Data Scientists Suck At Demand Planning! | Interview With Author Nicolas Vandeput
ฝัง
- เผยแพร่เมื่อ 20 ก.ค. 2024
- IBF on Demand Episode 9 (28.6.20)
Sponsored by Arkieva, your one plan S&OP software. Learn more about Arkieva's approach to demand planning here: arkieva.com/software/demand-p...
Learn about IBF 's world-leading consulting services to build or transform demand planning here: bit.ly/3d58dMT
Data Scientists Suck At Demand Planning! Or do they? Can they learn the required skills? Conversely, can demand planners learn new data science skills? I interview author of 'Data Science For Supply Chain Forecast', Nicholas Vandeput, and examine the differences and similarities between data scientists and demand planners, and how we need to learn from each other.
02:12 Forecast accuracy by university degree
04:30 Data scientists make horrible demand planners
06:16 Mindset of data scientists
07:35 Experimenting in data science
12:55 Differences between data scientist and demand planning roles
14:04 Will data scientists do the modeling then hand over to a demand planner?
16:45 Should we get data scientists into demand planning, or do we keep them as separate roles?
17:00 Human judgement, collaboration and data science
20:10 Will data science be part of supply chain or will demand planning be a data science function?
20:25 The value if an independent analytics function
25:48 The importance of knowing new models and techniques
29:13 Demand planners need to adapt
34:50 IBF maturity assessments to establish forecasting and analytics functions
35:19 The ultimate goal of a centralized planning and data scientist function
Find out more about IBF at www.ibf.org and www.demand-planning.com
Find out about Arkieva at www.arkieva.com
I like to say that I know enough about data analytics to get myself into trouble! I started my career as a SQL data analyst with a dabble of R, but ultimately what I was trying to do was to find a way to get my foot in the door with supply chain. I definitely view myself as a supply chain professional with data skills, not a data person with supply chain skills. Anyone who remains agile in learning knows that we are at a time where it is more difficult to thrive as a supply chain professional without data skills, and data has been known as the big disruption in supply chain for sometime now. Concerning the intersection of data science and planning, don’t you think it depends on whether you are demand driven planning vs forecast driven planning? We use a software product where I work that does a very good job using historical sales data to create demand driven forecasts, but the struggle and ambiguity is with forecasting for something with no history. This is where S&OP would need to take precedence over data and where, like in your soft skills video mentioned, relationship building and agile learning would be paramount. Perhaps S&OP needs to be redefined as S&DOP.
Excellent podcast and contribution from Nicolas.
First of all: Erik congratulations for the achievement of over 1000 followers on you tube.
Although the two functions are very different, I am convinced that between data scientist and demand planner there is a clear need today of a necessary cross fertilization. Mind set change is required. Also the new approach will require Discipline in order to facilitate the implementation of a new model /process as new standard way of working.
Creating a data scietific function in the supply chain is a valuable tool to facilitate the transition to new models that leverage the specific skills and knowledge (and relationships) of each functional group.
I'm also convinced that this is a transitional solution with the aim to create new skills for demand planners that will interact more and more with powerfull machine!
Be curios, ask why, look at interaction, stay knowledgeable.
Ciao, Giuseppe
This is an amazing interview, what about the future of demand planing being 90% automated, with optimisation in real time.
Can you share the source for the survey about finance, supply chain and engineers being the worst at demand planning? :) I'd love to see the full list of who they surveyed, and who appears to be the best
That vid was obviously nice :)
But at the same time I was looking for some more confrontation with more sticking to the title and in the end you end just being nice to each other and not questioning data science role in DP too much :>
What I mean is that even when I am on the data science role in the supply chain I think this is still a serious threat to present a data science solutions as a magical boxes that can fix all the problems. We all know that data science do not apply always and everywhere and denying that fact leads to the failures in many projects that are trying to, for example, "forecast unforecastable" (spectal entropy close to one :) )
As prof Hyndman writes: "Every week I get sent papers proposing new forecasting methods that fail to do better than even the simplest benchmark. They are rejected without review." and myself i have seen situations where LSTMs couldn't beat a random walk or seasonal naive forecast and yet had been presented to the client beforehand that it will be inevitably "silly old models"...
Cheers! :)
Did you happen to read Carlos comments for this same episode 😊 I know exactly what you mean and what you are talking about. I know it was last year and I am just now responding but maybe I can be more confrontational in 2021….
Good podcast, however I have to disagree on some things with you and Nicolas.
Both demand planners and data scientists have to have a holistic focus on their function.
This is part of the upskilling and reskillimg process we all have to go through in these days.
Everyone involved in any function of Supply Chain Management has to have a holistic focus, each one from their functions allowing the supply chain to function as a system.
If you tell them to focus in just one thing, the result will be sub-optimization, which is the common performance we have in inmature organizational cultures.
No, I agree with you. I admit I did over simplify and stereotype for the podcast and entertainment purposes. My hope is that it would make people think though and get out of their own comfort zones a little bit more and be even MORE holistic. Good comments and sorry it took me 7 months to respond.