Looking into the future – dealing with forecasts: a curse or a blessing?
Elisa Kreher, Thomas Franken, Marie-Catherine Peressini[1]
Suppliers to the aerospace industry benefit from advance orders from aircraft manufacturers and can plan their production at an early stage with the help of forecasts and production schedules (cadences) from the major manufacturers. What sounds like a blessing for many industries, however, presents aircraft manufacturers’ suppliers with the challenge of using the additional information correctly without getting bogged down in administrative work.
Due to the high level of complexity in the production and procurement of components in the aviation industry, the value creation processes often take a long time and are very demanding.
In addition, the high cost pressure and limited capacities mean that customer orders are produced as late as possible to avoid inventories and as early as necessary to avoid overloading capacities. Reliable short-, medium- and long-term planning is therefore essential in order to achieve an optimal alignment of procurement and production.
In general, production and sales planning in the aviation industry can be divided into three time horizons: The short-term horizon includes current customer orders. The subsequent period primarily contains the customer’s demand forecast (customer forecasts). As the customer’s demand forecast is not sufficient for corporate planning, procurement and production processes in many supplier companies, it is necessary in these cases to fill the more distant planning horizon with the suppliers’ own forecasts (own forecasts).
The structure of the three forecast horizons differs depending on the customer. Some aircraft manufacturers provide their suppliers with a delivery preview via a dedicated supplier portal in the form of CSV files. Others, on the other hand, transfer planning data such as forecasts and orders in the form of Excel files that have to be imported into the ERP system. While the time periods of orders and demand forecasts do not overlap for some customers, but are synchronized weekly, the time series of orders and demand forecasts do overlap for most customers.
In addition to the two major customers, it is above all the numerous small customers who increase the effort involved in the planning process either by providing their requirements forecast in different ways, e.g. in Excel or as a PDF document, or by not providing a requirements forecast at all. The provision via Excel and PDF also ensures a high potential for errors due to the media break.
At the end of the demand forecast, customer forecasts “fray” for most aircraft manufacturers. They then no longer reflect the expected quantity required per month and in some cases even show gaps.
Beyond the forecast period of the customer forecasts, the suppliers must prepare their own forecasts if required. The so-called cadences, i.e. the planning of the expected monthly construction figures of the individual aircraft models, can provide a good initial planning basis for this. The cadences extend much further into the future than the requirements forecast submitted by the customer.
If it is known which components are used in which aircraft models and in what quantities, these cadences can be used to create an initial forecast. Unfortunately, this “parts list” relationship is not always fully known to the supplier companies. In addition, many vendor parts are optional parts that are not installed in every customer order for a specific aircraft model. In such cases, it is not enough to know the “bill of materials” relationship; instead, usage probabilities must be calculated and used.
The weal and woe of the supplier industry depends on the quality of customer forecasts, and the topic is sometimes discussed with corresponding sensitivity.
In order to continuously improve data quality, some aircraft manufacturers have set up their own controlling system to determine the stability and accuracy of forecasts. The “accuracy” of a forecast indicates what proportion of the forecast quantity was actually realized as a customer order quantity in an analysis period. The “forecast stability” is used to measure the uniformity of all forecasts for a specific observation period and is used to measure the fluctuations in the forecast values transmitted each week.
Suppliers are usually faced with the challenge of regularly determining and archiving the required key figures for the data quality of a large number of forecasts. This controlling process is almost impossible to manage without system support and requires a concept for the automated calculation and monitoring of key figures.
On closer inspection, the supposedly comfortable forecasting situation in the aviation industry turns out to be much more complex, uncertain and incomplete than one might think.
Light into the darkness – approach to system-supported sales planning
As a supplier of pipe systems, structural components and auxiliary fuel tanks for aircraft manufacturers, PFW Aerospace GmbH (PFW), based in Speyer, is also dependent on its customers’ forecasts. The large number of variants of PFW components, their long procurement and production times and the relatively short delivery times to the customer made the planning process for procurement and production considerably more difficult and made the preparation of complete and consistent planning data very time-consuming.
As is customary in the industry, PFW also receives a large number of customer forecasts via various channels and then transmits these to the company’s own ERP system in an adapted form after an extensive, time-consuming and therefore error-prone manual preparation process. PFW lacked the technical support to solve the above-mentioned problem areas in forecast planning and to establish consistent 24-month demand planning for all customers.
To date, PFW has implemented sales planning using Excel as a planning tool, within which the customer’s forecast data was analyzed and adjusted if necessary. The aim of the forecast analysis was to identify the biggest deviations in the customer data. In addition, PFW checked during planning whether the incoming customer orders matched the forecast data. However, due to the large amount of data, only a limited amount of data was checked, as the manual effort involved massively exceeded the personnel capacities. Random checks of customer forecasts, some of which were known to be incomplete, entailed the risk that not all errors in the forecast data were recognized and that the resulting incorrect forecasts were passed on to procurement and production, leading to inadequate delivery readiness or excessive capital being tied up in inventories.
For this reason, PFW decided to improve the order fulfillment process by introducing a sales planning tool. The sales planning tool should
- a continuous requirements plan for a rolling 24-month period for all finished parts,
- all data models, from CSV data to Excel spreadsheets, are supported and uniformly displayed in one format as a whole, and
- thus all customer forecasts are consolidated and processed in one system and
- increase the transparency of incoming customer forecasts,
- Gaps in the customer forecasts are recognized and, if possible, closed automatically or reported by alert reports,
- manual control processes can be reduced,
- data transmission errors are eliminated and data processing errors are minimized and ultimately
- the optimized forecast values are passed on to the SAP system as planned independent requirements.
Closing gaps in the future – DISKOVER AERO as a tool for automated planning
Based on a detailed specification sheet, the PFW Aerospace project team identified the DISKOVER AERO planning and forecasting tool from SCT Supply Chain Technologies as the most suitable system, which impressed with its comprehensive and customizable functionality.
Thanks to DISKOVER’s flexible full-service rental concept, investment costs could also be saved and no IT personnel resources had to be freed up or hired for the ongoing operation of the system and support of the system users.
The DISKOVER AERO add-on solution communicates with PFW’s SAP system via flexibly configurable standardized interfaces, so that data exchange between the two systems is unproblematic.
The planning specialists from the management consultancy Abels & Kemmner (A&K) were consulted for the detailed design of the planning process and the resulting software requirements, as well as for the technical introduction of the system.
In the first phase of the project, the joint project team defined the three areas of work
- Standardization of the sales forecast,
- Standardization of the procedures for calculating the requirements plan and
- Controlling
defined.
The large number of types of forecast information provided in varying degrees of detail and the different ranges of forecast information led to a high level of manual effort in planning, which was to be minimized as part of the project. The standardization of the sales forecast was therefore a focal point of the project.
Thanks to the standardization of forecast formats and interfaces for CSV data and to SAP, all forecast information is now imported into DISKOVER AERO and displayed as consistent material number-specific time series over short, medium and long-term planning horizons. For this purpose, customer orders and customer forecasts are combined and supplemented with internal forecasts from DISKOVER AERO.
A planning horizon of 24 months is essential for PFW’s corporate and long-term planning. In order to achieve a 24-month planning period for all finished goods, regardless of the quality and length of the information provided, a standardized procedure was developed for calculating the requirements plan beyond the period of the customer forecasts, known as an internal forecast. A key component in determining the company’s own forecast is the calculation of usage probabilities for the individual finished products based on the aircraft manufacturers’ cadence plans. DISKOVER AERO uses its own forecast to ensure seamless long-term planning, which eliminates planning weaknesses in the form of missing or inadequate customer forecasts.
Depending on the structure of the information provided, the customers’ forecasts were entered as gross or net planning. A net planning mechanism is used when the forecast data has already been adjusted for customer orders. A gross planning mechanism is used if the time series of the order dates and the forecasts overlap.
To ensure the sustainability of the newly introduced sales and planning tool, continuous controlling was implemented to monitor the forecast data. DISKOVER AERO generally offers extensive reporting options as well as an integrated module for pivot analyses and presentations, which is used, among other things, to determine the forecast accuracy and stability of the customer forecasts provided. This means that the quality of the customer forecasts provided can be permanently checked. The sales planning system not only supports the evaluation of past forecasts, but also uses alert functions to provide early warnings if forecasts are missing or drop unexpectedly.
The implemented functionality was tested as part of a pilot phase with selected articles and then rolled out to all planning objects at PFW.
In DISKOVER, all forecasts and customer orders can be displayed graphically and transparently for each material on an aggregated monthly basis or in a daily grid, and transferred to the SAP system in the form of planned independent requirements after checking and any manual additions by the sales planner.
This has massively reduced the planning effort and shifted the focus of the sales planners’ work from the extensive and error-prone manual preparation of series of figures to their intelligent interpretation, checking and supplementation.
The unification, automation and standardization of the collection and preparation of forecast data in DISKOVER AERO and the associated elimination of media discontinuities led to better and faster material number-specific rolling 24-month planning at PFW Aerospace GmbH and thus to a more transparent demand situation.
To paraphrase a well-known bon mot by Mark Twain: “Forecasts are not always a difficult thing, even when they concern the future”.
[1 ] Elisa Kreher is a consultant at Abels & Kemmner GmbH, Thomas Franken is Deputy Managing Director. Head of Development at SCT GmbH. Marie-Catherine Peressini is Head of Sales Planning at PFW Aerospace GmbH in Speyer.
