December, 20208 IN MY V EWBY MICHAEL FAGAN, CHIEF TECHNOLOGY OFFICER, KMART AUSTRALIADRIVING COMPETITIVE ADVANTAGE FOR KMART GROUP THROUGH ADVANCED ANALYTICS, MACHINE LEARNING AND AIWhen Adapt surveyed more than 100 Australian CIOs in 2019 there were some interesting, if not surprising, results. More than 50 per cent believed they had too much data making it unwieldy and difficult to analyse, nearly 60 per cent struggled to hire good data scientists, and 70 per cent felt that they were behind global competitors as they struggled to generate value from their data and analytics efforts. Yet despite these difficulties, 95 per cent of Australian CIOs say Analytics is an important part of their strategy, and 54 per cent say it is one of their top 5 priority.The start of 2020 has been a tough one in our industry, but it is clear that the retailers who will flourish are those who can leverage analytics to deliver delightful customer experiences and optimise their retail operations. Innovating your offer is no longer optional ­ those who fail to improve will be overtaken by digital disruptors, many of whom are not based in Australia. This latter point is another feature of our digital age ­ problems that used to be solved on a national basis are now being solved globally, often by software-driven businesses. Uber revolutionized the taxi industry and took it global. Our customers are Uber's customers, and they expect the same personalized service. So, we must improve, and do it fast, for we are competing with the world's best ­ and advanced analytics is central to our competitive arsenal.The Kmart Group Data &Advanced Analytics team is committed to drive competitive advantage for the group by shaping and delivering better data-led decisions. We work in partnership with all business functions, including Merchandise, Online and Supply Chain. While this presents several opportunities, the challenges relate to identifying high impact use-cases, prioritization of work across verticals, and translation of business requirements. We don't have difficulty finding work to do, rather we have too much work to choose from and so must focus wisely In the current financial year, we have focused on customer outcomes related to demand forecasting and price optimisation. Our business stakeholders had a problem: they want to get the right stock to the right stores at the right time, so our customers can purchase what they want, when they want it. This is compounded by demographics, size curves and preferential tastes ­ twenty men and women will purchase the same drinking glass, but wear different shape t-shirts, in different sizes, and like different colours. Figure 1 - Semantic analysis of more than 4,000 social media comments received in one day last December showed customers talking about the new beach trolley that was recently launchedOur data scientists built a machine learning model which we fed with data from all products in all stores historically, and trained that model to recognize the attributes that are important to purchase decisions. After tuning and tweaking the model, we now have a powerful demand forecasting tool that is run on our own ML algorithm to accurately forecast full-price monthly demand for new and existing products during
< Page 7 | Page 9 >