store on Shopify, connect us with them
- If you run a Shopify store, we want to help you run it
Thanks for reading this far.
I would love to hear what you think of our story. Just reply! Or forward to whoever you'd like to.
By the next edition, we would have built out some of our ideas and put it out in the world. Tall claim, we'll see.
Maybe we would have walked a different walk. Maybe fallen flat. Whatever it is, I promise it will be interesting and exciting! 🤓
Meanwhile stay tuned and (please) stay safe.
Cheers!
Vinod
vinod@magicstudio.com
Founder & Cheerleader in Chief
Set up a chat: Calendly
Follow me: Twitter LinkedIn
About the team:
LinkedIn profiles of the founders: Vinod, Vivek, Akhilesh and Bharat
LinkedIn profile of our first team-member: Akvil
References: