Data-driven supply chains: A retail perspective

24 April 2018
From our 'Thinking outside the box' series
Supply chain management is undergoing rapid technological change. Gareth Jude and Behnam Fahimnia ask what the ability to collect, store, analyse and collaborate with data will mean for managing retail supply chains.

Supply chains have existed since the earliest days of human civilisation, but the idea of supply chain management is a relatively recent concept. At the core of supply chain management is the idea that more value is created by collaboration than by competition. One of the earliest examples of supply chain collaboration in action was between Walmart and Proctor and Gamble in the mid 1980s. Walmart agreed to share point of sale data with Proctor and Gamble and, in turn, Proctor and Gamble agreed to ensure optimal on-shelf inventory at Walmart. The results were increased sales, decreased inventory in the supply chain, lower costs for Walmart and better market intelligence for Proctor and Gamble.

Since then, supply chain collaboration has become a well-established practice in parts of the retail industry. Supply chain management is especially important in the retail industry as the Reserve Bank has calculated supply chain costs represent an average of 40 per cent of retail selling prices.

Supply chain management is about to be transformed by technology. Gartner predicts 5.5 billion smartphones and 20 billion Internet of Things (IoT) devices will be in market by 2020. It will be possible to collect data not just on sales and production schedules between retailer and manufacturer (as in the Walmart and Proctor and Gamble case), but also on the location and condition of goods throughout the supply chain, consumer and team member movements (both in and outside stores), stock levels in consumers’ fridges and appliances, the availability of consumers to receive a delivery, and much more.

Data storage is already being transformed through the increasing capacity and flexibility and decreasing cost of cloud computing. Blockchain and data exchanges, will mean information can be shared by all parties – not just point-to-point collaborators – easily and securely. Data analysis will also be transformed as autonomous or semi-autonomous platforms (AI) allow businesses to discover deeper insights, make predictions, or generate recommendations in virtual real time. In addition, high-speed data networks such as 5G in combination with dedicated low cost, low data rate networks like 4G Cat-M1 and Narrowband IoT (NB-IoT), will enable information to move between partners at a pace that allows timely business decisions.

What will this new capability to collect, store, analyse and collaborate with data mean for the management of retail supply chains? German online retailer Otto has introduced AI to predict demand and produce orders for 200,000 items per month. Overstocks are down by 20 per cent, returns are down by 2 million items a year and goods are being delivered more quickly leading to increased customer satisfaction.

Chinese convenience store retailer Bingobox has introduced IOT and AI to enable unmanned stores. Customers use their smartphone to gain entry and as they leave their goods are scanned via RFID and charged to their WeChat Pay or Ali Pay account. By removing checkouts and automating inventory management, Bingobox believe a team of four staff members can manage up to forty BingoBox stores. Scott Galloway has predicted that in the future, Amazon will send you two boxes a week: one containing everything they think you will need, and another to return what you didn’t use.

A recent survey shows Australian retail supply chain professionals are overwhelmingly convinced of the benefits of data driven supply chains and have begun to transform, but there are also obstacles. A reluctance to share key data, the need to properly secure it, skill gaps, challenges with current technology, organisational issues and relational issues relating to alignment of goals and processes with partners, were identified as barriers to deployment.

One thing that is easy to forget in an age that emphasises data analytics, is the significance of the human factor. There is a human element that links data to decision making. Better understanding and formulating this human element is a critical aspect of data driven supply chains. David Ferruci, IBM’s lead in the creation of the Watson computer system, views the future of decision making as a combination of human judgment and algorithms. The use of data is almost entirely dependent on human judgment in various forms.

The transition to data driven supply chain management will not be easy for the retail industry. Universities and researchers have a key role to play. Data driven supply chains require the traditional skills of supply chain professionals, but new skills from various disciplines will also be required. In particular, data scientists will be needed to interrogate the data and build the algorithms that turn supply chain data into useful business information.

Data science is predominantly the domain of university-qualified PhDs and research fellows. Online retailers like Amazon have long recognised the benefit of academic data scientists in their businesses. Target USA now employs 40 PhD data scientists in their IT team. Australian retail companies now need to fill this skills gap. Forming closer ties with Universities who have pipelines of PhD students and research fellows who have knowledge of the discipline and the supply chain sector would be a logical first step.

  • Gareth Jude is an Adjunct Lecturer at the University of Sydney Business School
  • Behnam Fahimnia is the Chair of Supply Chain Management at The University of Sydney Business School

Related articles

05 February 2024

How value adding is AI for strategic transport planning? Is AI Intelligent or simply a descriptive information dump?

Professor David Hensher reflects on the use of artificial intelligence (AI), particularly generative-AI (G-AI), in strategic transport planning, discussing its adaptability to diverse and unpredictable future scenarios, highlighting concerns about the limitations of G-AI in predicting situations with high divergence and emphasizing the need for utilizing hidden data not captured by AI.
08 January 2024

Autonomous Vehicles: Friend or Foe (or both)?

Abdullah Zareh Andaryana, Michael Bell, and Mohsen Ramezani provide an overview and critique of autonomous vehicles.
04 December 2023

Managing peak period rail travel: How fares should be constructed to spread commuter loads in the post-Covid working environment

Christopher Day looks at how the extension of the peak time period hasn't flattened the peak and suggests returning to the previous peak time period would have benefits in terms of reducing the transport network’s maximum peak utilisation and corresponding capacity requirement.