Analytics & Technology Boost FMCG Revenue by 35%
One of our clients was an FMCG company that sells aroma oils, potpourri, diffusers, candles and other such products in India, covering about 20 states in the country. They engaged us for a demand forecasting project for their number of product categories selling across 78 stores and about 500 store keeping units (SKUs).
Based on our assessment of their sales data, they wanted to prepare an inventory plan, and optimize some of their processes.
A forecast model was to be built by predicting weekly sales of SKUs, at city level. This is a requirement because the company has about INR 10 Crore of yearly turnover, and it is expanding, with a goal to reach INR 500 Crore over the next three to four years.
- Two years of store level sales data of SKUs
- Discount percentages as applicable to some SKUs at certain time in specific geographies
- Number of categories provided by client: 4
- SIBIA team created 12 categories based on product descriptions
- SIBIA team also prepared a holiday list, and seasonality index, to account for external drivers in the data modeling
Problem Solving Approach
A forecast model was built using LME (Linear Mixed Effect) model, to predict weekly sales for all categories at city level. There were a number of reasons for choosing the LME model, as mentioned below:
- Sales of a product varies as per specific factors unique to the concerned city, such as, local festivals, lifestyle of people, unique demand patterns
- LME allows treatment of the variations in data in phases and blocks, which helps to get a more accurate estimation
In order to build the forecast model, some of the external drivers were identified. These drivers were further used at appropriate instances in future analysis of data.
Moreover, there was a unique issue registered in the dataset – sales of some of the products discontinued after a certain period of regular sales. This called for additional scientific treatment of the dataset, to arrive at an accurate forecast model, as described below:
Various factors of sales growth were analyzed, for a time period of regular sales for SKUs that didn’t sell after the time period. In case of these SKUs, contribution weightage of the SKUs were calculated, so that the weights were distributed among rest of the selling SKUs during the time period (when the chosen SKUs stopped selling). These weights were multiplied with forecast sales numbers, to arrive at final forecast numbers.
Here is the final illustration of original vs. predicted (weekly) sales for aroma oils category:
In this manner, SIBIA helped the brand build intelligent inventory plan, so as to meet the demand of selected products at their target cities. This innovative beginning is a sign of growth mindset that sets up the momentum of the business in a designated direction of optimization and steady expansion. Enhancing analytical capabilities and use of technology helps a brand not only reach target users at scale, but also helps improve the customer experience over time. As a leading FMCG brand in India, the brand is certainly creating volumes of data; it is imperative they use it to maximize profits and yield powerful results. Here are a few recommendations:
Personalized Customer Experience
With India becoming the world’s fastest-growing market for mobile applications, in 2018, it’s high time that brands leverage this phenomenon to capture value by offering personalized services and offers, and creating value for customers in their lifestyle journey. For example, in Japan, customers who use McDonald’s app, spend an average of 35% more, because of the recommendations they are provided at the time they place an order. Mobile Apps can be used as an impactful channel to collect customer preference data, drive customer retention and growth based on novelty and loyalty, both.
Digital Products Menu
These aren’t just digital copies of existing product catalogues, but rather customized product recommendation lists delivered to target customers, over digital media, or App. Digital products menu changes with festivals, holidays, vacations, or weather. For example, Aroma Oils may be listed in the menu during a vacation that a customer has planned & added to the App., and upon ordering, be delivered to the customer at a desired destination. Special discounts on Diffusers and Potpourris may be listed during a festival season that people celebrate with family and friends.
Find out what happens when digital intelligence meets digital business
Top Five Data Science Tren...
The world of big data, machine learning and predictive analytics is actually one of the most vivid examples of a VUCA...
Did the Indian Consumers B...
GST (goods and services tax – the new value added tax regime in India introduced in July 2018) [http://www.cbic...
The role of predictive ana...
All eyes are on the Indian FMCG and retail sectors that “form the direct path to the India Growth Story - CII N...
Analytics & Technology Boo...
One of our clients was an FMCG company that sells aroma oils, potpourri, diffusers, candles and other such products i...