Accelerating Data-driven Value for Retailers
A Unified View of Demand Planning in Chaotic Times
by Paul Barrett
Reflecting on conversations with customers over the last six months, it's clear that the global pandemic is not in our rear view mirror - it's baked into our data. In fact, data is transforming retail. But across the retail industry, many retailers are experiencing "interesting times" when it comes to demand planning. The pandemic has disrupted historical data planning and is challenging the expectations of senior executives as global trends start and stop abruptly.
These challenges can easily be broken down into two key areas. First, historical data has become incredibly complex. The smooth curve and predictability gleaned from historical data was obliterated over the last 3 years with the pandemic driving demand sky high and other SKUs to new lows. Additionally, the recovery back to normal has seen unexpected spikes and lulls in normality. Secondly, this disruption further segregated the perspectives of executives, data scientists and planners. Planners are using demand data, executives are using their gut and data scientists are making decision with analytical data sets. We are now farther from a unified view of demand planning than ever.
At Accenture, we believe that for effective demand planning, it is critical that everyone within the company is using the same data, and working from the same information. When a spike in demand emerges, think back to toilet paper during early days of the pandemic, or a market issue like inflation changing, interest rates rising, companies need tools to help them identify the demand spikes and address them with pricing changes or evolving distribution plans. Executives want products on the shelves, data scientists want the most accurate data to make better decisions, and planners want to make sure the distributors are happy. If companies are going to extract the value that they need to stay competitive, they need to focus on driving not just data quality but a new level of veracity that elevates the signals and reduces the noise.
Shipping data from twenty-four hours ago may no longer be relevant if a port is shut down by the pandemic. Last year's seasonal patterns become less reliable in disrupted markets. The struggles around data quality are amplified for many retailers in their supply chain predictions, which by nature have the longest arc from signal to sale since orders may be placed months in advance. The siloed data, disruptions, and growing competition, while trying to keep an eye on overall sustainability, is an omnipresent obstacle that retailers are looking to overcome. With a process so complex, having a common set of facts and purpose built views for the key roles so that decision making can be fast, accurate and flexible.
What it takes to become a data-driven retailer
Retailers who will thrive in this ever-changing world will need to adopt a platform and strategy that provides total visibility, creating one version of the truth for their data. The solution will leverage internal data (e.g., from supply chain and trade), external data (e.g., consumption data, mobility, macroeconomic factors, brand sentiments, weather, COVID-19 cases) and advanced algorithms to forecast consumption. Retailers will be able to use this as their one point of truth – providing data that planners, data scientists and executives are looking for.
Over the last year, Accenture has been working with their customers to solve for this. Based on the recent launch of the Databricks Lakehouse for Retail, Accenture has created a Unified View of Demand solution that is an open, glass-box approach to demand planning. The Accenture solution is part of the Brickbuilder Industry Solutions program. With this solution, clients can increase forecast accuracy, speed and granularity, while shifting from debating forecasts to input alignment.
The business value retailers realize from data and AI
Accenture worked with one global CPG company facing multiple challenges for demand forecasting, based on consumption and shipment forecasting models. The customer was able to implement the Databricks Lakehouse Platform, solving for unconstrained demand and shipment forecasts. The solution also helped the customer standardize their way of receiving base price, average discounts, integrating the streamlined forecast to supply chain and finance. After Accenture set up Automated Demand Sensing utilizing the Databricks Lakehouse Platform, the client was able to generate weekly forecasts with 10-15% accuracy improvement. This bias improvement of 5-10% helped the supply planning team with better inventory planning and production schedules.
As companies look to prepare their data and AI strategy for the future, they need to keep in mind that retail is on a transformation journey, starting with their data.
To learn more about Accenture Brickbuilder Solution, Unified View of Demand.
To learn more about Databricks' Lakehouse for Retail, click here.
Frequently asked questions
Why has demand planning become so challenging for retailers since the pandemic?
The pandemic disrupted historical demand data by pushing some products to record highs while pushing others to new lows, destroying the smooth, predictable curves retailers once relied on. Recovery has been uneven, bringing further unexpected spikes and lulls that make last year's seasonal patterns less reliable. This volatility has also widened the gap between how executives, data scientists and planners each interpret demand, since they tend to rely on different inputs — gut instinct, analytical data sets or demand data, respectively.
What is Accenture's Unified View of Demand solution?
Unified View of Demand is a Brickbuilder Industry Solution that Accenture built on top of Databricks Lakehouse for Retail to give retailers an open, glass-box approach to demand planning. It combines internal data, such as supply chain and trade data, with external data like consumption data, mobility, macroeconomic factors, brand sentiment, weather and COVID-19 cases, and applies advanced algorithms to forecast consumption. The aim is to give planners, data scientists and executives a single, shared source of truth so decisions can be made faster, more accurately and with more flexibility.
How does a unified data platform help align executives, planners and data scientists?
A unified platform gives everyone in the organization access to the same data and information, replacing the siloed, role-specific views that widen disagreements about demand. Executives want products on shelves, data scientists want the most accurate data for decisions, and planners want distributors kept happy, and having one shared version of the truth lets each function work from consistent numbers. This shifts conversations away from debating whose forecast is correct and toward aligning on shared inputs.
What measurable results did the global CPG company achieve using the solution?
After Accenture set up Automated Demand Sensing on the Databricks Lakehouse Platform, the CPG company generated weekly forecasts with a 10-15% accuracy improvement and a 5-10% bias improvement. These gains gave the supply planning team better inventory planning and production scheduling. The implementation also helped the client standardize how it received base price and average discount data and integrate a streamlined forecast into supply chain and finance processes.
Why is supply chain forecasting especially affected by data quality issues?
Supply chain predictions have the longest arc from signal to sale, since orders are often placed months in advance, making them especially vulnerable to stale or unreliable data. For example, shipping data from twenty-four hours earlier may already be outdated if a port shuts down, and last year's seasonal patterns become less trustworthy in disrupted markets. Retailers need to focus on data veracity — elevating true signals and reducing noise — rather than data quality alone, to keep these longer-horizon forecasts accurate.
Get the latest posts in your inbox
Subscribe to our blog and get the latest posts delivered to your inbox.