The Demand Planning module starts with your historical shipments. These come from your ERP (NetSuite, SAP, Business Central, etc.) and are ingested at the customer, item, and location level on a weekly basis. Historical data (typically 2-3 years) is loaded during onboarding.
Note: The statistical baseline is generally based on the shipment history. Thus, the more historical data available to baseline the statistics, the better. Confido recommends 2 years, ideally at weekly granularity, of historical data per customer/product combination. |
3.1 Shipment History Ingestion
Confido connects directly to your ERP to pull weekly shipment actuals by customer, item, and ship-from location. If you are mid-ERP transition or have imperfect historical data, Confido can supplement with a CSV upload of historical shipment files. The onboarding team will work with you to align on the best approach for your data situation.
Note: If your historical shipment data has data quality issues — e.g., missing ship dates, incorrect item codes, or gaps from a system transition — Confido can work around this during the data load. It is better to load imperfect history than to start with no history at all, as even partial data helps the statistical models. |
3.2 Automatic Baselining and Outlier Detection
Once shipment history is loaded, Confido runs an automatic baselining process. This strips out large deviations — promotional spikes, pipe fills for new distribution, supply constraints, or other one-off events — to give you a clean baseline that represents your steady-state shipment rate.
Confido is uniquely positioned to do this accurately because it has access to both your shipment history and your sales intelligence (POS data and promotional calendar). When it sees a large spike in shipments and finds a promotion running in the corresponding POS data the following week, it can automatically tag that spike as promotional rather than treating it as a genuine baseline shift.
Deviations can also be tagged manually – or given a dedicated baseline number – based on a defined list of reason codes, if the automatic detection misses something or if the cause was internal (e.g., a supply constraint that caused a delayed order). Each tagged deviation can be labeled with a business reason, so the history of what happened and why is preserved. You can find the outlier section within the demand planning workspace. The default is set to the current customer/product combination; to toggle to all outliers, select “All grouping”.
Access the outlier wizard 
| Clicking on the magic wand auto-classifies outliers or mass-dismisses them   |
Manual overwrite with dedicated baseline value  | Manual overwrite with reason code
 |
Tip: Taking time to review and clean up the baseline during onboarding pays dividends throughout the year. A well-baselined shipment history means statistical models run on it will be more accurate from day one, and you will have fewer recommendations to clean up as new data arrives. |
3.3 Item Proxy for New Products
In addition to the sales module, new items can also be added to customers directly in the demand planning workspace. Click the “3 dots” next to a customer and select “Add innovation”. For the innovation item to appear, make sure it is in the product master (either through a direct ERP pull or manual addition).
New innovations can be launched manually into DP. Select manual entry and decide on the default weekly cases per DC, launch date and distribution centers to launch in. If you want to launch at all DCs hit “select all”.
For brand new items with no shipment history, the demand planner can designate a proxy item — an existing product with similar characteristics — and use its historical shipment patterns as a starting point. This is particularly useful when launching a new flavor or format of an existing product where historical performance of the existing SKU can inform the ramp expectations of the new one. To choose a like-product, click on the “3-dots” → Link product history next to the new product. Then select the products whose history to copy over incl. the transition date.
