Past Sales Adjustments
in progress
F
Fabricio Miranda
Adjust past sales of each product to correct it for anomalies such as stock-outs, price variations, ad spend and other outliers. Main objective is to feed the forecasting models with sanitized data, in order to increase the accuracy of the forecasts. Flieber will automatically pre-process some of the data and will surface other data for user to manually identify/tag. User will have the ability to make corrections to the pre-processed data in case the algorithms make wrong assumptions.
F
Fabricio Miranda
Experience:
- User finishes onboarding and is sent to the "Default forecast" modal, where they will see an option to enable/disable AI-based forecast adjustments (the default will be "enabled").
- User enters the "Edit forecast" modal and clicks on the "Past sales adjustments" tab -- Flieber will be showing on the chart the actual sales and actual forecast in grey, and the adjusted sales and adjusted forecast in yellow.
- The table below the chart will be transformed into an in-app spreadsheet (to allow quick editing), and will show 6 rows: actual sales, AI-adjusted sales, manually-adjusted sales, final sales, description (dropdown with options such as "stockout", "promotion", "ad campaign" etc.), and comments (free field for adding anything the user wants).
- Each column of the table will show one day of the sales period selected in the chart -- when a user selects a period, the chart will automatically show only the columns for the days on that period.
- When the user makes a change to the "Manually-adjusted sales" row, it will overwrite the "AI-adjusted sales"
- When the user clicks on the "Preview" button, the yellow line of the chart will change accordingly, including the forecast part, since a change in the past changes the future (the forecast will be re-run to account for those changes).
- The user will be able to click the "Interval" button to see the data in daily or monthly intervals -- if changed from daily to monthly, for example, the changes made daily will be consolidated into monthly and, if the user makes changes monthly and then selects to see daily, these changes will be distributed into the days of that monthly period.
- When the user makes changes in one tab and then moves to another tab of the "Edit forecast" modal, these changes will persist, so that nothing is lost.
- All sales-related data on the main "Sales forecast" page will show the adjusted sales (and not the actual sales anymore).
F
Fabricio Miranda
First release scheduled for February 2024. User will be able to see periods in the past sales chart where our algorithms have identified any anomalies and will be offered an option to tag those anomalies (examples of tags: stockout, influencer campaign, positive review, negative review, new version launched, among others). Tags will be used by the algorithms to improve the quality of the forecasts -- the more periods tagged, the more accurate will be the forecasts.
F
Fabricio Miranda
in progress