BIG DATA AND AI WORLD / 22-23 MAI IN FRANKFURT  •  Visit our Booth I140 at Hall 8! Find further information on event + registration.    
How to calculate return on technology investment in supply chains – in 3 steps
Blog

How to calculate return on technology investment in supply chains – in 3 steps

4 min read • Apr 15, 2024

Milan Listeš
Business Development Manager Data & AI

Discussing technology w/o giving it a business-relevant context is like standing outside a stadium to watch a game.

I have just made this comparison up and I’m not sure if it makes sense, but the whole point is that if you are a business organization, giving technology investment some context is crucial for project success. If you don’t do this, one of the following things will probably happen: 

  1. You will wave your hand when someone mentions “AI” and you’ll say “this is too complex for us” 
  2. You will wave your hand when someone mentions “Inventory optimization solution” and you’ll say “we already have it” 
  3. You will enter into projects w/o a clear goal of what are you trying to achieve and you’ll end up implementing just another solution which you don’t use 

As technology investment, like any other, is going to take up your resources, particularly in terms of time and money, the best way to give it some context is to understand and calculate the potential business benefits it will bring to your business. 

Let’s do this together in three steps. 

For the next three weeks we will reveal these steps to you, so don't miss the chance to stay up to date. 

Step 1: understand your current situation and especially your pain points

Many organizations in the commerce business rely on manual planning and they often do it w/o understanding market demand. Some typical pitfalls we see on the market include:   

  • Planning strictly by looking at min/max values listed in ERP;  
  • Looking strictly at the previous year’s sales data and using it to plan demand for the following year; 
  • Or only setting some fixed parameters like Stock Coverage in Days and doing all your activities in order to achieve this goal. 

Companies that do organize their supply chain planning in such a way often face the following challenges: 

  • They focus only on internal transactional data w/o understanding market dynamics which range from simple, easy to understand things such as weather forecasts or holiday seasons, up to more-complex-to-integrate datasets such as competitor pricing, market share or average purchasing power by market regions. 
  • After gathering basic data, they analyze it manually in various Excel-based “systems” and try to create calculations on how they should replenish stock in their central warehouse or one of their many stores. The problem with this step is that it’s prone to human error. Such mistakes are highly likely to happen, and it’s just a question of how big an impact they will make in your overstock / out-of-stock calculation. 
  • When the customer finally understands the requirements from the market, then there are many people involved in defining the complex matrix of the numbers of SKU’s which should be ordered to each store and when. This complexity almost always leads organizations into overstock situations in some stores, out-of-stock situations in other stores, and heavy discounting of some products etc.

If we split these challenges into more details that require improvement in complex commerce organizations, we have the following areas listed below ( an example of a retail organization). 

  • Highly accurate demand forecasting based on Machine Learning which will take into account all relevant 3rd party data such as weather, holiday season etc., and replace various manual planning activities such as copy-pasting from the previous year or using basic statistical functions to forecast demand 
  • Automated store replenishment recommendations which will not be based on static safe stock levels, and which will take into account all other relevant events like promotions that usually have high one-off impacts to your sales 
  • Initial store/brand allocation based on detecting similar items and stores. Key improvement in this area should come from detecting those similar segments or stores based on customer behavior and not only on customer demographics data 
  • Automated vendor replenishment recommendations for warehouses which will take into account various vendor-specific constraints like MOQ (minimum order quantity) or MOV (minimum order value), but also take into account all cargo specifics so that you can make sure that a container or a truck you order is optimally filled and not shipped half-empty. This will reduce your transportation costs, and your sustainability manager will also like this.

undefined

When you realize these pitfalls, the next step is to define and quantify your goals, but please be honest and realistic. Follow us on our social media platforms and we will share more details about the second step with you next week. 


Like what you read?

Subscribe to our Newsletter and get relevant updates …

Subscribe
Topics
Artificial Inteligence
About the Author

Milan Listeš

Business Development Manager Data & AI

Continue reading

Blog
Commerce

Data Science Adria Conference: The Future of Supply Chain from BE-terna's Point of View
Read more
Blog
Manufacturing

AI solution for your company: The AI Builder in action
Read more
Blog
Fashion, Distribution, Commerce

Recommendation systems in E-commerce: What’s the thing you’ve never known, but always wanted to?
Read more