New Functionality for Higher Forecast Accuracy

When it comes to inventory forecasting, seasonal items are some of the toughest for a traditional ERP system to forecast- especially in an industry like HVAC. Buyers usually have to do a lot of manual work in spreadsheets for each of these seasonal items, looking at last year's sales, in order to come up with inventory replenishment quantities.

As a solution to this problem, Thrive Technologies has just released new functionality called Extreme Seasonality, which analyzes prior years’ sales to accurately forecast extremely seasonal items.

For example if we look at the item below, we can see the demand history in the yellow rows for the last 3 years, and the 12 month sales forecast in the light blue rows. We can see that this is a very seasonal winter item with strong sales from August to March each year. This is a common pattern for heater parts, for example.

Extreme Seasonality Solution

We can see how the Thrive system created an excellent forecast for the next 12 months. This is because the functionality automatically identifies the strong seasonality by analyzing the demand history patterns, and therefore predicts the seasonal pattern going forward as part of its weekly forecasting process. This forecast then becomes the basis for much better inventory replenishment calculations.

Many ERP systems will use a moving average to predict usage or sales. If a 3 month moving average was used here, the usage number for December would be (8525 + 5263 + 2804 ) / 3 or 5531. If your buyers used 5531 as the basis for buying decisions for December, you would obviously buy far too much inventory, since the actual sales will most likely be much closer to 500 units, not 5531. This new functionality therefore results in fewer lost sales during the winter season and reduced inventory during the rest of the year, thus increasing inventory turns and fill rates and increasing profitability.

Other Features

The seasonal profile is automatically calculated for seasonal items, but then can be modified by users. In the example above, even though there are small quantities of sales in April – July, the monthly indices have been set to zero because this company does not want to buy this item at all in the off season.
Seasonal profiles can be copied to other items
Extreme seasonality also calculates the trend from year to year and incorporates that into the 12 month forecast
Extreme seasonality weights the most recent demand higher than older demand in terms of monthly index calculations
Extreme seasonality is recalculated each weekend in case the demand pattern or trend changes

To learn more about how our software can improve your company, or to schedule a free software demo, visit us at