How will the current COVID-19 pandemic situation affect the retail industry?
The epidemic has changed the face not only of the retail industry, but the functioning of the entire world. It is difficult to judge what effects the pandemic and the predicted crisis will have on the economy, and on the retail industry. We are facing a completely new market situation where knowledge and previous experience may not be enough to find solutions to current challenges. The realization of these new challenges can be supported by advanced data analytics.
From my observations of the industry, it seems that the companies which started using advanced data analytics and business process automation even before the pandemic are coping much better with the current situation.
From my observations of the industry, it seems that the companies which started using advanced data analytics and business process automation even before the pandemic are coping much better with the current situation. I have also noticed that retailers who were previously skeptical about the use of artificial intelligence in forecasting demand and sales are starting to ask for such solutions. They want to acquire tools to help minimize the effects of the pandemic and optimize processes in the new business reality. Data is “virus-proof” and can constantly provide valuable information. Data analytics based on artificial intelligence will definitely help retailers find their way faster in the new reality and make decisions based on facts rather than intuition.
How can artificial intelligence help retailers?
Artificial intelligence is not just about futuristic and self-service stores, interactive shelves or bots as store assistants. Such solutions, although they seem attractive, will be difficult for most retailers to implement, if only because of having to bear significant expenses of store infrastructure and equipment. However, using artificial intelligence to analyze data is relatively inexpensive and business efficient. Every company has data whose potential can be tapped. We’re talking about actual opportunities to build competitive advantage by minimizing out-of-stocks and overstocks, thereby increasing sales and reducing losses. Machine learning models are able to “learn” the new reality in two weeks, i.e. respond to changes automatically. In-depth statistical analysis backed by expert knowledge gives the business new insights into the current market situation.
Our analytical platforms dedicated to the retail industry make it possible to forecast sales individually for each store and each product, and generate automatic store stocking orders. In warehouses, they optimize the availability of goods and allocated capital based on demand forecasts and 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 generates alerts in the event of anomalies.
Our analytical platforms dedicated to the retail industry make it possible to forecast sales individually for each store and each product, and generate automatic store stocking orders. In warehouses, they optimize the availability of goods and allocated capital based on demand forecasts and 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 generates alerts in the event of anomalies. All this happens automatically, taking into account factors specific to the retail industry and new variables provided by the new retail reality.
What might such automation consist in?
Let’s take a store chain as an example. Automation becomes crucial when it is necessary to optimally replenish dozens of outlets on a daily basis with an assortment of up to thousands of items, based on the latest data. Most retailers involve their employees in the process of replenishing stores at the operational level. The pandemic has shown that optimization through increased automation is safer both for employees and for maintaining continuity of supplies. Statistical models can be trained daily, based on actual sales data. The result are sales forecasts generated for each product and store individually. Based on the forecasts, picking lists are created for each store. The entire process, from collecting receipts from cash registers, through model learning, to generating orders for warehouse workers is automatic, and reduces human involvement to an absolute minimum. Such a solution is safe and efficient.
What specific results do 3Soft customers achieve by using data analytics platforms?
When it comes to the retail industry, the key parameters we support are sales growth through minimizing out-of-stocks and reducing losses through minimizing overstocks. We’re talking about sales increases of several percent, which means millions of dollars in profits for our global customers. Added to this is a reduction in average store shortages of several tens of percent. The effectiveness of using our forecasts is 95 to 98 percent on average, depending on the type of assortment.
Using your experience, what can you advise retailers in the current situation?
From the perspective of using advanced data analytics in retail, I recommend considering implementing solutions from this area even in a small extent or for a single business process. I am a proponent of the strategy of picking the low-hanging fruit, namely implementing first those business cases that can bring the greatest benefit with the least amount of effort and low risk. Contrary to appearances, in the context of deriving business value from data collected by retail, such opportunities are numerous. What’s more, the first results can be obtained very quickly, in as little as two or three months.
From the perspective of using advanced data analytics in retail, I recommend considering implementing solutions from this area even in a small extent or for a single business process. I am a proponent of the strategy of picking the low-hanging fruit, namely implementing first those business cases that can bring the greatest benefit with the least amount of effort and low risk. Contrary to appearances, in the context of deriving business value from data collected by retail, such opportunities are numerous. What’s more, the first results can be obtained very quickly, in as little as two or three months.
