It is common practice for manufacturers to outsource large portions of their component manufacture to lower costs and reduce time-to-market. However, as Nitesh Jain writes, it’s not without risks.
As was the case with Boeing’s flagship 787 Dreamliner, outsourcing also carries risks which can result in poorer quality. In today’s highly competitive marketplace, taking high risks is unacceptable. Markets and legislative bodies are becoming increasingly unforgiving should a company breach product quality, especially when the brand promise creates such high expectations.
As the Internet of Things (IoT) gains pace, the opportunities on offer through intelligently utilising M2M data are increasing at an exponential rate. Manufacturers are now turning to M2M data – collecting and analysing it in order to create maintenance forecasts, predict failures and much more. Telematics in vehicles, smart grids, connected wearable technology and more are creating an overwhelming amount of data which manufacturers can harness to ensure quality.
Capitalise on machine data and analytics
In order to maintain high quality throughout the product lifecycle, manufacturers can capitalise on machine data and analytics to shape the quality curve at three points – component manufacture, assembly, and shipped product. Component manufacture and assembly rely on in-house factory data, whilst the third is collected from M2M Data. It is here that the Internet of Things is offering more and more opportunities.
To date, manufacturers have been largely reactive to quality problems that occur once the product has left the factory floor, but this is now changing. Failure at the shipped product stage can be due to a huge number of reasons: poor quality in shipment and installation processes, unsuitable product usage, weather and so on. However, it was not until the creation of interconnected devices that failure could be predicted and even prevented, thanks to sensors collecting data in real time and sending it back to the manufacturer.
What is now needed is a model to manage that data and translate it into meaningful insight. Once this model has been created, it will give manufacturers visibility to the product and customer in real time, meaning issues can be identified sooner, and even fixed in a semi-automated fashion, potentially even before customers become aware of their malfunctioning machines.
Data challenges
There are two main challenges to this, firstly the extraction and separation of useful signals and data from the vast quantities that are collected; and secondly, the final creation of a data structure that can meet new challenges and questions asked of the data collected. Manufacturers, therefore, need to start asking themselves the following questions now:
- Should all data be captured, or should smart filters be applied at the point of capture?
- How do we determine the key metrics so we know what data to capture and transform to the right level of granularity?
- Is our data platform reliable enough to analyse data in motion? Can it work in real time?
- Is our architecture future proof? Can it assimilate newer formats as device usage changes?
By not starting to look for answers to these questions now, manufacturers are putting themselves at risk of falling behind in the quality assurance race. The Internet of Things is here to stay and even if smart devices may not be the norm in all fields yet, they soon will be. By looking to take advantage of M2M data now, manufacturers can be prepared for when the majority of devices are interconnected, be that something as large as a tanker or plane or as small as a kettle, ensuring quality and reassuring their customers.
The author is Nitesh Jain, Wipro
Nitesh Jain is GM & Global Practice Head – Analytics & Insights, Advanced Technologies & Solutions, Wipro Ltd. He is highly experienced in creating analytics projects for a range of vertical sectors. At Wipro he is responsible for providing vision and strategy to the Analytics and BI practice.