Business leaders need to use various sources of information for their short term and long term decisions. Data analytics is one such source that is powerful and increasingly affordable for businesses of all sizes.
Most aerospace OEMs and some Tier 1s have major digitization initiatives under the garb of Big Data, IOT, Connected Factories, Command Centers or in Europe as Industry 4.0. MROs are also following suit. There are no longer just 'smart' factories as we strive for 'brilliant' factories according to these marketing campaigns. Current reality in aerospace manufacturing, however, is depressing on this front and meaningful applications are rare at best, even at big manufacturers.
On the other hand, small and medium sized businesses can not be faulted for not adopting it when Big Data evokes visions of terrabytes of data, expensive software packages, big investment in hardware or recurring IT support or cloud expense.
Let's start with a few facts. While computing power and data storage are cheap today, good amount of business knowledge is still needed to inform what the data inputs should be and what is actually being modeled in these data analytics. Nobody wants to be deluged by terrabytes of data and fancy graphics that do not generate fresh insights. Other risks in mindless data analytics include missing critical inputs not obvious to non-experts or relative weighting, reliability or accuracy of available inputs. And yes, running analytics in cloud does not make it any better if all important physics of the problem is not captured in the model.
Moore's law has made computing cheap but GIGO (Garbage In - Garbage out) law will always rule.
In short, data must be informed by insights and vice versa. A good understanding of factory physics is needed for manufacturing focused analytics and a deep knowledge of supply chain risks for supply chain optimization. If done correctly, it is a virtuous cycle indeed, resulting in increasingly deeper insights. You will get not just new ideas but a quantified model to examine what-if scenarios. The added value is the data based prioritization of options and complex combinations thereof that human brain or meetings of experts can not achieve. And it is remarkably quick, repeatable and can learn and improve systematically.
So, what does all this mean for smaller businesses that do not have the luxury of an overhead structure or funds needed to exploit this tool?
Allow me introduce the concept of Small Data, which is a way to analyze only 'relevant and reliable' relatively targeted and small data sets, focused on very well defined / bounded business problems, to drive actionable decisions and to further drive deeper data collection and analysis in turn. An example of such a bounded business problem would be - what combination of shift structure, asset utilization, outsourced operations and production scheduling maximizes cash flow for a desired fill rate?
No need to start by investing hundreds of thousands of dollars for servers, systems and software since an excel sheet or a no-frills discrete event simulation on your laptop can do more than you need.
I do not recommend jumping all in with a software package even if it sounds 'just right' until a) a successful pilot and b) a favorable ROI analysis are complete. You may start with hiring a full time or part time analyst who understands your operations (or has shown the ability to learn and is not just a numbers nerd) and have them partner with your operations leader, assuming they are versed in factory physics. Run a pilot with clear business objectives in mind before launching a big project which, if successful, will eventually require some investment to automate the solution, to plug it into your ERP / PLM tool and to build a user friendly interface for your operators.
And finally, when you do take the data analytics route, make sure it is accessible to and is actually used by people it is supposed to help and not just the analysts and strategists in the office. That's when your theoretical ROI will truly materialize.
In the meantime, enjoy the fruits of Moore's law!