Email
sinha.rk.dr@gmail.com
Is your Demand Sensing derailed by intermittent missing data?
I have seen many supply chain teams struggle with missing values in their Demand Sensing input data. Sometimes, demand signal has missing values in its data stream. Other times, it is for a demand driver.
Such missing values pose a challenge to the Demand Sensing algorithm. It is normally addressed by replacing such missing values by an internal estimate, which could be programmed as mean, median or a certain percentile of last so many days’ value. However, it still remains an estimate and not as good as the actual missing value.
While it’s difficult to go back in history and collect the missing values, can we do something to at least minimise its occurrence? That’s where AI agents play an important role.
AI agents can track the occurrence of missing values and decipher if there is a pattern to these occurrences. For example, a particular geography may not be capturing the demand signal properly. Or a particular demand driver like a specific promotion may have gaps in its data capture. Once we know the specifics, the information gap is easier to plug. The agent may even take action to repair the faulty data pipeline.
Improving the quality of input data through AI agents is a key step in improving the efficacy of Demand Sensing.