A Better Forecast???

Predicting what customers will buy and when, has always been a challenge to businesses. Our goal remains to have what the customers want when they want it, but forecasting demand still  creates challenges.  The software industry has responded with advanced planning modules, a multitude of forecasting algorithms, and even the introduction artificial intelligence into the forecasting process.  With all these high-powered tools and the time and money invested by businesses, why do we still struggle to realize a high level of forecast accuracy?

The challenges we face in trying to create an accurate forecast center around the product leadtime and Stock Keeping Unit (SKU) level detail.  Measuring the accuracy using total sales or at the product family level may show a better than average result, but when calculated at the level in which we sell (SKU), we find we’re not that good.  But this is the level that we must plan to in order to meet our customer’s needs.  This fact is compounded when there is long product replenishment leadtimes.  The results are a common mantra in business today, “We need a more accurate forecast.”

Instead of trying to build a better forecast, we should shift the focus to reducing the replenishment leadtime.  The shorter the leadtime, the more accurate the forecast.  If our leadtimes were one day, our accuracy would be high. Several factors contribute to overall leadtime including purchased component processing time, transit time, and manufacturing time.  Therefore, using different strategies to reduce the total replenishment leadtime may be more advantageous than trying to improve the forecast accuracy.

#orderreplenishment #leadtimemanagement #improvedperformance #planning #dataaccuracy #strategy #predictiveanalytics