Insights from the 2014 Institute of Analytics Winter Symposium: Building a Culture of Analytics

I had the privilege of leading a session at the Winter Symposium of the International Institute of Analytics, with a host of great companies discussing how companies are embedding hardware into products and trying to drive improved supply chain performance in their products.  The topics ranged on a variety of areas, but focused on several important insights.  One of the discussion points that came up is that people need to initiate an analytics project either from the point of view of solving a problem, but also considering the upside of opportunity and innovation, and using analytics as a means to identify unmet needs and market innovations.  Another discussion came up around managing risk, and the need to establish a range of risk that is often narrowed over time as new analytics reduce the window of uncertainty.  An analytics officer from a major consumer manufacturer emphasized how at their organization they valued failure, and in fact would have a celebration of lessons learned around failure.  Sustainability in analytics was another important element, and the challenges of building sustainable supply chains not so much to drive cost savings, but because “our customers expect it.”  Companies who do not pay attention to sustainable supply chains are ultimatey exposing their brands.

Much of the discussion also focused on the “culture of analytics”.  Many executives shared how a company’s culture would drive the appetite for analytics investments, as well as how they respond to analytics.  This can range from “that’s interesting” to “this drives action”.  Part of this is driven by the opportunity and starting with the right project to launch an analytics project.  A large insurer noted that they started with a small problem, and in every case, it involved a 6 week period of pure data cleansing.  But even once the analytics results are identified, there are always questions about whether it is representative of the phenomenon, as often proxy data is required.  Secondly, there is often a group of people who will always question the validity of the data, especially when reporting on risk.

The natural point of departure for the discussion on analytics was organizational agility – or how companies react to risk and intelligence information derived from the supply chain.  Analytics can help drive scenarios and a list of risk mitigation issues.  But what happens if there is a 39% risk?  How do you interpret that?  And what is the organization’s appetite for risk?  Some companies will happily take on that level of risk, while others will wilt and shy away from any scenario that includes that level of risk.  It depends to some extent on the agility of the organization and the ability to quickly shift plans and activities in response to new information.

“Perhaps it is really about the story we tell”, mused on executive.  We need to embed the presentation of analytics in a context that establishes not just risk and problems, but also in terms of opportunity (dare I say “prophecy”?).  Many analytics will be driven into future generations of products, which may trigger maintenance requirements, sales efforts, parts inventory management, and improved product design decisions.  Electrolux was at the meeting, and talking about the “real estate” of the fridge in the kitchen, and how the front of the fridge could have smart sensors used for anything from grocery lists, to consumer preferences, replenishment reminders, and multiple other “smart” applications in the future.

One large manufacturer of harvesting equipment talked about how farmers want optimal seed and planting rates, but is this really in the sandbox of the manufacturer?  “We want to make sure our equipment has 100% uptime during the harvesting and planting cycles, and need to embed analytics in the equipment to be able to service a product within two hours.”  But there may be a need to partner with organizations we wouldn’t have worked with in the past as we explore decisions that farmers need to make, and is there truly a monetization opportunity there?  “Smart Harvest” will monitor the flow of information from sensors on a combine that may change the harvesting rate of its products, but which the operator may or may not choose to use.  These embedded analytics that drive operational settings for harvesting are driven by algorithms that drive the equipment in a manner that never existed in the past.  If people choose to override the combine’s analytics, resulting in a suboptimal decision, how do you measure the performance of the equipment?  The value of analytics can be destroyed by those who choose to ignore them.

A computer manufacturer discussed their new data products in their ecosystem where there are multiple smart connected devices that collect usage information and build data driven products on top of that, as well as applications on usage, settings that will help battery life, and carriers we need to partner with.  This would lead to remote maintenance, and predictive failure analysis.  This is already happening in the aerospace industry, but issues are arising on who owns the data from the engines, with Pratt and Whitney battling it out with the likes of Boeing on data ownership.

In this context it makes sense that a large sports apparel company has a culture of risk taking and in fact have a “Best Failure Award”!  Companies who are able to establish “lessons learned” in a positive way that has a quick agile loop, which in turn builds the confidence you have in your operational analytics, and how you absorb the analytics into your day to day decision-making.  In the end, decisions are made based on both the analytics that support the decision, as well as the executives’ “gut feeling”.  But the gut feeling is a function of how well you know the data, which in turn requires a balanced approach relying on the models, the confidence in the models, and the ground-level understanding of human behavior and supply chain partners.

Another best practice involves presenting the data in a way that convinces the non-quantitative individuals in the audience, to make the data come alive and speak.  Countless problems and challenges and issues lie ahead.  Some of these include “who owns the data in the ecosystem”, “how much is data shared with those in teh supply chain”, “how to partner with other products to create aligned analytics that drive value”, “how to mine the mountain of data exhaust being generated by sensors, bands, vehicles, credit data, sales data, etc”, “how do we establish the legal agreements with clients that ensures they are okay with us monitoring their usage of our products?” and many other issues.

Organizations who are serious about analytics need to think about how to deliver value to customers in a way that can be easily absorbed, and also be able to harvest information in a timely manner that drives insights to someone who creates revenue.  No one wants to be “commoditized”, so the evolution of managed services will be highly dependent on the analytics that puts people in a position to own the service channel.