- Home /
- Products /
- SCM /
- Epicor IP&O /
- Demand Planning
Epicor IP&O for Demand Planning
Meeting customer demand while keeping inventory costs manageable has never been harder. Today’s supply chains are more volatile, customer expectations higher, and demand patterns less predictable. With Epicor IP&O’s powerful forecasting capabilities, you can take the guesswork out of demand planning.
An AI-Powered Demand Forecasting Solution
Epicor IP&O empowers enterprises to automate inventory forecasting and identify the right policies to improve service levels and minimize excess stock. With the help of IP&O, you can accurately gauge complex demand scenarios and test proposed policies, empowering you to create profit-maximizing stocking policies.
- Predict future performance and costs given current and proposed stocking policies
- Automatically identify optimal service levels that yield the lowest cost
- Develop a transparent, efficient, and repeatable planning process
- Leverage statistics and business knowledge to build consensus demand plans
Benefits
Field-Proven Statistical Forecasts
Epicor IP&O addresses all types of demand including seasonal, trending, intermittent, and cyclical patterns and selects the right technique automatically.
Probabilistic Modeling
IP&O generates thousands of realistic demand and lead time simulations for each item to reflect real-life uncertainties. It stress-tests current and proposed policies yielding accurate predictions of future costs, inventory value, and service levels that enable you to make fact-based inventory policy decisions.
Patented Intermittent Demand Forecasting
IP&O’s patented bootstrapping method for intermittent demand has been added to the automatic forecast selection along multiple proprietary enhancements.
Frequently Asked Questions
It is recommended to have at least 1 year of data, ideally 2 years of data or more. It is not always true that more data is better and you can control how much history is fed to the forecast engine on an item specific or group specific basis if desired using the “forecast parameter” feature in IP&O.
All of your historical transactions including shipments, sales orders, transfers, returns, and supplier receipt data is imported. In addition, all standard attributes and master data is imported including product descriptors, item costs, supplier lead times, and stocking policies (min/max/safety). While all data is imported into IP&O’s database, you can configure in the end-user applications how much of the data will be used by the end-user demand planning, inventory optimization, and reporting applications.
Demand data and attributes are imported into IP&O daily. This ensures that product attributes, costs, and other fields are always synchronized with the data in your ERP. Note that demand data is collected in IP&O but not rendered to the demand planning and inventory optimization until either a week or month of actuals is available (if you are forecasting in weekly buckets then once per week a new stat forecast will be produced, if using monthly buckets then a new forecast will be generated automatically once per month). You can choose to save forecasts back to Epicor ERPs as often as you prefer and the cadence that works best for your business will be agreed to during implementation.
That’s OK. Epicor IP&O can import data via file imports. Simply send files of historical transactions extracted from your legacy systems and IP&O will automatically import them. When you go-live on the ERP system, IP&O will append your data with transactions from the new ERP system.
The software conducts a tournament with each forecast model competing to win. Current historical data is held out and each model uses the prior history to forecast the hold out period. The forecast model that has the lowest error wins and is used to forecast the future using all the history.
There are dozens of models that are dynamically evaluated in the tournament. There are 7 classes of models that are used. The tournament will automatically try out many different variations of each model using different averaging periods, weights, and other consideration.
- Winters Multiplicative (Seasonality or Seasonality & Trend)
- Winters Additive (Seasonality or Seasonality & Trend)
- Double Exponential Smoothing (Trend)
- Linear Moving Average (Trend)
- Simple Moving Average (No trend or seasonality)
- Singe Exponential Smoothing (No trend or seasonality)
- Bootstrapping (Intermittent)
Yes, you can select a desired forecast method from the available options. You can even specify the specific parameters for the model if desired including averaging periods, weights, and more. You can even run a tournament using a user defined selection of methods. This is particularly useful if you want to guarantee the forecast includes a certain behavior. For example, if you always wanted the forecast to show a seasonal pattern, you may elect to run a tournament with just the seasonal models. In addition to selecting your own model, you can configure the amount of history to be used and other forecasting parameters.
You can use “freestyle” forecasting that enables you to specify a user defined forecast quantity and trend. For example, you can the starting forecast and growth rate or ending quantity.
Yes, for no additional implementation fee you can use the native “snapshot” import feature to load external forecasts. Forecasts must be in the format required by IP&O. All snapshots are then accessible from the snapshot drop down within the graphical view. If you’d like this process automated, then we can do so as part of the implementation with some additional time.
Yes, you can apply overrides to any forecast at the product level or group level. Changes made at one level are automatically reconciled at other levels.
Yes, every time an override is saved it is recorded. The audit trail is logged and maintained in IP&O according to the project settings. Simply click on the audit trail icon within any of the cells and you can see the prior changes.
No. All forecast overrides are preserved even when the system auto reforecasts. If an override has been applied the newly computed forecast for that period will be disregarded.