On March 25, I listened in on a COVID-19 special interest group webinar which discussed strategic planning with the corporate board. After the Gartner analyst concluded her remarks, a question and answer session with supply chain executives followed. It wasn’t long before one of the more perplexing, but common challenges was raised. How do you solve for poor forecasting when COVID-19 has turned upside down any semblance of normalcy and when history is no longer a predictor of future demand?
The discussion centered on use of social media as a solution, but the conversation soon moved on to the next topic of discussion. In that moment, I thought to myself there was a missed opportunity for a deeper discussion.
Impacts From Demand Volatility
We do not have to look far to understand how companies have been impacted by demand spikes, less demand and changing demand variation. Demand volatility is certainly not a new challenge for supply chains, but its impacts have perhaps never been more greatly felt than during the coronavirus pandemic. Here are some recent headlines that point to the impacts:
- Last week, Nike called out China stores seeing double-digit increases in traffic week over week, with “some stores having already returned to prior-year levels”
- Church & Dwight seeing a surge in demand for its household goods
- H&M Hennes & Mauritz AB April 3 earnings call — “In the second half of the first quarter, growth was held back by the rapid outbreak of coronavirus, particularly in China where sales declined by 84% in February. Excluding China, Hong Kong, Singapore, Macau, Japan and Taiwan, sales increased by 7% in local currencies.”
- Walgreens Boots Alliance April 2 earnings call — “At present, we’re seeing a number of puts and takes. In the U.S., we saw very strong sales in the first three weeks of March as customers accelerated their purchases across multiple categories. However, we are now seeing declining sales trends, especially in quarantined areas. While it is difficult to predict future sales, realistically we expect to see future destocking by our customers and short-term sales may be negatively impacted by lower foot traffic. This is particularly evident in discretionary categories such as beauty and photo as customers redirect spending to everyday essential categories.”
Demand Sensing as a Solution
My colleague Marko Pukkila, in his recent Power of the Profession blog published March 30, suggested a new dawn for demand planning is upon us. I share his sentiment and at the heart of this “new dawn” is a demand-sensing capability. While known in some industries, it’s still not fully or widely used. The potential for demand sensing is still an untapped opportunity for most.
Gartner defines demand sensing as “the translation of demand information with minimal latency to detect who is buying the product, what attributes are selling and what impact demand-shaping programs are having.” There are common and less common data sources used with demand sensing (See Figure 1).
While demand sensing solution providers will tell you they can improve the accuracy of your demand plan utilizing all these inputs, the reality is most companies use only the left side (common data sources) in automating their forecasts. Relying only on common uses does not go far enough to understand the unprecedented impacts of COVID-19. In today’s war against the coronavirus, what also matters are the last two data inputs on the right side (social sentiment data and consumer/shopper/user insights). These inputs must come from geographies, local markets, stores or end users to get an early read on demand changes and understand the underlying causals of purchase intent.
Demand Sensing Benefits
- Daily statistical forecasts vs. weekly or monthly forecasts using daily point of sale and customer inventory data to understand trends sooner.
- Higher forecast accuracy. Improvements of 15-40% increase in forecast accuracy have been reported.
- Shorter forecasting lag periods (seven days vs. a traditional 30-day lag forecast) to faster understand sources of error.
- A prerequisite for an agile supply chain response.
The coronavirus pandemic is unprecedented in its supply chain impact. A demand-sensing investment when combined with an end-to-end supply chain capability can be a powerful tool to best combat black swan events such as COVID-19 (See Figure 2).
It’s not just a good demand signal that counts, but it helps. Combining an accurate forecast with the ability to make the demand plan visible across the organization and with suppliers is the first step. When you can deliver it with enough frequency (daily) to address changing market conditions you are then in position to support your supply chain reliability and agility initiatives. When joined with the right strategy and technology investment, only then can you expect to create a resilient end-to-end supply chain capability.
When the dust settles, and we move to recovery and lessons learned from COVID-19, you may have the best opportunity to improve your demand sensing capability. Use this opportunity to build the internal business case for a demand sensing investment.
- Build a roadmap of end-to-end improvements across your supply chain, demonstrating how an improved forecast enables critical supply chain functions that increase agility and resiliency during times of change.
- Identify and quantify the impact of an improved forecast to your organization using measures relevant to the business. For example, focus on those impacts that can improve service, reduce inventory and obsolescence, improve cycle times and reduce network costs.
- Use the Gartner benchmarking service to identify your forecasting performance compared to best-in-class organizations. Leverage benchmarking outcomes to reinforce the need for further investment given that differences between median forecast performance and best-in-class forecast accuracy performance are significant across most industries.
Managing Vice President,
Gartner Supply Chain Research