The Buy Low, Sell High Brain

Hello Gleamers:

Over the last few weeks I’ve been investigating online arbitrage and the many tools that are available for individuals to seek out and find items to buy and resell on Amazon.com. I had a pretty intense dream where I had developed a program that was making me millions of dollars doing just this thing. So, when I awoke and remembered this idea, I quickly wrote down all the details I could recall about it. I know I called it the Buy Low, Sell High Brain.

At the core of any business, the idea is to expend an amount of money that will give you a greater return at a later date and time. The faster and greater you can get that return the more profitable your windfalls shall be.

The problem with online retail arbitrage are many but the one thing that can never be solved is the involvement of a human being handling all the research legwork. This is a huge bottleneck in the process. I assume I was dreaming about this because I had come across a program like OAX-Ray in my arbitrage research.

So, in my dream I wrote an AI program that scours a set list of websites for items for sale and cross compares them with the buy box and cost thresholds for resale arbitrage via Amazon FBA.

Pretty simple right? However, my dream went further. I literally was spec’ing this out and I’m surprised I remembered these specs. Here’s a general outline of what this program does:
Not everything would be desirable to sell on Amazon FBA. The fees are pretty high so there’s no way to sell low cost items and still make a profit in my mind without some sort of additional deal to get favor from Amazon. So certain threshold gates needed to be established via program configuration:

Configuration would allow threshold gates for the following:

• Limits on profit after costs
• ‎Limits on sale velocity via a formula based on the item’s sales rank and category.
• Amazon ‎Competition
• ‎Item size
• Amazon Restricted sale of item

I’m sure there are more, but these were the ones I could remember from the dream.

This program would act very much like the costly screen scraper browser plugins but take an additional step in providing additional AI logic to make decisions based on the pricing of items across all retail websites.

Purchase accounts would be setup on all websites allowing to accommodate for e-bates, Rise.com, and/or other retail store discount codes and rewards systems. Full knowledge of these systems and account balances would be known by the AI and would allow for it to consider these when making best purchase decisions.

The first phase of this program would prepare the items daily in a web browser, with the first tab page showing a complete report of the day’s findings. Calculated fields would be manually adjustable to account for any changes that occurred since the report was generated or after inspection of the shopping cart results.

Each additional page would have the page of the shopping cart of the related retail site with the items found with an automatic refresh of 5 minutes to keep the cart session alive.

This would all me to review the findings and make the final purchase decisions for each find/shopping cart.

The AI will enter suggested purchase quantity based on expected sales velocity and budget constraints. The AI will know account balances of purchasing methods as well as thresholds for limits on balances. In turn it will be able to estimate quantity to order.

Over time, the AI can form patterns on retailer sales and discount schedules to further assist in decisions as well as suggested order quantities. Aside from these built-in heuristics data, outside sources for determining annual sales velocity and discount rate patterns can be useful to assist in determining when items should be purchased and when they should be sold. For example, Christmas ornaments are extremely seasonal and are typically sold only in the last quarter of the year. However, they are best purchased after December 24-25th when retail stores move to clear out inventory.

These pattern-checking algorithms can then be used to assist in notifying the best time to buy and sell an item down to a month. If possible, google trends data could be utilized.

In my dream the program actually was doing all the purchasing for me. My only job was to receive the boxes, inspect and label them for Amazon FBA and forward them on to Amazon. In the dream, I was making thousands of dollars a day. It was crazy.

I guess the dream interpretation is that I want passive income, I struggle with finding a niche to sell in on Amazon (like most people thinking about selling on Amazon), and my solution is to write a program to do it for me. I’m very much a data analytics guy. Filtering numbers in a spreadsheet to sift out the gold gets me pretty excited. The problem is the sea of data is getting larger and larger each year making manual analysis harder and harder.

I’m curious about this program. It wouldn’t really be that difficult to write.