“We do not wait for the future. We build it.” – lines from a T.V commercial
With all the hype and buzz around data science and advanced analytics, it almost feels like a robotics, automation and machine learning renaissance these days. Much of this hype is well deserved. Remember that technology has always played a major role in transforming our world and it seems we are on the cusp of what we like to call “The Digital Transformation” age.
This article (the first in a series of articles that I plan to write on topics pertaining to application of Machine learning in Supply Chain) focuses on leveraging machine learning in Inventory planning-The Why, the How and the When.
Before you read on further, please keep in mind that these are my views and perspectives. The Data Science domain is rapidly innovating and still in a state of flux. Different practitioners have different perspectives, and this is what makes interactions great as we get to learn from each other. Please feel free to chime in if you don’t agree with any of my views but my request will be to keep the discussion professional.
The Data Science frenzy and its adaptation in Operations and Supply Chain
The Digital age has left its impact on all the key business functions, like Marketing, Finance and Operations. Operations-specifically Supply Chain, from my perspective, has been a unique function, as far as the adoption of “Digital” goes. Supply Chains were the first to embrace technology and automation. Much before there was a discipline called Data Science and before the existence of the word “Big Data”, giant warehouses embraced automation systems and automated production lines worked with precision that could be compared with that of swiss made watches (I have a thing for swiss made watches).
However, I believe that the same can’t be said about adaptation of machine learning methodologies in Supply Chains. Though Industry 4.0 is already here creating its own buzz, and we are actively discussing how Big Data technologies can “transform” Supply Chains, the real adaptation has been rather slow. There obviously are best in class companies who are taking the lead and working hard and strong on getting that competitive edge by being the first ones to harness the power of data science, and more specifically, harness the power of Machine learning.
Since this article is focused on Inventory, I will keep that focus this point onward. We will talk about application of machine learning in Inventory analytics, but before that, we need to get a quick perspective on where we currently stand in that arena.
A very very brief history of Advanced analytics in Inventory management
The inventory analytics journey that started in early 1990s with the classic deterministic EOQ model, unfortunately, has not come far enough from there, as far as the Inventory analytics methodologies go. As we know, the primary trade offs in the classic EOQ model are holding cost vs fixed costs and economies of scale. Inventory management in the real world is much mode than just there parameters. Then the Newsvendor model was adopted, primarily to address the variability aspect and from that evolved the classic Inventory models that take variability into account (Stochastic Inventory model-if you like technical jargon).
The current widely used Inventory models like Base stock, Order upto and additional variations, are all focused primarily on Safety stock and build upon the variability aspect addressed by the Newsvendor model. However, in my perspective, this is where the journey slowed down. We got too fixated with Safety Stock calculations. While managing Safety Stocks are crucial, the standard models that are used are very standardized, with only a handful of parameters being used/considered to calculate Inventory requirements and manage inventory.
The only primary modifications to classic inventory optimization/analytics processes that I have observed recently are:
- Better demand categorization/classification: Focus on leveraging data science has led to some more sophistication in Inventory models, by defining the demand distribution better. Rather than force fitting every demand pattern into a Normal distribution, we can be more precise in defining the distribution type. However, the calculation parameters are still the same.
- Forecasting is evolving: Since we typically use Forecasted demand, and forecasting algorithms are evolving fast-the models are getting better than before but the pace is not significant enough (my perspective).
So what is not exactly right with the current methods?
My belief is-it is our fixation with calculating Safety Stocks. This very fixation then gyrates us into using the same models again, and again and again, without putting too much thought into what goes into these calculations and how these calculations can be improved/refined.
I have always been passionate about applying advanced analytics in Supply Chain and operations domain, but the analytics I love applying is the one that would deliver results and would consider the real world challenges. A couple of years ago, I wrote few articles on LinkedIn and my blog site, focused on practical and real world aspects of leveraging Inventory models. If you are interested in exploring further, the links are below:
The gist of these articles, and the crux of the problem with current Inventory management modeling approach (in my belief), is that the real world challenges extend way beyond the simple parameters and assumptions that form the underlying basis of the classic Inventory models that we use.
Why this has not been challenged enough?
I don’t have the full answer to it but I believe that at the heart of it is the plethora of Off the shelf “Multi-Echelon Inventory Optimization” tools (Phew….so many buzzwords there) that are out there in the market. The companies selling these have done a great marketing job in claiming each new solution to be significant improvement over others/previous solutions.
The fact is, the underlying logic in most of these tools is still the same. No matter how many fancy functionalities are incorporated in these tools, the underlying calculations are still pretty simple, and in my opinion, archaic. However, these providers have been able to create enough buzz in the ecosystem, that makes us believe that significant advances are being made in the field of Inventory planning analytics.
There has been advances though-but they have been mostly on the technology front-not on the logic side. With advances in technology, even a powerful laptop can host a “Multi-Echelon Inventory Optimization” tool , which in turn can perform millions of calculations. This should excite us-but not to the extent of believing that we have made advances in the approach of Inventory modeling. All we have done is to enhance the capability of calculating millions of values using an archaic model.
Millions of calculations are great-but how useful are those values? Will those values make an impact on your Inventory? Will it help you become a best in class organization as Inventory planning and management goes?
We seem to forget the key aspect-the win is not in successfully implementing a solution. The win is getting the maximum value out of that solution, win is leveraging that solution to solve your business/operational problems. The world is full of stories of “suceessfull” ERP implementations (back when ERP was the buzz) that delivered little to no value, and in some (unique) instances led companies to file for bankruptcy.
So how can Machine learning help?
In order to avoid this article being a mile long, I have decided to split it into two parts. In the second part, I plan to cover:
- Example of how machine learning algorithms can help do more effective inventory planning and why it is better than classical approach
- What are the key challenges in implementing such solutions-going beyond the “perfect world” theories
The second part can be accessed on my blog using this link: