3. Product investigations are inefficient because data is isolated, difficult to access, hard to clean
Simple Questions, Hard Answers.
• Which batches of component ended up in a failure batch?
• How many other batches have been contaminated with a suspect component?
• Do bad parts take longer than good parts to manufacture?
• Is there a specific parameter that prescribes product success more than the rest?
• Which components have significant variation in failure rate?
• Is our failure issue material related? And if not, what is driving the failure?
Problem Statement(s)
4. Supply Chain Mapping
Product
Qty: 100
Part A
Qty: 100
Failure
Qty: 1
Part A
Qty: 100
75
Issued
25
Issued Results In
Part
Part
Part
Part B
8. Live Graph Benefits
7:50 AM 9:30 AM
Score/Rank
All material batches were scored
based on relationships to scrap.
The shared batch had the
highest score for batches of that
material, and it was the
component batch used in 7/10
of the worst scrap ranking sub
assemblies
Explore Graph
The graph was queried to
immediately find shared
components from the high scrap
sub assemblies. High scrap
batches were found to all share
a component batch
Issue Found!
Parts from the shared
component batch were
examined and found to have a
physical defect. The component
batch was scrapped and yield at
the sub assembly level returned
to high levels
Scrap Spike
Manufacturing saw a spike in a
specific scrap code on several
batches of sub assembly
The graph database enabled
these steps to be performed with
speed and at scale