IR Guest Comments

GUEST COMMENT Retail’s leap into Prescriptive Analytics

prescriptive analytics

Channie Mize, senior vice president & global general manager for retail at Periscope By McKinsey, gives her view on why prescriptive analytics is the new way for retailers to make insight-driven decisions.

Retailers have always worked hard to understand their businesses through data and to make better decisions for the categories they manage. But a new chapter is opening: prescriptive analytics is transforming how retailers raise profits and engender customer loyalty. But what is new about prescriptive analytics and what makes it better than other approaches?

Traditionally, at the start of the big data era, retailers employed data scientists who used descriptive analytics to understand the causes of previous successes and failures to help improve customer loyalty, pricing, promotions and assortment. Today, category managers at retailers of all types, from DIY to groceries and high fashion, use predictive analytics to get directionally accurate forecasts for a handful of scenarios. But since both these analyses focus on past results and don’t provide proactive recommendations, they hinder making the rapid changes required in every store to maximise revenues and profits.

This is where we get to Prescriptive Analytics – which is the latest and most advanced technology to accelerate category managers’ success. It differs in that it uses machine learning, pattern recognition and models to proactively suggest next steps or actions based on analysis of complex criteria and data. The objective of prescriptive analytics is not only to predict future outcomes but to make recommendations based on those outcomes. It is reliant on new transactional and non-transactional data tracked via the consumers’ digital footprints across multiple platforms.

Machines can now do more than crunch mountains of data – they can also recognise patterns humans can’t, learn from their mistakes and make specific, real-time recommendations that are easy to understand. Using the new approach, retailers can tap into in-house and external data to identify the handful of products with the biggest impacts on basket size and profit and then make simple weekly or even daily recommendations to category managers to adjust pricing, promotions and assortment in each brick-and-mortar and online store to boost revenues, profits and customer loyalty.

Prescriptive analytics takes the heavy lifting off the category manager and allows the science/machine to do more of the work enabling greater operational capacity. This means that as a category manager, rather than fighting the data, which many have done, she can focus on the job of making the decisions that improve her category performance.

Prescriptive analytics is not just for big retailers either, it is helping level the playing field for all retailers, but especially bricks-and- mortar stores because the tools make timely, tailored, granular sets of recommendations that category managers can accept or reject based on their real-world experience and business knowledge and the tool’s decision support.

For the customer, prescriptive analytics helps simplify their shopping experience by pre-empting their needs and suggesting products that they want before they are yet aware of wanting them – thus motivating even the most reluctant of shoppers.

Harnessing prescriptive analytics takes time. To get the full value of the new tools, workflows and methods, most retailers need to change the way they think and work. In our experience, widespread adoption also requires testing, patience and the right champions if it is to be a success.

The best category managers are experts in the products they sell. Most also have strong relationships with vendors and believe they know shoppers well. Naturally, they tend to resist changing the way they make decisions based on the advice of a data scientist in headquarters who’s never created an assortment or negotiated with a vendor.

So how can a company persuade the organization to incorporate prescriptive analytics into their decision-making process – especially when recommendations go against their gut instincts? In our experience, most companies need to start small and gather success stories.

For example, if a manager wanted to consider the impact of cutting the “long tail” of SKUs in a category, the prescriptive analytics algorithm might recommend keeping certain low-selling specialty products on the shelves because they attract some of the best shoppers, and that taking them off the shelves might cut profits in other parts of the store – something the manager might not be able to anticipate on her own.

In our experience, when managers accept recommendations that go against their gut feeling and quickly see clear improvements in metrics such as basket size and same-store sales, they become the champions of prescriptive analytics tools that an organisation needs to accelerate the transformation – they tend to have more influence over their peers than data scientists.

Making hundreds of small, steady improvements every week won’t require armies of analysts – the prescriptive systems are scalable because the rocket science is built in. The pioneers will grow their bottom lines and gain subtle but powerful competitive advantages, building customer loyalty and changing the nature of bricks-and-mortar retailing.

Prescriptive analytics is much more than software and data science – it’s a way to combine the strengths of humans and machines to make better, faster decisions. Using it to make significant performance improvements, therefore, requires organisational and mindset changes. But it does not need to be an overnight change, planned incremental steps in individual categories can help build confidence in the technology and lower the risks.  For adventurous retailers prepared to try something new, there is a real opportunity to impact the bottom line and “wow” customers.

Photo credit: leowolfert (Fotolia)




Leave a Reply

Your email address will not be published. Required fields are marked *