Coming to terms with the new analytics-driven retail environment

The retail environment is a challenging one, even at the best of times. Before products reach the shelf, the retailer must plan the ideal assortment mix, orders and number of units. Merchandise teams do their best to predict what the hot new products will be, which items customers prefer to shop for online and, most importantly, dial in the ideal price to be competitive. Of course, this merchandise planning and all other aspects of retail have been disrupted by events of the past two years.

From stock-keeping unit rationalisation, assortment optimisation, store clustering, category management and demand forecasting, to gaining a better understanding of customers, the entire supply chain has required a rethink. To orchestrate decision-making at the pace and scale dictated by the huge amounts of data being generated, retailers have become more reliant on high performance in-memory applications powered by analytics to get the job done.

Thinking differently

The pandemic has had a huge impact on the supply chain with retailers, consumer good firms and manufacturers having to think differently about how best to use analytics to enhance their value proposition for these new operating conditions.

Today, there are many touchpoints that must be considered when making crucial merchandising decisions, and analytics are needed to drive this. Historically, retailers looked at store sales and e-commerce sales separately. Using advanced data management capabilities along with analytics allows retailers to have a complete view of demand, regardless of channel, for a local market. This will enable retailers to understand what products resonate within a market, understand areas of opportunity, and then slice it all the way down to knowing the right sizes and prices for a given market.

Looking beyond historical patterns, anticipatory analytics will allow planners to predict demand, especially when there is no precedent. Thanks to rapidly evolving technology, this begins with intelligent demand planning which uses machine learning that delivers insights driven by artificial intelligence (AI) for real-time decision-making.

Customer first

Over the past 18 months, retailers have realised the need to forecast from consumer to supply, not from supply to the consumer. Companies have forgotten that without demand there is no need for supply.

Demand planners have no real view into demand, and as such, are not able to react to shifting consumer demand patterns.

So, when it comes to demand planning for retail, the ideal future state is where predictive, AI-driven demand planning is used. This empowers retailers to deal with whatever happens next, including the unknown. At the optimal future state, demand planners within the retail space bring more value to the business through automated, accurate and fast forecasting activity. They can process improvements and collaboration across the supply chain thanks to having access to more advanced analytical capabilities.

Main stages

This requires retailers to develop a road map to reach this level of analysis. It is through this demand planning maturity model that retailers can create an enabling environment to be more driven by data and analytics using the sophisticated technology available to them.

The four stages to achieve improved demand planning are as follows:

Stage 1: Trusting your gut

In this stage, demand is estimated based on experience and what has typically happened in the past. However, future events cannot be accurately modelled with planners relying on best guesses and judgment.

Stage 2: Past learning

During stage 2, demand planning is more formalised. It is based on existing shipment transactions typically captured in an enterprise resource planning system. Unfortunately, this means that planning is supply-centric and not consumer-centric, making it difficult to cope with unpredictable events such as the pandemic.

Stage 3: Beginning automation

Here things are starting to change. The planning process is more efficient and sees much of the analytical process being automated. Forecasting becomes more flexible and there is greater accuracy due to the incorporation of predictive planning models.

Stage 4: It is a matter of science

This is the ideal future state any business in the supply chain should be striving towards. Here, all types of planning are generated through an integrated platform. Cloud-native demand planning solutions support continual optimisation and multi-team collaboration. More importantly, data scientists work alongside demand planners to assist them in extracting more insight, resulting in more advanced decision-making than was previously possible.

By embracing an environment where the retailer can use forecasting models that include underlying trends, seasonality, promotions and so on, the business can be better positioned for growth heading into 2022 and beyond.

For more insights about the four stages, download SAS’s free e-book, and click here for more on transitioning to analytics-driven demand planning. 

This article was paid for by SAS.