When you upload your sales order history data, Crisp assess the quality of the data provided for each product. Three important categories are assessed: history, density, and sales frequency.
History refers to the historical length of time that the data covers. Density refers to the number of data points provided. Sales frequency will only apply to products that are sold periodically or intermittently.
Crisp then generates feedback for you based on the data that you provided. For example, products that have a longer historical length and more dense data will have a higher quality ranking than products with a shorter history or sparse data points.
This feedback has two purposes. First, it is meant to give your forecasts some context. Forecast accuracy relies partially on the data that you provide, so the more data Crisp has, the better the forecasts will be. It also enables you to identify which products you may need to upload more data for.
Follow the steps below to view the data quality feedback for a specific product.
On the Products page, click the product ID that you want to view the data quality feedback for
Click the Data Quality button in the top right above the Sales / Forecast plot to review the feedback. Products with quality concerns will display a red notification in the top right of the button
- On click, Crisp will display a description of the data quality concerns. Descriptions will vary across products. Below are some examples of data quality feedback.
The product you have selected has fewer than 2 data points uploaded at this time. Crisp will not be able to generate a forecast for this product.
This product has a short history of data uploaded. Crisp will not be able to gauge the impact of seasonality and holidays until at least a year's worth of data is uploaded.
This product has sparse data. Upload more data, or relate this product to another product (if applicable), to improve Crisp's forecast accuracy for this product!
This product has somewhat sparse data. It is not ideal, but it is not as negatively impactful as truly sparse data either.
Crisp has flagged this product as being sold only periodically. This is not necessarily negative, but if you know this product is sold consistently you will want to upload more sales order history data for this product.
This product has not been selected for forecasting.
- You can address most data quality concerns by uploading more sales order history data for each product.