Advanced Analytics for Product Portfolio Optimization in Manufacturing
Updated: Feb 6
Industries around the world are innovating new products and services to make a difference in the continuously evolving world. Although the manufacturing sector has seen a very steady and slow growth in demand for the goods at around 3-3.5%, it still forms an essential part of most industries across the globe. With the digitization and technological revolution of most industries across the globe, manufacturing industry can serve as a good canvas for advanced analytics to play out.
How can Big Data aid in manufacturing?
A Forbes report published in 2016 stated that around 70% of the manufacturers had already started incorporating Big Data and advanced analytics into their regular operations. Most manufacturers strongly believe in the concept of Industrial Internet of Things (IIoT) and claim that data analytics is a vital part of the same. Most industry players already believe in analytics and its ability to improve the process of producing better goods. Hence, the application of advanced analytics should be scoped out and implemented at each stage of manufacturing.
A typical manufacturing company follows more or less the same procedure to convert an idea to a market-worthy product. The first stage is essentially finalizing the concept or the idea behind the product. This is the R&D phase where the function, design, and usability of the product is put to place. What follows next is the engineering and design phase wherein we get to see the prototype and blueprints of the products. This is then followed by the development and testing phase where various industry standards, product safety requirements, functional constraints are checked for after the actual product has been manufactured. The commercialization and launch phase after testing helps manufacturers make the product available to the market and hence, assess its early performance. One can monitor the early sales trends of the product at this stage.
Having stated the various stages of the manufacturing process, it is a safe claim to make that data analytics can improve the efficiency and productivity of each of these stages. It is important to note that advanced analytics might not be required to glean insights from the data available for each phase. However, a few tweaks on the data and some basic analysis can help unveil the scope of improvement of multiple processes.
Advanced Analytics in Product innovation and ideation
In product innovation, advanced analytics can transform innovation and development in multiple fields like CPG, materials science and semiconductor industry, synthetic biology, life sciences and even heavy machinery and automotive sector. A lot of leading pharmaceutical companies, like Bristol-Myers Squibb are already using data and analytics to help them with drug discovery and prediction of drug performance in the market once launched.
To be specific, data from various sources, like research papers, clinical trials and even reviews from patients and doctors can be used to better determine the chemical compounds that could be used as drug treatments for a variety of diseases, even the unconventional rare ones. A lot of other companies like the ones in the consumer electronics use crowd-sourcing and data from social media channels to design their next product. Advanced analytics on such data-sources could help companies predict such upcoming trends and stay ahead in the industry.
Manufacturers like Rolls Royce, Tesla and Aerospace are using sensors to collect data on various parameters. This further helps to identify features that shall aid in the next products’ development or revamp process. Moreover, insights can be derived by doing some basic analysis on suppliers’ manufacturing processes and materials can aid in better design and accelerate the time-to-market.
Such analyses help companies gain an understanding of their ability to complete a certain process, foresee the complexities in the process, create timelines and thus, meet their budget and time-to-market targets. Statistical modeling of data on the complexity of projects, for instance, the addition of a certain type of material in IC chips can inform teams in the semiconductor industry about how it could affect the production time and even the performance of the chip when simulated for different real-world scenarios.
Most CPG companies analyze the performance of their product portfolio using historical product information. They use this data to innovate new products and remove a few suboptimal products to optimize their existing product portfolio and thus, maximize their gains. Firms can also identify similar products using their functional and physical attributes to make better pricing and purchasing decisions. Time series forecasting or advanced decision trees can then be used to predict if a newly introduced product with similar attributes could perform better given different time frames or market scenarios.
This analyses could also feed into simulations and recommendation engines for industry players to indicate what kind of products the market is favorable towards. Last but not the least, analysis of intellectual property, research patents and whitepapers provides critical information to design a product that is legally sound, innovative and potentially successful.
Data Analytics in Product engineering, development, and testing
Failure modes and effects analysis (FMEA) is a systematic and standardized approach for identifying all potential failures in a design, a manufacturing or assembly process, or a product or service. It is normally done at the early conceptual phase of designing a prototype. Analysis of product-level data can provide insights on when a particular prototype could fail and what corrective measures should be taken to address such problems. Logic and rules designed on the basis of this data can help manufacturers perform a more effective FMEA. This will help them design better parts and assemblies. Analysis of product level information and the corresponding manufacturing equipment and processes used also promote standardization by harvesting old modules from existing databases.
Data from manufacturing instruments and processes is one of the most significant information a manufacturing plant can capture. Multivariate analysis of metrics for processes, products, and machine-state can help firms optimize many important manufacturing dimensions like yield, throughput, equipment availability and operating costs. IoT plays a major role in this phase as sensor data and machine readings can form the foundation of such analyses.
A lot of companies dealing with heavy machine manufacturing conduct stress tests on their machines and use this data to predict machine failures. Running simple statistical models like regression and correlation tests can help them identify the most optimum conditions for the machines’ operations. Predictive models can also help floors to identify how to improve yield and throughput of machines. This will also require a lot of mathematical tests and algorithmic transformations on data before building a model.
Analyzing the bill of materials (BOM) can also help firms set the optimal product cost. Additionally, analysis of the composition of a product, quality-inspections, third-party lab testing of a products’ BOM can also give insights on safety standards, product quality, and hazardous materials. This can help in ensuring regulatory compliance and can aid in more efficient and faster product development. Enterprises use supply chain analysis to minimize resource wastage while transferring the product from the production line to the stores or warehouses.
Advanced Analytics in Post-production and sales
Advanced analytics can be used in the post-product launch phase in making product pricing decisions, identifying the target audience and managing the post-sales operations like, product recalls, discounts, product improvisations, and even customization. Strategic product pricing is one of the biggest application of data analytics in the manufacturing industry. Automotive giants like General Motors, Nissan and Toyota have used competitor price data, cost insights from the production phase and insights from lifecycle pricing of similar products in the past to come to an optimal price point. Pricing post audits and A/B testing are some statistical tests and procedures that industry experts use to recommend optimal prices for products.
Furthermore, analyzing local laws, competitor data and customer feedback could also help firms design product-packaging and labeling-specifications. This would aid in designing effective marketing campaigns and ensure maximum market penetration. Customer lifetime value (CLTV) and customer-portfolio optimization are some common statistical problems that are solved across industrial verticals. Such processes can help firms identify the best customer set for a particular product and even predict how long the customers are going to stay loyal to the product or the brand. Insights from assessing local-markets and customer preferences help in designing and launching customized variants and tweaking products to satisfy consumer segments across various demographics.
After-market services, warranty analysis, spares and service facilities generate enormous amounts of data that, if analyzed, can provide recommendations on how to optimize the product portfolio, improve customer loyalty, and in effect maximize the market share and brand reputation. Product recalls, prove to be extremely expensive to a manufacturer. However, when analyzed using customer data, warranty claims data and even recall notices issued by governments and firms provide crucial insights into flaws in the existing product development process. Timely analysis of the same provides the opportunity for correction and improving the brand-image before the losses have magnified.
Although there are a lot of tests and mathematical proofs for various stages of the manufacturing process, it is necessary that firms acknowledge the need for such steps in the first place. Firms then need to either build an analytical muscle within the organization or sign contracts with firms that are capable of solving such problems. Data availability might be another issue that most enterprises might need to battle. However, one can solve that by setting up good data-capture and storage infrastructure across all subgroups in an organization. Once all these challenges, beyond others, are won, companies should ensure that all their processes and decisions are data-driven. This would enable them to maximize their earnings with minimal losses.
Wish to leverage big data analytics to grow your business tremendously? Contact Datahut, your big data experts.