SAPSAP in practiceStop 2 of 10
Demand planning in SAP
How the forecasting ideas on this map show up in SAP IBP, and how the agreed forecast reaches MRP in S/4HANA.
Read atChange any time
In the SAP world, demand planning usually happens in SAP IBP (Integrated Business Planning), a cloud planning tool, while execution happens in S/4HANA. The flow mirrors the Demand line on this map:
- History in: sales history and customer orders flow from S/4HANA into IBP.
- Clean the data: pre-processing steps correct outliers and fill gaps, so a one-off bulk order doesn’t distort the baseline.
- Statistical forecast: IBP offers forecast models such as moving averages, simple, double and triple exponential smoothing (Holt-Winters), automated exponential smoothing, Croston’s method for intermittent items, regression, ARIMA and gradient-boosted trees. Planners assign models by segment, or let the system choose.
- Measure the error: IBP calculates error measures, such as MAPE and others, so planners can compare models and track accuracy.
- Consensus: sales and marketing add their input in planning views, often in Excel, and the S&OP process agrees one demand plan.
- Short-term sensing: demand sensing uses recent orders and other signals to adjust the next few weeks.
- Hand-off to execution: the agreed forecast goes back to S/4HANA as planned independent requirements (PIRs), the demand that MRP plans against (in S/4HANA they can also be maintained with transaction MD61).
From there, MRP takes over: it nets the PIRs against stock and open orders and creates planned orders and purchase requisitions.
The concepts don’t change because the tool is SAP. Classification, method choice, bias and value-add all still apply; IBP just runs them at scale.
Words to use
- SAP IBP
- SAP’s cloud planning suite for demand, supply, inventory and S&OP.
- Planned independent requirement (PIR)
- Forecast demand stored in S/4HANA for MRP to plan against.
- Pre-processing
- Cleaning sales history, such as outlier correction, before statistical forecasting.
Planned independent requirements
A planned independent requirement (PIR) is a forecast quantity for a material, plant and period that MRP plans against. PIRs are created by IBP integration or entered with MD61 (changed with MD62, displayed with MD63). Each PIR carries a requirements type, derived from the material’s planning strategy, which decides how PIRs and sales orders interact.
Strategy 10 versus strategy 40
The strategy group is set in the material master’s MRP 3 view.
- Strategy 10 (make-to-stock, net requirements planning): sales orders are delivered from stock and don’t add to MRP’s requirements on their own. The PIR of 1,000 drives production, and it is reduced when goods are issued for deliveries. Total requirement for MRP: 1,000.
- Strategy 40 (planning with final assembly): sales orders consume the PIR. The 300 ordered reduce the open PIR to 700, and the orders themselves become requirements. Total: 700 + 300 = 1,000. If orders reach 1,200, they exceed the forecast, and MRP plans for 1,200.
Without consumption, the forecast and the orders would be added together (1,300), double-counting the same demand.
Consumption mode and periods
For strategies that consume, the MRP 3 view also sets the consumption mode (backward, forward or both) and the number of days to look back and forward. A sales order for 15 June with backward consumption of 30 days can consume PIRs dated between 16 May and 15 June. Getting these right decides whether an order early or late in the month finds “its” forecast.
How IBP sends its numbers
In IBP, demand planners work with key figures (actuals, statistical forecast, consensus demand) at planning levels such as product, customer, location and month. Forecast models are set up in steps: pre-processing (outlier correction, filling missing values), the forecasting algorithms, and post-processing (error measures). When the consensus is agreed, an integration job, typically through SAP’s cloud integration services, sends the released demand to S/4HANA as PIRs at the right material, plant and period.
Algorithms and drivers in IBP
Beyond smoothing methods and Croston for intermittent items, IBP offers regression with external variables and gradient-boosted decision trees that can use drivers such as price and promotions. Models are assigned to product segments, often based on ABC-XYZ or demand-pattern classification, and IBP can pick the best model automatically by comparing backtested errors. Forecast error key figures (MAPE, MAD, RMSE, bias and others) are stored so planners can compare models and track accuracy over time.
Demand sensing
IBP’s demand sensing uses recent orders and shipments, plus other signals, to adjust the forecast for the next few weeks, at a finer level such as daily or weekly by location. It sits on top of the monthly consensus, not instead of it, and should be judged at short lags against the plan it adjusts.
Cleaning up old forecasts
PIRs dated in the past that were never consumed or reduced still look like demand to MRP, inflating requirements and creating unnecessary orders. S/4HANA provides reorganization transactions (MD74 to adjust requirements, MD75 to delete old requirements, MD76 to delete history) to clean them up. Running them regularly, or making sure the integration replaces old PIRs correctly, is part of good planning hygiene.
Choosing a strategy by business type
| Business | Typical strategy | Behaviour |
|---|---|---|
| Make to stock, finished goods sold from stock | 10 | Forecast drives production; orders served from stock |
| Make to stock, sales orders should net against forecast | 40 | Orders consume forecast within the consumption window |
| Make to order | 20 | Each sales order drives its own production |
| Assemble to order with forecast-planned components | 50 or 70-type strategies | Components planned from forecast; final assembly waits for orders |
| Gross requirements planning | 11 | Plan production regardless of stock (e.g. to keep a line running) |
From APO to IBP
Many companies used SAP APO Demand Planning (part of SAP SCM) for years. APO’s mainstream maintenance also ends in 2027, so companies are moving to IBP, which runs in the cloud and covers demand, supply, inventory and S&OP in one model. The concepts (forecast models, consensus, release to execution) carry over; the data model, user interface and integration are new.
Interview kit
Practise it out loud
Answer each question yourself before opening the model answer.
In SAP, demand planning typically runs in IBP. Sales history from S/4HANA is cleaned in pre-processing, statistical models such as exponential smoothing, Croston or gradient boosting generate the baseline, error measures compare models, and sales and marketing input feeds a consensus plan in S&OP. The agreed forecast flows back to S/4HANA as planned independent requirements that MRP plans against.
How does a forecast in IBP reach production planning in S/4HANA?
The released consensus forecast is transferred to S/4HANA as planned independent requirements. MRP nets them against stock and open supply, explodes BOMs, and creates planned orders for production and purchase requisitions for bought materials.
Which IBP forecast model would you use for spare parts with sporadic demand?
Croston’s method, which is designed for intermittent demand, rather than standard exponential smoothing. I’d also check accuracy with measures that handle zero demand, and size stock from the demand distribution.
Connected stops
Where this idea feeds in, or gets its inputs from.