The recent and unexpected emergence of (and fallout from) the omicron variant has deeply impacted retailers and created even more supply chain hurdles, including labor shortages. This follows in the footsteps of panic buying and the historic out-of-stocks in 2020. We know that loyal shoppers leave when they consistently cannot find what they need when they need it.
Retailers are currently in an environment where it is no longer about getting supply chains “right,” but a continuation of the “feast or famine” environment we have been living in throughout the pandemic. Everyone is asking the same question: what’s next? What adjustments should be made when experts predict there are more disruptions ahead in the coming months and years?
One of the simplest ways to prevent out-of-stocks is to know when and where consumers will want products. In order for retailers to react to today’s supply chain problems, they need to employ impactful demand planning, which begins with high-quality data.
While the most applicable data sources required to stay in-stock vary by product segments, the characteristics of demand-sensing data are the same — timely, consistent, and regional.
Advertisement
“Timely” means that the data is up-to-date enough to make accurate predictions about what will occur in the near- and mid-term. Real-time data is the gold standard for demand planning. In addition to future plans, retailers can make real-time changes. With a clear picture of the ideal state they can work to shift inventory quickly and efficiently as soon as conditions change.
As an example: some retailers with illness-based demand rely on data from institutions like the Centers for Disease Control. The CDC publishes aggregated claims-based data from the prior week. On the mornings before publication (typically Fridays), the data available is approximately 12 days old. This lagging data can have serious consequences during times of rapid change.
In contrast, predictive illness insights informed by real-time illness data can reduce forecasting error because of its timeliness. Predictive illness insights can show where illnesses and their attendant symptoms will rise and fall.
In the case of Kinsa, these insights begin with collecting real-time illness data from millions of households across the country using Kinsa’s app-enabled smart thermometers. Kinsa collects temperature, symptom and demographic data completely anonymously and aggregates this data with other inputs of publicly available health data, weather data and more. The output is the creation of an illness Data Hub with predictive power to forecast when and where illness will rise up to 20 weeks in advance, while simultaneously monitoring in real time for symptom spikes at a hyperlocal level.
For retailers with illness-based demand in drug, mass, grocery and more, real-time illness insights can be leveraged for decisions around shifting inventory to distribution centers or stores. This ensures that customers find the products they need to feel better on the shelf at their time of need. While this example is illness-specific, the impact real-time data has on forecasting is universal.
Consistent data helps reduce errors and allows retailers to move swiftly and confidently. With consistent data, retailers can compare demand across time periods, helping them know whether demand is higher than similar periods in the past. Staying with our example of retailers with illness-based demand, this would include having knowledge of what a typical flu season looked like in years prior plus real-time data that shows where and how fast illness is spreading. This enables retailers to understand how exceptional illness levels are relative to previous time periods, which ultimately helps them make informed decisions about things like demand plans and staffing.
While topline data shows the whole picture, regional data is critical to capitalize on geographic nuances. For example, a store-level product forecast could be overlooked or simply “filled in” from a top down forecast with less defined datasets. One area of the country could be experiencing decreasing levels of the flu while other areas could be nearing their seasonal peak. Both require different inventory strategies.
CDC illness reports might show low levels of illness in a certain area, while Kinsa’s data might show that a distribution center in that same area is expected to see illness spread rapidly in the weeks ahead. If a retailer is relying solely on data like the CDC’s reports, they might miss an opportunity to make inventory adjustments and ensure enough products are on shelf to meet a community’s upcoming demand.
High-quality data is essential to planning. While no one has a crystal ball to predict the future, using the best data available is by far the best way to plan and then manage through highly variable demand.
Brad Pope is VP of Customer Success at Kinsa Health. He has over 20 years of experience in using retail data to solve supply chain and planning problems. Prior to joining Kinsa, Pope was the VP of Analytics and Data Science at Retail Solutions Inc., where he built and led two technical teams, guiding the delivery of solutions that customers used to make decisions about supply chain, ecommerce, shelf assortment, on-shelf availability, sales and category management. Pope has his MS of Data Science from Indiana University, instructed retail data analysis at Northwest Arkansas Community College for over five years, and is Scaled Agile Framework (SAFe) certified. At Kinsa, Pope ensures Kinsa’s Insights solutions and unique data signals are used to curb illness to the fullest extent possible.
