Dell’s Supply Chain Data Analytics Center (SCDnA)

Dell and EMC merged last year, producing a power combination of products including laptops, PC’s, and high end servers.  As part of this shift, Dell moved towards creating a series of Data Analytics Centers across the organization, including one for Sales, Finance, Operations, and Supply Chain.  The term DnA emerged as it was viewed as a hub for unravelling the DNA of data, and the Supply Chain DnA team was born from this effort.

The SCDnA team is led by Shawn Canal, who worked for 7-8 years at EMC on the production floor, and transitioned into an analyst role.  His position is as a senior technical program manager for the center, which has as its mission to explore digital supply chain initiatives, including how suppliers build products, how we integrate suppliers into our processes, and how products perform in the field to establish predictive analytics across the entire life cycle of Dells products, thereby enabling customers.  At any given time, the program has any number of different analytic initiatives going on. One of the larger technology initiatives is focused on creating a Supplier Visibility Tool, originally designed to enable suppliers to manage more inventory on the floor, thereby supporting the “hockey stick” of orders.  With the Dell acquisition, the end of quarter push to sell products is no longer present in a privately-held company, so the focus now of the SVT is more on driving down inventory, increasing working capital velocity, and converting inventory into revenue faster.

The Supply Chain DnA team was the result of bringing together people who were already doing analytic work in different parts of the organization, and centralizing them into a virtual team.  Some were statisticians, others had a masters in math, some were analysts who had done programming in school, but the one common denominator among the team members was that they understood how the organization operated, could perform advanced statistical calculations, and were a natural fit for an analytics team.  The team currently has 21 individuals located around the globe, who report in directly to the supply chain organization, supporting over 93 applications across the business.  They also are able to “pull data” to allow supply chain managers to do their own “self-service / KPI’s” on applications like Excel and Microsoft BI.  They will also work on building new applications for the organizations that perform advanced analytics that drive serviceability and performance.  Applications include part requests, field inventory returns, and others, as well as controlling of information security.  A broad skill set resides in the group.

Shawn notes that the SC DnA team is not part of the IT organization, but recognizes that the team works closely with IT to ensure that the data engineering infrastructure is in place.  “One thing we learned is the importance of working with both the IT organization, as well as with stakeholders in your own organization closely, or it will take much longer to product useful applications.  You cannot just take their requirements input and hope you wind up with something they want.  There is an art to working with the organization to really understand the requirement, but because what they are asking for is often not what they really need.  We have . people from engineering, finance, operations, sales, asking for things, and when questions come in and they say ‘I need this’,  we counter with them that ‘I think you are really asking this question.’”  Some people say I want a raw data dump and we want to look at the data. Others say I don’t know statistics, but I want to know what is interesting!”

In one case, our sales force told us that “We have good telemetric information that comes in every week from the machines at our customers’ sites.  (Telemetric data provides information on each server’s condition and provides clues as to uptime.)   However, we don’t know what to do with this data from the field. Our science team then asked them  – what is it that you looking for?  They discovered that they could use the data to identify which servers would fail using a predictive model, allowing them to act ahead of time to replace servers before they failed.  The data analytics team had to pull the telemetric information and work with the engineering team to understand how the data indicated potential failures, and built out a statistical model. Other projects include building dashboard portals, inventory modeling, and other applications.  There were over $16M of cost savings from such projects.  In each case the analytics team will work with internal functions, but never with customers or suppliers directly.  They are agnostic when it comes to using third party integrated solutions or internally developed tools, whichever works best for the situation, but IT is always part of these discussions.

A good way to portray the analytics problem-solving methodology is designated as “Farm to Table.”.  The following steps are involved.

  • Farm – This involves about taking the raw data and ingesting it into the data lake.. Data is pulled from ERP or internal systems, and the raw data is absorbed into a format where it can be processed.
  • Kitchen – Data engineers will “cook the data”, to standardize and enrich it, and prepare it so that it can be loaded into visual tools such as Tableau or Klik.. There is a lot of self-service analysis done in the supply chain team, and many different web apps are used in this process
  • Order – Business Analysts will develop the application or tool, and present the results to the customer.
  • Value – A feedback loop provides information back to the team, and the value created by the team is quantified.

To drive visual analytics, Shawn notes that “we use a lot of tools such as Tableau, Microsoft BI, and others.  But we don’t try to feed everybody, only those people in the supply chain.  In fact, we report up directly to the President of Supply Chain for our division.  There are other groups such as a Sales DnA, and Finance Dna, that work with other groups.  An external data science team will also work with external data, and we may come together on projects that overlap.”

The DnA also features a Business Intelligence Competency Center (BICC)  which breaks up the tasks into operational 40% and strategic 60%. Two IT representatives serve different roles.  One is our relations manager, and the other is our strategic manager on the technical, infrastructure and technical competency. The Relationship manager works with our IT group to help address collaboration with functional organizations – inventory, sales, engineering, finance, logistics, manufacturing.  The team conducts a task review to look at requests, which occurs once a month to review projects, as well as weekly “deep dive” meetings to vote on priorities for different tasks.

Shawn concludes that collaboration and prioritization based on return on investment for analytics projects is key to success.  “In some cases, we may scope out a project and hand over to IT who tell us how long it will take.  Once a project is approved and we are working on tasks, we give the business weekly updates on the task, and the effort and resources it will take.  When we deliver to the business, we follow up with them to ask “did you realize your value?”  Our funding is driven by the value savings, which also serves as the basis for prioritization.  help them drive down into bite sized things as you go through the project, and not waste resources.  Value is qualitative and quantitative, and we view what we do as enabling another group to make better decisions”.