Think of big data analytics (or any type of analytics for that matter) as a product. It follows a product life cycle just like any other product. Early in the product life cycle there is a lot of variability in the product being delivered and volumes are low. The correct production process is a job shop where skilled technicians (e.g. data scientists) can produce one-off analyses. Cost is not an order winner, but an order qualifier. The order winner is “get me insights from my data as quickly as possible” which means quicker than my competition and may be measured in days or weeks (i.e. not necessarily real quick).
As the product moves through its product life cycle volumes start to increase and product variability decreases. Some standardization is occurring as we learn more about what our customers truly need in terms of analyzed data. Order winners start to shift from delivery and quality towards cost. But it is interesting to note that even though delivery and quality start to fade as order winners, delivery speed and analysis quality are actually improving. The data analysis process needs to move out of the job shop and into batch production. The workers doing the analysis do not need to have the same level of technical skills as in the job shop. This standardization allows for more specialized machines (software applications) to start taking the place of skilled technicians.
As the product starts to mature in its product life cycle the volumes are even higher and the analysis product is more standardized. Cost is the order winner, but even though delivery and quality are order qualifiers, the order qualifying level is very high for both – “I want the analysis now and it needs to be right as the customer is on the phone”. Not only must the data analysis be automated, but also some or all of the decision making needs to be automated. A standardized production process incorporating significant automation through IT is required to be competitive.
Understanding the product life cycle and how it impacts operations strategy is critical to an organization’s success. All products move through the product life cycle. Some move slowly while others move quickly. It is important to recognize that the production process that was successful when the product was in its infancy will not be successful as it matures. Skilled labor will be replaced by specialized machines freeing up the skilled labor to work on new products. The maturing product will move out of the job shop and into a production line (automated process using specialized machines). This is true for all products – it does not matter if it is a manufactured good or a service such as data analytics.
Note that a line process produces products faster, with higher quality and at lower cost than batch processes or job shops if the volumes are high enough. The faster you can move a product along its product life cycle, the sooner you can take advantage of automation. The effective use of IT can actually help us to move products through the product life cycle by implementing mass customization.
Mass customization when done well accomplishes two things. It makes the product look customized to the customer with high variability and low volume placing it early in the product life cycle where higher prices can be charged. At the same time it makes the product look standardized to the production process with low variability and high volume placing it late in the product life cycle where a line process can be used reducing the cost of production.