Forecasting for Retail and CPG in 2020

Use Cases & Projects Catie Grasso

While the backdrop of 2020’s global health crisis and economic uncertainty makes heading into the holiday season quite unlike years past, the U.S. is still slated to drive online sales growth. According to eMarketer, both Black Friday and Cyber Monday shopping days are positioned to surpass $10 billion in e-commerce sales, with their projected totals up 39% and 38% from last year, respectively.

How, then, can organizations operating in the retail and CPG sectors make the most of the unprecedented shift to online? One key answer lies in incorporating demand forecasting into a greater data and analytics strategy. The more accurately retailers can forecast demand, manage inventory and lead times, and improve relationships with suppliers, the better positioned they are to reduce costs and more effectively invest their resources where they will generate the most impact. We’ve rounded up five ways organizations can begin (or optimize their existing forecasting capabilities) to accomplish this:

1. Leave traditional forecasting and planning methods that are full of manual processes and, resultantly, unintended bias, in the past. Instead, leverage machine learning-based demand forecasting which is fully capable of incorporating the wide range of data sources needed to produce results precise enough for the modern enterprise and an ever-changing environment.

2. Incorporate and use said ML-based forecasts to allow for greater accuracy based on the predicted buying habits of net new customers. Further, they can be used to better understand the interplay between brick-and-mortar and online operations, something that will be front of mind for retailers globally this holiday season as they determine what to prioritize — physical or digital shelves.

3. Use real-time data to move inventory where it’s needed to avoid delays and, ultimately, unsatisfied customers. Retail analysts can use Dataiku’s use case-specific plugins for inventory demand forecasting to predict required stock and optimize SKU placement without the need to code.

4. Yet again, harness predictive analytics to decide what product to stock and where (and also let inventory management teams know how many units of product they have on hand for the next 30, 60, and 90 days, for example) based on data about regional differences in preferences, weather, and so on.

5. Use forecasting as a catalyst for other use cases. For example, using machine learning to forecast revenue can help define and allocate staffing for certain physical store locations and can simultaneously be used to monitor any fluctuations in inventory levels. Companies will be able to identify which factors are the most important for predicting revenue, such as the day of the week.

Although the 2020 holiday season will likely have less people rushing out to stand in lines for physical store experiences this Cyber Five (the five-day stretch between Thanksgiving and Cyber Monday), that doesn’t negate the attention retailers should place on the consumer. By leveraging data science and machine learning effectively, retail and CPG organizations can more effectively understand shifting consumer behaviors, market dynamics, and supply chain relationships.

You May Also Like

Demystifying Multimodal LLMs

Read More

At Saur, Water Leaks Are Stopped by AI

Read More

8 Core Steps for Your Predictive Maintenance Strategy

Read More

How to Address Churn With Predictive Analytics

Read More