Zunair Ameer

I find it beneficial to first group SKUs in categories based on their demand patterns and then see which approach works best. It is also important to categorize based on markets and soft input from sales to further hone the model used. Then the most important bit is to review and renew on a regular interval through tracking KPIs like MAPE or even WMAPE if possible.

In fast-paced goods, this is true, you need short-term forecasting to take a leading role in your operations and long-term to be more of a strategic importance. In certain other scenarios eg. selling automotive wear parts, this may not work. There are many parts which need replacement after 2, 3, or 5, years of working. Hence, it is also worthwhile to consider which SKU or market you are focussing on before considering short over long-term and vice versa.

Reliability is subjective to the market and potency of our data and models used. I've seen seasonality spanning years and months alike. So I would say, short-term forecasting is just a tool which has its benefits in certain situations. If we have a smooth demand curve, this approach would be useful but If we have erratic, lumpy, or intermittent demand, then we need a more robust method in our arsenal.

Every Business and market need is different and thus requires a more boutique approach to get the best results. For example, in a variable demand with a short, stable lead time environment, one can get away with less MOQ and lower safety stock. Contrarily if market variability is lower one can always opt for a higher MOQ to reduce transportation costs. Turnover is important to manage based on the market. We may see some SKUs sell once in three years but they can be important for customer retention. In my experience, what works best is a holistic approach to consider service levels and define them based on demand variability, customer focus, and lead time deviation.

With growing supply chains and increasing SKU counts per inventory, it is increasingly hard not to recommend software. Companies over a threshold of SKU can find it very hard to effectively manage without the help of inventory management software. It also reduces the cost since you need very few individuals to manage all the inventory. I remember a decade ago, it would have taken me days to analyze local inventory using Excel while now global inventory requires only opening a tab with Power BI connected to our inventory management software. Hence, the benefits are undeniable and vast. In terms of what kind of software, it is better to have cloud-based when you consider global supply chains as we can manage global inventory from one spot

barcode scanners and RFID readers are helpful but I believe it is better to start with education. It is often said tools are only as good as the user. I experienced in a summer job that the warehouse had all these space age methods of inventory management but the colleagues working with the tools were not fully educated and it ended up creating more delay than efficiency. When we are using barcodes or RFID, we need to be accurate with our data. For example, if the worker managing a small batch isn't careful in stacking or shelving, we would end up with the wrong numbers and since we are using computers we likely won't realize the mistake for a long time.