Advanced analytics team up people and machines to create dynamic supply chains

embracing advanced analytics, the venture magazineAdvanced analytics have taken over the world of sport. Big data completely changed the way we look at cricket, rugby, and footy. Batsmen are rated by wagon wheels, fly halfs by points per minute, and full forwards by probability of goal. The data doesn’t just give viewers more to think about during the match or give fantasy players something to agonise over when choosing their teams. It helps the players perform better. The case is the same in supply chains, where statistics and predictive modelling are helping to increase efficiency.

Big data measurements

Much like a sporting match, a supply chain is full of things that can be measured statistically. There’s inventory, production, demand, speed of delivery. When perishables are involved, there are expiry dates. Order too much or fail to get something shipped before it’s expired and you’ve wasted money. Fail to order enough, and you have impatient customers waiting for their orders. Big data is here to help.

Measuring month-over-month or year-over-year growth helps businesses notice patterns and be able to predict how much inventory will be needed or project sales. Spotting these trends is the first step in analysing why a pattern developed. Was there a specific event that sparked demand? “Fiji Water Girl” Kelleth Cuthbert made headlines by photobombing celebrities on the red carpet at January’s Golden Globe Awards. While that caused a surge in sales, it’s not something Fiji can count on occurring every January.

Once all that data is gathered — through data mining and machine learning — supply chains can insert them into predictive models that help prepare for the future and lend insight into what would happen if certain changes were made. Would improved customer satisfaction caused by faster shipping times yield a large enough sales increase to justify hiring more couriers? Would a different warehouse layout improve efficiency? With advanced analytics, companies can figure out those answers before making costly changes that might backfire.

New careers

While artificial intelligence and machine learning are instrumental in the collection and analysis of all this data, they aren’t doing it alone. Rather than eliminating jobs performed by humans, advanced analytics have created new ones, such as digital engineer. Digital engineers are big data scientists who are responsible for managing the algorithms that guide automated decision-making. In other words, digital engineers tell AI what to think and how to learn. They train AI to better understand human language and behaviour to run more smoothly. Conversely, they relay the information gleaned from machine learning to executives in layman’s terms. Think of them as robot whisperers.

These digital engineers are in high demand as more companies shift to a flexible, dynamic supply chain model with end-to-end segmentation. Universities across Australia are offering degrees in digital engineering, and average salaries hover around $80,000. Companies such as Chainalytics and WSP provide outside analytics consulting complete with digital engineers who design AI that can do everything from coach workers on the shop floor on correct procedures to identifying the most effective ways to engage with important suppliers. Other companies, such as fashion retailer The Iconic and Ricoh electronics have in-house engineering teams to handle large multinational supply chains.

A study by Accenture found that 63 per cent of business leaders believe AI will result in net job gains for their organisations in the next three years, and that 62 per cent of people think AI will have a positive impact on their work. This embracing of big data analytics and collaboration with machines will improve efficiency with supply chains and spur sales growth. Accenture’s study determined that companies investing heavily in AI and human-machine collaboration could boost growth by 38 per cent and increase employment by 10 per cent over the next four years. If speedy delivery times make customers happy, they’ll buy more. If warehouses operate more efficiently and are better organised to store more inventory, they’ll be able to meet that demand. If there is good engagement with key suppliers, there will be fewer weak links in the chain. People partnering with machines to run advanced analytics make that happen. Now that’s a formula for a winning team.


The five levels of analytics

  1. Descriptive: Making sense out of the numbers by comparing them to previous numbers, as in year-over-year sales; answers the question, “What happened?”
  2. Diagnostic: Drills down into the numbers gathered; answers the question, “Why did it happen?”
  3. Predictive: Uses the diagnosis to build a model accounting for a number of different scenarios; answers the question, “What will happen next?”
  4. Prescriptive: Advises what course of action to take for each possible scenario; answers the question, “What should we do next?”
  5. Cognitive: Uses AI and machine learning to evaluate the execution of the previous steps and optimise processes; answers the question, “How did that go, and what can we improve?”