Please enable JavaScript.
Coggle requires JavaScript to display documents.
How to identify areas of accuracy improvement for demand forecasting ? -…
How to identify areas of accuracy improvement for demand forecasting ?
By better capturing inherent attributes of demand
Is the trend of SKU level demand fluctuating significantly or stationary?
Do SKUs' demand exhibit high seasonality or cyclicity ?
Are the forecasts inaccurate consistently vs in specific weeks across SKUs ?
Has there been a baseline shift in demand curves of SKU demand ?
What level of seasonality does SKU level demand exhibit ?
By capturing outside influencers of demand
Do external events influence demand ?
Do certain SKUs observe spikes/dips and hence low accuracy during holidays ?
Did certain SKUs experience spikes/dips/base line shifts during covid and hence experiencing low accuracy post it ?
Does SKU demand and hence accuracy vary due to competitior events and promotions ?
Does Alamarai driven activities influence demand ?
Is there high influence of marketing and merchandizing on SKU demand ?
Does inventory and supply of SKUs impact demand significantly ?
Do promotions of products influence demand of SKUs ?
Have prices for SKUs changed over time with corresponding changes in demand and hence accuracy drops ?
By improving the forecasting process
Are other algorithms giving better SKU demand accuracy ?
Are the current models always under / over forecasting for specific SKUs ?
Are supervisor adjustment factors improving accuracy for certain SKUs?
Are supervisor forecasts (adjustment factor) always under / over estimated for specific SKUs ?
Is the process delivering low accruracy consistently vs in specific weeks ?
In weeks of low accuracy, is it because of actuals fluctuating significantly vs model not being able to capture minor variance ?
In accuracy low across all SKUs vs speciic SKUs ?
In weeks of low accuracy, is the trend / baseline shift, seasonality , impact of other variables captured accurately by the model ?
Which external influencers change significantly in weeks with low demand forecast accuracy ?