The rise of e-commerce changes everything. But the tools and processes needed to forecast e-commerce demand reliably can increase your accuracy no matter how you do business. See how you can improve efficiencies across all channels and future-proof your business.
According to a recent COVID-19 survey of consumer sentiment, nearly three out of four consumers in America reported using e-commerce/contactless delivery services and switching brands during the pandemic in 2020. Additionally, retail market analysts predict e-commerce will continue to find favor (~10% year-over-year growth between 2021 and 2025), with e-commerce projected to account for ~17% of all global retail sales in 2021. If your enterprise planning competency begins to recognize the rapidly growing importance of e-commerce, you’ll be in a better position to satisfy customers and successfully support your business objectives.
Accurately forecasting demand must be a key element of your planning process. However, due to the lumpiness of demand, information asymmetry, and changing customer preferences, it can be – and most often is – very challenging for any organization to accurately predict customer demand. To meet this ongoing challenge, it’s essential that your planning organization continually revitalizes its human and digital capabilities.
Here are some ways to begin the journey towards forecasting excellence.
Build a great supply chain network
Evaluate your network. Determine if opportunities exist to improve your responsiveness by reducing lead times. Seek out and identify partners that can help you pool risk in a meaningful way. Leverage network consolidation to reduce your demand variability.
Identify the products and customers that matter most to your business
Identify the most important products and groups of customers based on strategic or monetary value (e.g., revenue, cost of goods sold, margin) and demand properties (e.g., demand frequency, volatility). You may also find additional criteria such as intelligence on product affinities, cannibalization, and product introductions quite useful.
Gather as much demand data as you possibly can
Understand the drivers of demand for different products and collect as much data as you can. Explore variables as diverse as weather, inventory levels, promotion/markdown events, customer reviews, product recalls, competitor pricing, product attributes, and macro/micro-economic indicators. A good big data governance process is an essential foundation to building a successful machine learning toolset.
Pay attention to your gross sales and returns data
Capture stock-outs (or even better, backlog) and website downtime events as they contribute to the under-representation of your actual customer demand. Also, pay close attention to e-commerce returns, which can amount to 10-15% of gross e-commerce revenues. Understand the drivers of product returns to plan your reverse logistics, and give importance to product availability.
Collaborate with your collaborators
In the true spirit of S&OP, work closely with sales, merchandising/marketing, finance, and the C-suite to align operational and business forecasts. Make Forecast Value Added (FVA) a part of your consensus discussions and carefully examine how different teams impact forecast accuracy.
Leverage e-commerce for traditional slow-movers
Use e-commerce to push those products that are traditionally difficult to forecast at the SKU-store location level. Leverage the benefits of channel consolidation to improve forecast accuracy and reduce your inventory costs.
Use demand sensing to dictate deployment strategies
Leverage consolidation points in your supply chain network to achieve better forecast accuracy. Set your deployment rules based on a combination of customer/channel priority rules and “sensed” short-term demand forecasts.
Augment forecasting capabilities with boxed solutions
Weigh the pros and cons of employing a boxed solution vs. a familiar homegrown tool to enable your enterprise forecasting processes. Focus on applications with robust analytical capability such as anomaly detection, time series models, newsvendor models, machine learning, demand collaboration modules, and reporting templates.
The growth of e-commerce, alongside your existing distribution channels, brings new levels of complexity to forecasting demand. Reach out to us and learn how Chainalytics can integrate the right processes and technology into a reliable forecasting methodology. Using one-of-a-kind tools and approaches like digital assets and managed analytics services, we consistently deliver actionable insights and measurable outcomes to our clients.
Suraj Vissa is a Sr. Consultant in Chainalytics’ Integrated Demand and Supply Planning practice. He is passionate about enabling supply chain planning teams and leveraging solutions at the intersection of business, process, and technological best practices.
