Return to news

INTERVIEW: Data-Driven Retail – how to make better decisions in times of uncertainty

28 April 2020

Michał Koziara, CEO, 3Soft S.A., talks about how advanced data analysis can help the retail industry enter a new business reality.

3Soft: How will the current situation connected with the COVID-19 pandemic affect the retail industry?

The epidemic has changed not only the face of the retail industry, but the way the entire world operates. It is difficult to say for sure what effects the pandemic and the anticipated crisis will have on the economy and, consequently, also on the retail sector. We are undoubtedly facing a completely new market situation where knowledge and experience to date may not be sufficient to find solutions to current challenges. The response to these new challenges may be supported by advanced data analysis.

According to my observations of the industry, companies which had opted for the Data-Driven Retail approach and started using advanced data analysis and business process automation even before the pandemic started, are coping much better with the current situation.

According to my observations of the industry, companies which had opted for the Data-Driven Retail approach and started using advanced data analysis and business process automation even before the pandemic started, are coping much better with the current situation. I also note that retailers who have so far been skeptical about using Artificial Intelligence and advanced data analysis in sales are starting to inquire about these solutions. They want to obtain tools to help minimize the effects of the pandemic and optimize processes in the new business reality.

 

 

3Soft: To what extent will the pandemic and the expected crisis change the retail industry?

Retail is characterized by high specificity in particular sectors. The current situation will have a completely different impact on the FMCG segment, and other challenges will have to be faced by fashion retailers. Nevertheless, in my opinion, there are three main areas of change.

Firstly, the importance of online channel sales will increase. The current situation means that many customers who previously were not convinced to shop online are starting to use this option. They have been forced to do so, in a way. So we can think about using the potential of the real omnichannel. Combining the online and offline worlds means not only tracking customer journey, but also effectively implementing operational tasks related to warehouse stocking or logistics.

Secondly, the current situation clearly shows the importance of employees in the retail sector. Ensuring their safety and facilitating the performance of their daily duties is what may be crucial for the retail sector to bounce back. Many retailers reiterate that in order to be able to meet the “customers first” requirement, they must focus on the implementation of the “employees first” strategy. Using Artificial Intelligence to optimize and automate business processes in retail will facilitate the work of people, including those working from home. Thus, employees will be able to focus on strategic activities, leaving the handling of routine operational tasks to systems.

Thirdly, the use of advanced data analysis will increase. The pandemic, or rather the effects it has caused, has shown that the Data-Driven Retail approach is the right direction. The concept, which used to be described as a megatrend before the pandemic, will now gain the interest of retailers and will be put into practice. Data is “virus resistant” and can provide valuable information on an ongoing basis. Artificial Intelligence-based data analysis will surely help retailers to find themselves in the new reality faster.

Data is ‘virus resistant’ and can provide valuable information on an ongoing basis. Artificial Intelligence-based data analysis will surely help retailers to find themselves in the new reality faster.

3Soft: How can Artificial Intelligence help retailers?

Artificial Intelligence is not only about futuristic autonomous shops, interactive shelves or bots as shop assistants. Although seemingly attractive, such solutions will be impossible for most retailers to implement, if only because of the need to bear significant costs for the infrastructure and store equipment. However, the use of Artificial Intelligence for data analysis is relatively cheap and business-efficient.

The implementation of the Data-Driven Retail approach involves analyzing billions of receipt data, information about thousands of products and dozens of retail-specific factors in order to predict future business events. Thanks to the use of advanced statistical methods on a large scale, it is highly probable that trends can be identified, demand for individual products can be predicted, marketing activities can be optimized and sales can be increased by adjusting the offer to individual customer needs “just in time”. Machine learning models are able to “learn” a new reality within two weeks, i.e. react to changes in an automatic way. In-depth statistical analysis supported by expert knowledge gives the business a new perspective on the current market situation.

Artificial Intelligence makes it possible to forecast sales individually for each store and each product, and to generate automatic orders to replenish stock in stores. In warehouses, it is possible to optimize the availability of goods and allocated capital on the basis of demand forecasts and to generate automatic orders for suppliers. The use of the latest Big Data technologies ensures effective monitoring of data flow and sales processes in real time and the generation of alerts in case of anomalies. All this can happen automatically, taking into account industry-specific factors and new variables that will be provided by the new retail reality.

We are talking about specific opportunities to build a competitive advantage by minimizing out-of-stocks and reducing over-stocks to increase sales and reduce losses. Artificial Intelligence makes it possible to forecast sales individually for each store and each product, and to generate automatic orders to replenish stock in stores. In warehouses, it is possible to optimize the availability of goods and allocated capital on the basis of demand forecasts and to generate automatic orders for suppliers. The use of the latest Big Data technologies ensures effective monitoring of data flow and sales processes in real time and the generation of alerts in case of anomalies. All this can happen automatically, taking into account industry-specific factors and new variables that will be provided by the new retail reality. If we add the automation of business processes to this, we can talk about a transformation towards Data-Driven Retail.

 

 

3Soft: What can such automation consist in?

Let’s take a chain of stores as an example. Automation becomes crucial when it is necessary to optimally stock dozens of points of sale with an assortment of thousands of items on a daily basis, based on the latest data. Most retailers engage their employees in the process of stocking their stores at an operational level. The pandemic has shown that optimization by increasing automation would be safer for both employees and the continuity of supply. Statistical models are trained daily on the basis of current sales data. As a result, sales forecasts are generated for each product and store individually. Based on the forecasts, orders are created to replenish stock in each store. The entire process, from collecting receipts from cash registers, through model teaching, to generating orders for warehousemen, runs automatically, limiting people’s involvement to an absolute minimum. Such a solution is safe and effective.

3Soft: Based on your experience, what advice would you give to retailers?

From the perspective of using advanced data analysis in retail, I recommend considering the implementation of solutions from this area even in a limited scope or in relation to a single business process. I am in favor of the strategy of picking low-hanging fruit, i.e. implementing those business cases first which can bring the greatest benefits at the lowest cost and at low risk. Contrary to appearances, as regards deriving business value from data collected by retailers, there are many such opportunities.

At 3Soft, we always start our cooperation with clients by holding technology and business workshops. During these meetings we analyze the situation, identify opportunities and discuss possible scenarios. Together with the client, we study the available data and consider processes which can be optimized by using Artificial Intelligence. We try to inspire and turn the inspiration into concrete actions. This is how we understand our role.