Overcoming Poor Quality in Manufacturing with Advanced Analytics


By Arthur Lee, Vice President, Product Management, ETQ. He is responsible for leading the product management team, product strategy, go-to-market plans and everything product-related.


We live in a world where brand value is defined by the quality of the product. More often than not, the average consumer is judge, jury, and executioner, deeming poor quality products to be a direct reflection on the brand. As such, it’s critical that quality is kept top-of-mind at every stage of the product lifecycle – from design and manufacturing to shipping and distribution. However, this can be an extremely challenging and complex task.


In the United States alone, for example, there were 7,423 recalls in the 2018 fiscal year.1 Each recall, regardless of product, has a detrimental and potentially devastating effect on brand trust, consumer loyalty and, ultimately, the company balance sheet. The good news for manufacturers is that these recalls can be either minimized or eliminated by introducing advanced analytics into their Quality Management Systems (QMS).


In recent years, disruptive technology trends such as big data and automation have given manufacturers a tech-centric-window into their business practices. The sheer volume of data being produced by machines or digital assets means that manufacturing companies can take a proactive approach to quality, using advanced analytics to identify quality issues before the product begins its customer journey. In addition, manufacturers can utilize software tools to visualize and organize the data itself, gaining operational insights that can prevent a poor-quality product from an almost inevitable recall.


The advanced analytics market is forecast to reach 29.53 billion by 2019,2 thanks to increased adoption rates and a shift toward a proactive, rather than reactive, approach to product quality. The caveat to the increased availability of actionable data is that it creates its own set of pain points, each of which must be addressed if issues arising from poor quality are to be alleviated.


With that in mind, let’s take a look at four quality management challenges that manufacturers are likely to face, both now and in the future.


Challenge 1: Manual Data Collection

The U.S. manufacturing industry is one of the world’s largest markets, producing around 18.2 percent of the physical goods available to consumers.3


Over the last five years, the Industrial Internet of Things (IIoT) has allowed manufacturers to integrate connected devices into their process. These “smart” machines capture valuable data points at each step of the product journey, providing the manufacturer with a holistic view of not only the manufacturing network and industrial assets but also the quality of employee input throughout the process itself. As a result, these IIoT-generated insights can overcome visibility challenges and inform both smarter business decisions and organizational performance.


As noted above, the IIoT generates a phenomenal amount of real-time data. And while the volume of data can seem overwhelming, it provides manufacturers with a significant amount of actionable information. On the flip side, somebody has to manually sift through this data pool to identify and analyze possible pain points.


Without dedicated resources to manually collect data – time, staff or an allocated budget, for example – manufacturers will struggle to create an insightful advanced analytics model with enough relevant data points to reference.


The solution to this conundrum is to invest in data collection automation software tools such as a QMS. Depending on whether the QMS is an independent solution or a collaborative effect (including other supply chain elements, for example), the integration of this software frees up the workforce to focus on implementing the data-driven insights.


Challenge 2: Building Trust with Automated Solutions

Automating the data collection – and subsequent organization – is a crucial part of an advanced analytics solution. However, these insights rely heavily on the extracted data being accurate and, more importantly, trustworthy.


It is somewhat of a no-brainer to say that manufacturers need to be able to rely on the generated data in order to gain quality-related value from this information. Bad or inaccurate data is of little use in terms of product insights, and automated solutions need to be installed correctly from the start.


An automated solution must be set up and installed to remove unintentional bias, while the software itself needs to be “trained” to produce trustworthy results. Filtering out potentially critical information on asset quality, for example, could skew the data in the wrong direction, a pain point that might lead to manual intervention at a later date. Advanced analytics requires a smooth-running system to be effective, otherwise, the insights that it provides will adversely affect product quality.


To avoid this potential barrier to product quality, manufacturers must identify third-party software vendors that not only build and deploy accurate advanced models but also offer dedicated support teams who have expertise in building trustworthy technology solutions.


Challenge 3: Assigning Ownership

We have already highlighted the fact that the process of collecting and analyzing data can be time-consuming.


Taking that into account, it is vital that the manufacturer assigns a specific department or project leader to oversee the data generation itself. This will ensure that the required quality standards are being met, with the onus on that individual or team to intervene if poor quality issues are identified. More often than not, however, manufacturers fail to decide who should have ownership of a data collection project or are limited by the aforementioned time and budgetary restraints.


Depending on the size and structure of an individual manufacturer, data collection and analysis ownership will naturally vary between companies. However, one factor is likely to be the same across the board: consistency. In other words, the people assigned to the task must ensure that the data is managed effectively. If not, then the data will remain unorganized, unstructured and, potentially, unprepared to provide product quality insights.


Challenge 4: Time

Ultimately, the biggest challenge that most organizations face is time. Integrating advanced analytics into an established manufacturing process doesn’t happen overnight. Manufacturers need to be prepared for both a gradual rollout and numerous rounds of testing before they can determine that their solution is ready to go.


In a consumer market where product recalls are both detrimental on numerous levels and have the capacity to go viral, manufacturers must pull out all of the stops to ensure quality standards are met at every step. Taking that into account, advanced analytics is a tool that can prevent the factors that lead to poor quality products, with the added bonus that it enhances the relationship with brands and, importantly, the end user. At the end of the day, our disposable society might not always provide the levels of quality that we want, but there is no reason for manufacturers to not use the quality management tools that are currently available.



1 FDA. (2018, September 30). FDA Recalls Dashboard. Retrieved April 15, 2019, from https://datadashboard.fda.gov/ora/cd/recalls.htm

2 Markets and Markets. (2014, April). Advanced Analytics Market. Retrieved May 10, 2019, from https://www.marketsandmarkets.com/Market-Reports/advanced-analytics-market-58104148.html)

3 Amadeo, K. (2019, February 11). US Manufacturing, Statistics, and Outlook. Retrieved April 15, 2019, from https://www.thebalance.com/u-s-manufacturing-what-it-is-statistics-and-outlook-3305575

1 User Review

Add your review:
Great analysis of the challenges! A transformational Quality Leader once told me that bad news does not get better with age.