Online Grocery Retail: A Case Study

online grocery retail

The Background

Online grocery retailers experience a higher degree of demand fluctuations compared to traditional retailers. This is primarily due to the options provided on the online marketplace.

While demand unpredictability can lead to inventory-related concerns, it also severely impacts logistics planning in online retail.

Therefore, understanding demand, especially with respect to marketing, category management, and other operational parameters is critical for any online retail business.

Our Customer 

Our customer is one of the largest online grocery retailers in the world which does several millions of deliveries each year.


  • Ineffective logistics planning due to inability to forecast orders accurately on a daily basis
  • Trial and error’ approach to marketing and product management activities due to the lack of understanding of demand dynamics
  • Monthly and weekly forecasting not being adequate for daily operational planning
  • Low agility in the forecasting process not being able to support a dynamic business planning process

The Approach

The conventional approach of forecasting the final outcome (number of orders) fails to produce the kind of forecasting accuracies that are needed to address the above challenges. Moreover, it fails to capture the drivers of demand in a meaningful and actionable way which is necessary to optimize the marketing and category management activities.

Therefore, we invoked the stepwise forecasting function available in FORECAST² in which the final outcome is derived through a series of sequential forecasting models. We call the waterfall approach!

The customer journey

A typical ordering app takes the user through the following steps as part of the ordering process. The corresponding data is used in the forecasting process.

  • Open APP: The number of users who have opened the app
  • Add to cart: The number of users who have added items to the cart
  • Checkout: The number of users who have proceeded to check out
  • Delivery and payment: The number of users who have entered delivery and payment details
  • Order: The number of people who have placed an order

Accuracies: waterfall forecasting approach

93%   OPEN APP




Factors that influence each stage of the customer journey are incorporated into the corresponding forecasting model.

Accordingly, factors such as brand and awareness spend in both mass media as well as in digital media are part of the model that forecasts app usage.

Similarly, product discounting-related information is captured as part of the add-to-cart model. Influences such as time to delivery are captured under the delivery and payment model.

Finally, cashback, bank discounts, or any other order level discounts are captured as part of the order model

The outcome 

we achieved highly accurate demand forecasts across multiple cities
accuracies for each city
  • 2% reduction in delivery costs
  • Forecasting agility by reducing forecasting time to 30 minutes
  • Higher ROI on marketing and category management activities through optimal usage of demand levers and resources
  • Accurate foresight of the business outlook for the leadership

To know more about how your business can benefit from using Forecasting at Scale, please reach out to us or simply sign up for a free trial.

Sri Lanka
No. 63/1,
Dhrmapala Mawatha,
Sri Lanka.
No. 9, Raffles Place,
Republic Plaza,
Singapore, 048619.
Menara Digitaraya,
Jl. Raden Saleh Raya,
No. 46A, Jakarta,