Scope, Granularity and Scale

We need to see both the details and the larger picture; people in business need software to help.

We need to see both the details and the larger picture; people in business need software to help.

Can you understand and solve a broad or large-scale problem with a high-level, low granularity view? Usually the answer is no.

That makes for some major challenges, as we have only a limited capacity to work with many items at the same time. Granular detail across a broad scope is very difficult to fathom. Fortunately, that is where technology such as models, algorithms, and software that includes those can come in handy.

First, let’s look at how we humans tend to view issues. Often we use less granular views as the scope increases, and more as we drill into a narrower area. However as the old saying goes, “The Devil is in the details.” Truer words were never spoken.

So what is the way forward? How can you effectively solve broad-scope, large-scale problems? By capturing all of the details needed, often in a smaller-scope view, and boiling that up into the broader scope.

Sounds easy – like reducing a liquid when you are cooking…

Yet it is not so easy to do in decision making.

The trick is to have all the detail available, but in a context where it’s become meaningful. This meaningful aggregate or meta-data can be a starting point for viewing the broad scope challenge, but to actually make sound decisions, the details nearly always come into play. Fortunately, we have technology that allows us to stay at the level of problem that is holistic and still base decisions on accurate and complete details.

Examples of software approaches to this:

  • complete digital twin models of products and processes during design – so that every element of something physical is modeled virtually before it becomes real (Dassault Systemes and Siemens PLM are strong in this)
  • algorithmic approaches to supply chain planning – based on the uncertainty and variability of demand and operations (GAINSystems, Llamasoft, ToolsGroup have some leading solutions)
  • multi-level models of operations as are available in modern global MES/MOM software that accommodate process, cell, plant, enterprise, and supply partners (Dassault DELMIA/Apriso, Siemens PLM Camstar)
  • big data analytics and discovery can be used in many ways, but the core is bringing diverse data together into a meaningful context, then conducting analysis to pinpoint sources of performance challenges, finding patterns, and conducting what-if scenarios are a powerful new frontier for manufacturing companies. A key concern in this arena is that the data be clean, normalized and comprehensive. (Siemens Cloud Omneo does all of that plus the analysis for product data; IBM SPSS, Microsoft, Qlik, SAS, and many others provide more general-purpose big data analysis)

As the world becomes more tightly connected and complex (think streams of data coming in from IoT everywhere), all of these types of solutions (large scope but based on fine detail) will become more critical.

People must have such tools to make good decisions. More and more decisions are large scope and scale, but must be based on finely granular data. We must look at both each grain of sand and the entire beach. The time for trade-offs is long past.

Tags: , , , , , , , , , , , , , , , , , , , , , ,

No comments yet.

Leave a Reply