Predictive analytics in the supply chain Handfield

By Rob Handfield, Director, Supply Chain Resource Cooperative, Poole College of Management, NC State University

Robert HandfieldPredictive analytics in the supply chain is a “hot” topic.  I have put together what I believe are some real opportunities for driving improvement across a number of organizational areas that will prove to be important.  In my experience, organizations pursuing improvement in these areas are going to see results beyond the traditional “cost savings” focus of many procurement areas that continue to pursue traditional leveraging opportunities.

 

Prediction 1 – Spend Analytics Will Become More Predictive (Not Backward Looking)

My recent interviews with a number of executives leads me to believe that driving where the organization is going is a lot more important than looking in the past.  Most spend management looks at historical spending – which is interesting, but doesn’t tell you where the puck is going!  As one executive I met with said:

 

“It is important to have meaningful data – but what you do with it is the issue.   Can you form insights that are actionable? That is the real question!”

What this individual was emphasizing is that asking the right questions about our data will become important.  Spend analysis can certainly provide one set of insights – but only provides a partial view.  Organizations need to be able to use this information to provide greater insights into future spending requirements, and provide views that will support business strategy and greater views on what will be important going forward.

 

In one company, the analytical team explored the relationship between economic activity, consumer demand, and demand for their products. They discovered that there were several leading indicators of demand, including coal (a 10 month leading indicator). Other indicators included data pulled from the Federal Reserve economic database, the Producer Price Index, the Purchasing Managers Index, as well as internal metrics such as the size of the company’s own sales force. Using multiple regression models, a predictive forecasting model was developed which allows users to input data into the model, and develop specific forecasts for different categories of products purchased from overseas suppliers. The model is also being used to adjust company revenue and budget growth estimates – and can effectively be used to either curb or expand budgeted growth estimates based on these economic forecasting models.

 

Prediction 2 – Incident prediction and workflow management systems will replace supplier risk monitoring on major projects and indirect areas

Incident prediction means understanding what issues are on the horizon, not just what are current risks.  To ensure monitoring of current workflow systems, real-time systems will be needed to collect worker feedback, from those that are closest to the action. In addition, it will require simple, holistic metrics that capture all areas of business risk for contract factories, and which encompass both current and future risks.  We may also see forums for industry collaboration that identifies best practices, shared insights, and an opportunity to drive a standard approach to what are emerging as very large and complex risks in the global supply chain.  On-going research is needed to capture social media, “big data”, qualitative reporting insights, and other non-traditional data to enable predictive insight, and build a shared source of truth for factory and supply chain risk.  New insights are needed into risk mitigation practices that go beyond simply “avoiding” risky production locations when a decision is required , but instead drive better business decisions and improved community impacts for sustainable supply.

 

Prediction 3 – Corporate responsibility (diversity, environment, labor and human rights) will become an integral part of the sourcing and risk management process.

The recent publicity generated by the Rana Plaza disaster in Bangladesh, and other labor risks, has elevated the importance of integrating these issues into the sourcing decision.  There are many different data resources that are potentially available to augment and create a rich set of metrics that can be used to drive insights and warning to procurement and senior management.  Organizations will focus on creating centers of excellence tasked with creating indices that provides a quantitative and visual representation of the supply disruption risks that exist in the global supply chain, as well as the related financial cost impacts associated with these issues. Such indices should be proactive in nature, and provide an early warning system as well as an estimate of potential financial impact of diversity, environmental, and human labor rights violations to procurement.  These centers should be not tasked not just with mitigating risks, but to provide early warning and a dashboard that can be used to alert management and serve as an early warning mechanism of possible threats to the supply chain, and the relevant financial impacts to the organization.

 

Prediction 4 – Organizations will build stronger modeling capabilities to plan and manage future supply chain talent requirements.

As we’ve noted in prior posts, talent is the key to procurement organizations.  However, most organizations don’t think of talent as a critical input – the assumption is that you can find the right people “on demand”.  Unfortunately, this is proving to be a big assumption that is not panning out.  Organizations are finding a critical shortage of talent for many of the roles they are seeking. Talent should be part of building a procurement transformation, and not an after-thought.

 

Some of the research we’ve done has identified predictive models that consider the future requirements and skill sets  in different areas of supply management.  Given this future state, the model considers the current state, which includes a number of parameters including Staff Pipeline, Candidate Pipeline, Min/Max/Most Likely Inputs for each Recruiting Activity, % Interviewed, % Hirable, and % Accepting.  In addition, the model should consider retention rates for employees.  This is particularly challenging, given the many different companies that “Millenials” will move around with over a period of time.  Efforts for building talent should be a partnership between human resources and supply management, to truly think about how to achieve the right long-term outcomes.

 

Prediction 5 – Post-Award Contract Management (SRM) will become the Biggest Single Source of Sustainable Cost Reduction

One executive at a major oil and gas company stated this very succinctly:

Our existing systems and ways of gathering data and information is adequate for the category management or process – the pre-award work. We can you what is happening in terms of how much we spend in this category, what business unit level spending we have, what types of things we are buying, and derive “good enough” information to do strategy work and enough consumption information to negotiate volume tenders around the world. But were we fall down – is where we believe 80%+ of our opportunity for continuous improvement exists – which is in the brownfield post-award stuff. For example, do we have the information on when we are buying energy – during peak hours or not? How is the service or consumption information being used in real time at the asset to drive savings, and where is the analytics for that?

 

The opportunity to drive post-award contract management will continue to be critical.  Organizations I’ve spoken with recognize that this is the key.  One company I met with last week has a focus on improving productivity in its Taco Bell, KFC, and Pizza Hut brands, by considering not just the price of food, but also the operational characteristics of food preparation that drive a total cost solution for many of the purchases they are making.  This level of operational and price-based management will be a key focus for procurement in the future.