Home © Rinat Abdullin 🌟 AI Research · Newsletter · ML Labs · About

Business Amnesia

At Lokad we focus on big data analytics and consulting for the retail. This is about doing some small things (sometimes simple, sometimes not) that would companies to be more efficient in their business (e.g.: reducing lost inventory, getting better service level, optimizing replenishment for an inventory of 500k products, etc).

More often than not, I observe two patterns:

  • First, companies have really refined and deep business models. They are really good in what they are doing.
  • Second, business processes of these companies tend to be hindered by their software and it's limitations coming from the previous century.

This hindrance comes in following forms:

  • Business practices in a company don't even have a chance to evolve and adapt simply because legacy software is so fragile and limiting. Something that was once designed to support and serve now resists change by being expensive and risky to change.
  • A lot of money and resources is wasted inefficiently to work around such limitations. Companies can still be extremely profitable, but a few billions of euros less than they could make.
  • A lot of vital business data is lost irreversibly, simply because before for years it was too expensive or impossible to persist it within the existing setups.

So we are arriving at an interesting paradox here. On one hand, a company can be really good in what it is doing in business. On the other hand it can continuously suffer from the progressing amnesia - loosing fine-grained information about actions and decisions linked to individual customers, supliers, products or promotions.

The fact that you can't process your business data today, does not mean, that you will not be able to benefit from it later. It might be useless or it might increase your revenues. However, if you don't save it - you will loose it forever.

Imagine what a sad world we would live in, if health records of every patient were discarded. Detailed history of business operations is such a set of health records. It can be used to find problems or explore opportunities for improvement, which would be unique for each unique business. You just don't throw out such vitual information.

Such general approach to discarding extremely valuable data can be even more surprising, given the fact that these days it becomes increasingly cheap to store data for future reference. Prices start from 15 cent per GB per month (any cloud provider) and go down to 1 cent per GB per month (Amazon Glacier). Keep in mind that 1 year of sales history of 1000 supermarkets can easily fit on a single smartphone.

More often than not, you don't need any rocket science to store your business history for the purposes of the learning and analysis. So please, do it - this a cheap and low-risk way to potentially derive additional value later. Business amnesia, on the other side, can be an extremely risky thing, especially in competitive markets.

Published: September 11, 2012.

🤗 Check out my newsletter! It is about building products with ChatGPT and LLMs: latest news, technical insights and my journey. Check out it out