Global supply chain planning at IKEA

Patrik Jonsson1, Martin Rudberg2 och Stefan Holmberg3

1. Chalmers University of Technology Department of technology management and economics 412 96 Gteborg, Sweden +46 31 7721336 patrik.jonsson@chalmers.se 2. Linkping University Division of Production Economics, Department of Management and Engineering, 581 83 Linkping, Sweden +46 13 281566 martin.rudberg@liu.se 3. IKEA of Sweden 343 81 lmhult, Sweden stefan.holmberg@memo.ikea.com

ABSTRACT
IKEAs supply chain has a global spread with growing sales and purchasing in all major regions of the world. This article outlines IKEAs global supply chain planning concept and describes the roles of the planning organization, data quality and software support in the concept. Its cornerstones are mutually integrated planning processes, a centralized planning organization, focus on data quality, use of advanced software support and structured change management during implementation.

1. INTRODUCTION
IKEAs supply chain is global with sales in more than 250 own stores in 24 countries and 32 external franchisees in 16 countries. The stores are supplied through 31 distribution centres, or directly from the 1,350 suppliers in more than 50 countries. IKEAs supply chain consequently has a global spread with both sales and purchasing in all major regions of the world. IKEAs growth has been tremendous and sales are still growing. Currently IKEA plans to open 10-20 new stores every year with a goal to double sales within the coming five year. Considering the pace of growth in sales, the many stores and warehouses, and the fact that some business areas change up to 30% of its assortment every year, supply chain planning is a real challenge. The supply chain needs tight control and high levels of visibility to keep costs down and avoid obsolete inventory and/or stock outs.

91

The IKEA supply chain is mainly make-to-stock (MTS) and only a few products are made to customer orders. Consequently, the entire supply chain is heavily dependent on forecasts. The regions and the stores have traditionally had a strong power and a high degree of local freedom in terms of planning and placing replenishment requests. This has led to a fragmented supply chain planning with local optimization and a lot of manual intervention with plans throughout the supply chain. Furthermore, due to frequent shortage situations some regions have purposely overestimated demand to ensure delivery, which in turn has led to imbalance in terms of demand coverage. Hence, some markets have suffered from stock outs during long periods, whereas other markets have ended up with obsolete inventories. Forecasting has been done on a regional level with approximately 120 users striving for different goals and using different methods. Part of the explanation to this is that IKEA has lacked a common and structured tactical planning of demand and replenishment. In terms of capacity planning, all different parts of the supply chain (stores, warehouses, regions, etc.) tried to optimize its own part of the supply chain, leading to a set of imbalanced supply plans with a low and unstable total throughput with long replenishment times for the supply chain as a whole. The above mentioned situation led to a number of problems with direct impact on performance in a number of ways. First of all, the supply chain had a functional orientation with limited transparency, leading to a reactive behaviour with fluctuating goods availability (sometimes stock out situation and sometimes over stock). IKEA has also used extensive manual work in its planning processes and the planning was based on fragmented and unreliable planning information. Hence, there was a lack of trust between different parts of the supply chain, which even further have enhanced the bullwhip effects in the IKEA supply chain. Other problems related to the supply chain performance was difficulties to get enough attention of data maintenance, the lack of proper follow-up tools to monitor forecast deviations, hard to change mindsets among users, no synchronization of order and stock data, to name but a few. To overcome the difficult situation, IKEA initiated a program (cluster of projects) aiming to taking better control of its supply chain, and enhance performance in terms of delivery service and costs. A new global planning concept was developed and is currently being implemented. Its cornerstones are mutually integrated planning processes, a centralized planning organization, focus on data quality and use of advanced software support. The purpose of this article is to outline IKEAs global supply chain planning concept and describe the roles of the planning organization, data quality, software support and project and change management in the concept and its implementation.

92

The article is structured as follows. First, the basics of supply chain planning processes, centralized planning functions and supply chain planning software is outlined. Then IKEAs planning concept is described followed by a discussion about the roles of the planning organization, data quality, software support and project and change management in the concept and its implementation.

2. FRAME OF REFERENCES
2.1 Supply chain planning processes
Planning is conducted on different time horizons and levels of detail. In practical terms, the problem of different time perspectives and the level of detail required are dealt with by planning material flows and production successively in a hierarchical structure of planning processes. For long-term planning, a somewhat lower precision and detail levels may be acceptable, while short-term planning demands a high precision and level of detail. The planning structure often used in manufacturing companies contains four planning processes; sales and operations planning, master planning, materials planning, and execution and control of material flows. Forecasts and customer orders are important input information to the planning processes. Figure 1 illustrates the relationships between the four basic planning processes.
Strategic/ Long-term

Demand management Forecasting Sales & operations planning Master planning

Tactical/ Mid term

Order fulfillment

Materials planning
Operational/ Short term

Execution & control

Figure 1: Relationships between basic planning processes. Several of the planning processes consequently have mutual relationships. From a supply chain perspective, there are also relationships between different functional and organizational planning processes. For example, the forecasting process may integrate individual forecasts from several markets and products, the sales and operations planning and master planning processes may involve several production and distribution sites, the materials planning may integrate several warehouses, for example by the use of distribution requirements planning. However, these decisions are often distributed among a multiple number of independent decision-makers along the supply chain which leads to incongruence and imbalances on a global perspective (Pibernik and Sucky,
93

2007). Hence, companies typically look for ways to control these imbalances, e.g. through coordination and/or centralization of planning responsibility.

2.2 Centralizing planning functions

Global supply chains are often discussed and analysed based on two major aspects; the configuration and the coordination of the network. It has been established that vertically focused networks striving for low cost production typically should establish a centralised planning organisation (Rudberg, 2004). This is due to the fact that vertically focused organisations operating on a global scale normally have sales, productions and distribution facilities spread around the globe. Under such configurations a centrally coordinated planning organisation facilitates cost efficient material movements, replenishment and order handling. Consequently, theory suggests that a vertically focused network requires a relatively high level of coordination, and therefore relatively high degree of centralisation of the manufacturing management. The corporate staff must play a much more active role in making the vertically focused organisation work and theory furthermore suggests that centralised management offers better cost-effectiveness due to better coordination, possibilities of higher utilisation, and avoidance of duplication of activities (Rudberg, 2004). The performance of a supply chain as a whole is to a large extent determined by the way the different planning processes are coordinated and synchronized (Pibernik and Sucky, 2007). Supply chains coordinated on a central basis will lead to better results in terms of, e.g., overall costs, compared to supply chains with decentralised management (Rudberg, 2004; Pibernick and Sucky, 2007). To facilitate the smooth coordination based on centralizing the planning functions, companies typically develop common policies regarding manufacturing structure, but also in terms of common (standardized) working methods (Rudberg and West, 2008). Such common working methods also facilitate the introduction of standardised software support, which is necessary to support decision-making in global supply chains.

2.3 Software support

When it comes to coordinating a vertically focused but dispersed supply chain, information and communication technologies can be of great help. Advanced planning systems (APS) have been put forward as tools adopted to aid the complex web of decisions that have to be made in a global network. APS is designed to deal with multiple sites and also to include various supply chain decisions in one central planning engine. Hence, stock replenishment, distribution, production and sourcing decisions can be balanced in a centralized function in order to avoid bullwhip effects and also to strive for the optimal use of resources throughout the supply chain (Stadtler & Kilger, 2005). Combined

94

with point-of-sales (POS), stock level, in-transit, current work load, and capacity information, APS can excel supply chain planning in the following ways; Establish integrated supply chain planning processes since it addresses activities throughout the supply chain. Handle vast amounts of data at reasonable run times, and the possibility to run multiple scenarios and what-if-analyses before a central plan is established. Present results in an easy-to-understand environment through visualisation of KPIs, stock and load levels, and so forth; hence increasing the understanding of the effects that the decisions will have on the supply chain performance. A higher degree of automation of planning, which leaves more time for the decision-maker to analyse the supply chain and the actual decisionmaking instead of using the time to find a feasible plan. Since APS used in supply chains have to deal with multi-site environments and is based on a standardised planning process with a high degree of automation, the quality of input data (both master data and transaction data) is even more important than in other types of planning systems. Hence, the organisation has to put a lot of effort into safeguarding that planning data is of high quality. In terms of planning, information is normally distinguished from data, where information is the content and meaning of the data. Information quality is a generic term and has several different dimensions. Normally, distinctions are made between the extent to which information is valid, reliable, up-to-date, complete and straightforward to understand and use (e.g. Jonsson and Gustavsson, 2008). APS requires high data quality to work properly, and in such cases provides the possibility to provide information of good quality to the decision-makers which thereby can make better informed decisions.

3. IKEAS PLANNING PROCESSES

3.1 The supply chain
IKEA is one of the leading home furnishing companies in the world. With its vision to create a better everyday life for the many people, the company has reached annual sales of close to 20 billion (FY 2007) and has close to 120,000 employees. IKEA has more than 522 million visitors per year in its stores all over the world. In addition to the visitors in the stores, some 450 million visitors are tracked entering the IKEA website. IKEAs main marketing channel is its catalogue that is distributed world-wide in 191 million copies (in 56 different editions and 27 different languages) displaying some of IKEAs 9,500 selling articles. IKEAs growth has been tremendous and sales are still growing.

95

IKEAs supply chain is global with sales world wide. The stores, which are divided into three geographical areas, are supplied through 31 distributions centres, which in turn are supplied by some 1,350 suppliers in more than 50 countries. In terms of supply (purchasing), Europe stands for 69% (Poland being the largest purchasing area with 16% of total purchasing), followed by Asia with 28 % (China 22%), and North America with 3%. The majority of sales is in Europe (82%) with Germany as the top selling country (16%), whereas North America accounts for 15% of sales (USA 10%). Asia and Australia together account only for 3% of total IKEA sales.

3.2 Overview of the planning concept

As a first step in developing the new planning concept, IKEA centralised all forecasting activities and need calculations to control stock levels and replenishment throughout the whole supply chain (see Figure 2).
Suppliers IKEA of Sweden
Strategic/ long-term 1. Sales planning (GM, BA, IoS)

IKEA
Distribution services Retail Market intelligence

Tactical/ mid-term Production planning 4. Supplier capacity planning

2. Forecasting (Demand planning) 3. Need planning (DRP)

In-stock In-stock In-transit

Operational/ short-term

Execution

Delivery schedule

Replenishment orders

5a. Transport planning

5b. Warehouse planning

5c. Store planning

Figure 2: IKEAs global supply chain planning concept and planning processes. The new global planning process starts with the corporate sales planning (1) which sets the frames for the tactical demand planning activities (2) at the 12 business areas. The forecasts are thereafter input to the global need planning process (3), which in turn drives the supplier capacity planning process (4) and the planning of the distribution supply chain (transport, warehouse, and store planning; 5a-c). In the following, the sub-processes in IKEAs centralised global supply chain planning process are described into more detail. Note that this article does not cover the operational distribution supply chain, i.e. transport, warehouse and store planning (5a-c). The global planning process is owned by a central function at IKEA of Sweden (IoS), where decisions concerning the

96

number of articles, purchasing, suppliers, distribution, store coordination, and so forth, are made. As such, IoS is the centre of all planning activities in the new planning process.

3.3 Sales planning

The sales planning starts with the overall sales forecast made by the corporate management (Group Management, GM) at IKEA. The forecast is made on an aggregate level in terms of total sales volumes in monetary units for IKEA in total. It expresses the expected sales increase in percentages. The forecast is related to the strategic business plan, involving business cycle and market intelligence issues, and includes the remaining part of the current fiscal year plus five years into the future. The corporate sales plan is updated three times per year. At the other end of the sales planning process, Demand Planners at IoS provide forecasts for each of the 12 business areas in terms of sales volumes taking into account the business areas (BA) growth plans and ambitions for the future. The responsibility for merging the two gross forecasts lies within IoS, where Demand Planners explode the corporate management sales plan into BA sales and compare the two forecasts. When any differences between the two forecasts are reconciled the forecast is broken down into sales frames per regions (a region is a group of countries), and further down into sales frames per product area. These frames are thereafter compared to the forecasts provided at the tactical demand planning level. The sales plan and the frames are reviewed three times per year, and the forecasts have to be adapted to the frames also three times per year. If there is a difference between the forecast and the frame, the Demand Planners should adjust the forecast accordingly. One alternative is to adjust the forecast in a specific country. If this is not appropriate, the forecast is proportionally adjusted in all countries.

3.4 Demand planning

Some 32 Demand Planners are active in the tactical demand planning process, each responsible for forecasting a certain part of the assortment. The tactical forecast resides with IoS and is done on a rolling 84 weeks planning horizon on store level, with new historical sales data loaded once a week (see Figure 3). The operational forecast is a manual forecast (for the most of the time replenishment needs in the stores) registered by the respective sales unit (i.e. store) for the coming three weeks, whereas a tactical forecast (based on sales history) is used for weeks four to 84. The operational forecast and the tactical forecast are combined to create a final forecast for each article on the selling unit level (i.e. the store level). Thereafter the forecasts on store levels are aggregated, reconciled, and compared with the sales frames on the retail forecast group level (i.e. normally country level) and on the distribution services region level

97

(several countries), cf. Figure 3. A Retail Forecast Group (RFG) consists of one to several stores located geographically close to each other. In Europe, a Retail Forecast Group normally corresponds to a country. The Retail Forecast Groups are usually grouped and served by a fewer number of Distribution Services regions, in Europe for example, there are six such regions. The market input on the two upper levels concerns every countrys activity plan (for example campaigns) and estimated price changes on product level. IKEA encourages the individual countries and stores to have local campaigns and activities. But country specific activity plans must be decided at least six months in advance. Activities on store level can be planned on shorter notice. Demand Planners review the forecast on the regional level each week in order to identify those products for which the forecast deviates considerably from actual sales. In case that sales deviate considerably from forecasted figures, the Demand Planner looks for the reason and adjusts the sales figures or forecast model accordingly.
Tactical demand planning
Market input FC on region level
Reconcile forecast

Aggregate sales history

Sales frames

Market input
Aggregate sales history

FC on retail unit level

Reconcile forecast

Need calculation

FC on selling unit level Statistical forecast Operational forecast

Figure 3: The tactical demand planning process.

3.5 Need planning

The need planning process follows traditional distribution requirements planning (DRP) principles (see Figure 4). The stores provide a forecast for each article for the coming three weeks (after the product lead-time), whereas weeks 4-84 are the responsibility of the Demand Planner. The forecasts are netted against current stock levels and safety stock requirements at the stores, and also netted against goods in transit. Thereafter the stores net requirements are aggregated into distribution centres (DCs) and also here netted against DC stock levels and goods in transit to replenish the DCs. Each DC Group is thereafter aggregated and the calculated forecasted demand for the coming 84 weeks is established, of which the coming 26-52 weeks are communicated to the suppliers (depending

98

on the quality of the plans). Volumes are divided between suppliers based on a so-called Supplier Matrix that determines the split of volumes between different suppliers (see Figure 4). One DC SKU could for example be sourced from two or three predetermined suppliers. The Need Planner is the person responsible for the need calculations, and also takes care of the exceptions messages that might occur. Examples of exception messages that the Need Planner must resolve are stock exceptions (low, high, stock-out, etc.), transport exceptions (late-in-transit), and supplier exceptions (capacity, commitment, etc.).
Article 1 Article 2 Article n - stock - in-transits +/- safety stock
DC Group

- stock - in-transits +/- safety stock

Store forecasts

Store 1 Sup 1
DC 1

60%

Store 2 Store 3

Cap Com

Sup 2

40%

DC 2

Store 4

DC Group Forecast Aggregated total IKEA need

Figure 4: The need calculation process (simplified example).

3.6 Supplier capacity planning

The need calculation is used to plan capacity requirements at the suppliers. In the general agreements between IKEA and its suppliers, IKEA often commits to provide a certain volume to a supplier. This is to make the supplier willing to invest in plants and equipment to produce the desired products. Furthermore, the supplier communicates a capacity limit to IKEA up to which the supplier can guarantee delivery of volumes. The supplier capacity planning at IoS includes load leveling between weeks to always fulfill the committed volumes (which also includes shifting volumes between different suppliers), and also to stay within the capacity limits (see Figure 4). This could be conducted by a load leveling function in the planning system or manually by the supply planners.

99

4. IKEAS PLANNING PROCESS ENABLERS

In order to realize the new global supply chain planning concept, it has been possible to identify four main enablers for the implementation; planning organisation, data quality, software support and project and change management, which all are described below.

4.1 Planning organization

The organizational design and function is a key for IKEA to carry out its global planning concept in a standardized way world wide. The supply chain planning responsibility has been centralized to IoS, and two specialized planning positions (Demand Planner and Need Planner) were developed to take the responsibility of the main processes in the global planning concept. There are about 32 Demand Planners and their role is to secure that the sales planning is translated into a sales forecast on article level with a global responsibility. The responsibility includes the global sales forecast accuracy, involvement in the sales planning process, product range changes and development of the sales forecast methodology. The approximately 70 Need Planners focus on securing that the need planning continuously is matching the capacity planning with a global responsibility. Need Planners are responsible for service levels and stock levels in stores and DCs, the global need planning process, balancing the global need and capacity per supplier/category/material, actions on capacity exceptions, supply plan accuracy, and so forth. The more long term planning issues, such as checking and securing supplier capacity, involve the strategic purchasing and trading business support organizations. The establishment of the two centralized planner positions is important for creating necessary specialist competence in demand and need planning, but also for making it cost efficient and practically possible to carry out the planning processes in a standardized way, world wide. Despite of the centralized planning organization there have been some problems implementing the new global planning concept and working methods, partly because of insufficient end-user training and support, insufficient knowledge and involvement by line management and that the planning software are not the main applications for some organizations.

4.2 Data quality

The importance of improved data quality was early identified as an important cornerstone in order to make the global planning concept successful. Several data quality problems have been eliminated but IKEA still has some issues to

100

deal with in order to further improve the planning and supply chain performances. The most severe ones are discussed in the following. Insufficient maintenance of lead time data gave wrong input to need calculation and caused stock out problems in stores. Process improvement was difficult because of incompatible data capture and lead time measurements throughout the supply chain. As a response to these problems, a new lead time concept that assigns clear responsibilities to different actors was implemented. Furthermore, a work group was established with members across the supply chain deciding on working methods and lead time issues and a web-based application (based on a data warehouse solution) was created to visualize lead times and exceptions on missing lead time data. There were also some problems with in-transit and stock data synchronization, which sometimes resulted in double-counting of stocks. Status updates of intransit data were not always available throughout the supply chain. This resulted in incorrect need calculations and delayed order generation which may result in stock out risks. Also the master data quality was sometimes of insufficient quality, for example, due to an incorrect updating process. In summary, incorrect data figures resulting in poor planning accuracy, may decrease the trust in the plans and also result in extra manual work. Extra manual work may also lead to more manual intervention with plans, which is counter productive to the main ideas with the common working methods concept. As it is now, the supply plan accuracy is still quite low, mainly because of low forecast accuracy of new products, manual intervention with plans, and that everyone doesnt follow the established working methods. Also, the project is still in process, wherefore the concept so far only has been implemented for parts of the supply chain (mainly for suppliers to DCs and partly for DCs to store).

4.3 Software support

IKEA has an old and wild grown patchwork of systems and applications, and it is difficult to change any of them. But to facilitate the new planning concept APS software from JDA was implemented to support most of the planning process in Figure 2. The JDA Networks Demand module was implemented to support the forecasting processes. The software, in combination with organizational changes, made it possible to reduce the number of forecasters from 120 to around 30, and at the same time the average forecast accuracy increased from 60 % to 80 %. The JDA Fulfillment was implemented to support the 70 Need Planners in the need calculation and supplier capacity planning process. It also support other roles in IKEAs planning organization with

101

accurate and up-to date information net requirements, stock levels, safety stock calculations, and replenishment needs. In summary, the demand and fulfillment software supported the forecasting and the need calculation with functionality to carry out frequent quantitative forecasting, distribution requirements planning, aggregating and disaggregating forecasts, but also with a user friendly interface allowing for customized visualization, report generation and exception based working methods.

4.4 Project and change management

IKEA has over time struggled with achieving consistent result from its implementation efforts. Although several projects have resulted in successful implementations, the opposite has also occurred leaving the organization with some co-workers not fully adapted to new working methods and tools. The Change Process has not been sufficiently recognized and the affected coworkers have from time to time been brought in to the change process too late, and left on their own too early by the implementing projects. Part of the recent change efforts, the change process has been taken seriously. A Four-Step model has been defined clearly recognizing the need to create awareness in the first step, create interest in what is coming in the second step, making users try out the solution in the third step and finally adopt the changes in the fourth and last step. The experience thus far is that the receiving organizations truly appreciate the implementation approach and the result is very good. An approach like this requires that a staff of resources skilled in Communication and Learning are actively supporting the rollout resources. This might at first glance be perceived as an extra cost, but the true belief is that making things right from the start saves money at the end.

5. CONCLUDING REMARKS
Implementation of the global planning concept has lead to several improvements in IKEAs supply chain, such as reduced stock levels and improved service levels. Figure 5 illustrates characteristics of and main differences between the old and new planning concepts.

102

Old planning concept Functional orientation Limited transparency Reactive behaviour Extensive manual work Unreliable planning information Lack of trust Fluctuating goods availability Over stock

New planning concept One integrated planning process providing reliable planning information A common working-together environment creating supply chain visibility Working methods and tools to detect and deal with problems at an early stage A coordinated and balanced Supply Chain

Figure 5: Characteristics of the old and the new planning concepts. Experiences from implementing the new planning concept are for example that working methods and the planning organization are at least as important as the software support when integrating the supply chain. The focus on data quality and project and change management were vital for making the new planning concept work satisfactory on a global basis.

REFERENCES
Jonsson, P. and Gustavsson, M. (2008) The impact of supply chain relationships and automatic data communication and registration on forecast information quality, International Journal of Physical Distribution and Logistics Management, 38 (4), 280-295. Pibernik, R. and Sucky, E. (2007) An approach to inter-domain master planning in supply chains, International Journal of Production Economics, 108, 200-212. Rudberg, M. (2004) Linking competitive priorities and manufacturing networks: a manufacturing strategy perspective, Journal of Manufacturing Technology Management, 6 (1/2), 55-80. Rudberg, M. and West, M. (2008) Global operations strategy: Coordinating manufacturing networks, Omega, 36, 91-106. Stadtler, H. and Kilger, C. (2005) Supply Chain Management and Advanced Planning Concepts, Models, Software and Case Studies, 3rd edition, Springer

103