Fourth’s sales forecast is modeled taking into account items, rather than sales amount. This prevents discrepancies with sales or discounts, as the sales and items are forecasted independently. Historical item data continues to be accurate.
The algorithm automatically determines the ratio of sales between this year and last year, applying a 60 day window (last 60 days this year vs equivalent 60 days last year). Therefore, if there is any upturn or reduction in sales this year vs last year as a result of anything Covid-19 or VAT change related, the trending ratio will reflect that and factor in to the forecast.
Any years beyond last year are only used to identify which days are closely related to others, which in theory should continue to be correct, post Covid-19 and VAT change.
As the Covid-19 impact is ongoing, the algorithm is automatically picking up that recent sales are more representative.