Reverse BOM recommendation based on Attributes

sanjeev

Hi,

Consider this scenario for a Diamond Industry supply chain

Demand for Diamond from customer comes with various attributes [cut, carat, clarity, etc etc]
Prospect has made over the years diamonds with various attributes. This historical data is available [both BOM and Recipe].
There is a unique Stage 1 BOM and Stage 1 Recipe. This gives a yield. Based on the yield, Stage 2 BOM and Stage 2 Recipe is decided. And so on till Stage 5 - the output actually of Stage 5 is a saleable diamond
Now the situation is this
1. if there is a demand for a diamond with certain attributes
2. check the historical BOM and Recipe [over 2000 historical BOMs and historical Recipe's available]. These are available Stage 1 to Stage 5 wise
3. based on the yield, Anaplan needs to recommend Stage 1 to Stage 5 BOM and Recipe - Stage 1 gives a specific yield and that determines what BOM and Recipe to pick for Stage 2 and so on

Is this possible to do via Anapan? Any recommendations?

Accepted answers

seymatas1

Hi @sanjeev,
Here are my first thoughts:

There should be a match score to compare demand with the historical BOM and Recipe module for Stages 1-5.
There will be a RANK formula to rank these match scores.
Anaplan will not naturally learn the logic from historical patterns. You should clearly define the logic as mathematical calculations and inputs.
If this is not possible, you can consider using a machine learning algorithm.
If you have enough structured historical data, Anaplan's PlanIQ can create a forecast.

I can share more insights if this seems similar to the solution you have in mind.

Seyma🌷😊

Prajjwal88

Hi @sanjeev ,

Agree with @seymatas1 thought about ranking the historical BOM and Recipe for stages, to add up on the same you can establish a multi staging planning process where firstly it will do Historical data analysis & matching stage by stage followed by Scenario-based recommendations.
The overall process could follow this sequence:
1. Filtering out all historical BOMs matching demand attributes.
2. Stage by stage yield calculation (Like For a selected Stage 1 BOM, calculate the output quantity (based on demand attributes and then Use this output to determine applicable BOMs for Stage 2 via some mapping logic).
3. Build logic to select the “best fit” BOM and Recipe at each stage (could be selected considering the historical probability, efficiency, or closest match).
4. Recommend final recommended BOM/Recipe chain for a given demand line.

Note that some of the key considerations could be- Sizing of data (Considering you have 2000+ BOMs × 5 stages × many attributes), Performance part (For complex logics) and lastly as Seyma mentioned the Scoring Logic (ranking BOMs/recipes using a scoring mechanism based on fit % or historical success rate)

Hope this helps.

All comments

seymatas1

Hi @sanjeev,
Here are my first thoughts:

There should be a match score to compare demand with the historical BOM and Recipe module for Stages 1-5.
There will be a RANK formula to rank these match scores.
Anaplan will not naturally learn the logic from historical patterns. You should clearly define the logic as mathematical calculations and inputs.
If this is not possible, you can consider using a machine learning algorithm.
If you have enough structured historical data, Anaplan's PlanIQ can create a forecast.

I can share more insights if this seems similar to the solution you have in mind.

Seyma🌷😊

sanjeev

hi @seymatas1,

Thanks for replying. Can you pls elaborate on this?

Prajjwal88

Hope this helps.

sanjeev

Thanks @Prajjwal88 for the insights. I have been thinking on similar lines for a while now but been a little not-sure of the approach and needed alternative ideas and thoughts. The scoring mechanism does look very insightful and something that can be made dynamic based on client's criterion.

Just like to point that the approach will actually be a little different than what you suggested. We will have to start with Stage 5 and based on the scoring logic recommend a BOM. And then work backwards to get to the Stage 1 BOM and Recipe.

And thanks for the recommendation on Sizing and Performance. That is something that would be pivotal to the whole solution.

Quick Links

Can you help?

Unanswered questions

Azure blob connection Cloudworks vs ADO
I am trying to connect to Azure blob, the credentials work on cloudworks, but do not work on ADO. Has anyone else experienced this issue?
Berapa nomor Tokopedia?
Tokopedia Menyediakan layanan WhatsApp? Untuk menghubungi layanan Tokopedia melalui (WA)62_08174_9080_16 kamu bisa mengikuti.
Looking for Anaplan End-Users to take a short survey!
We are looking for Anaplan end-users to provide feedback on their experiences with the Excel add-in. Interested individuals will respond to this 5-minute survey to help us understand personal needs and behavior when using the add-in. The feedback provided by survey takers is essential to the roadmap of Anaplan's products.…