UK: Pile it low, sell it fast.

UK: Pile it low, sell it fast. - An ongoing revolution is transforming the world of inventory management: make to order, not to forecast.

by Malcolm Wheatley.
Last Updated: 31 Aug 2010

An ongoing revolution is transforming the world of inventory management: make to order, not to forecast.

Have you ever played 'the beer game'? Developed at Massachusetts Institute of Technology (MIT) in the 1960s, it was popularised by Peter Senge in his early 1990s book, The Fifth Discipline, and is now an indispensable part of many a management consultant's toolkit. The players in the game - each playing either a retailer, wholesaler or brewer - must place orders for the beer (or in the case of the brewer, the ingredients) that they think they're going to need, based on the judgments that they make about likely demand up or down the supply chain. It's a microcosm of the real world of inventory management; a world full of guessing, second guessing ... and a lot of hoping and praying. On most occasions on which the game is played, even a small movement in the demand in beer from the consumer can cause the supply chain to oscillate into chaos, inventories to pile up and costs to spiral upwards. The game is a sobering reminder of just how difficult inventory management can be.

It could all be so different. Slick inventory management can transform a business. Professor Daniel Jones, co-director of the Lean Enterprise Research Centre at Cardiff Business School, is one in a long line of experts pointing out the startling transformations in business performance that are possible when a business takes a stern enough stance on inventory management. Jones cites improvements of up to 90% in both throughput times and defect levels.

The problem in dispensing advice on inventory management is that its starting point varies markedly according to which business you are in.

A manufacturer of automotive components might wish to eliminate inventories, but a retailer of the same components would perhaps be unwise to adopt the same goal. Even at a given stage of the production process, generalisations are difficult. A shop with no stock is seemingly laughable but try telling that to successful Internet bookseller, which (contrary to stories about the company) does not possess a huge warehouse stuffed with an unimaginable number of titles but it has a cosy relationship with a book wholesaler which does.

The importance of inventory management is of course unevenly spread across industry sectors. Some industries are much better at it than others, notes Colin New, professor of manufacturing strategy at Cranfield University School of Management and chairman of the Management Today Best Factory Awards' judging panel. 'Repeatedly, the Best Factory Awards process has highlighted the fact that one of the strongest drivers of excellence in inventory management is the market discipline imposed by harsh, demanding customers,' says New. For evidence, he suggests, 'look no further than the supplier base of the UK's large supermarket groups, or the first-tier suppliers of the automotive giants: you either get good, or you get out'.

These days, 'getting good' generally means looking at a problem and then simplifying it - or at least changing its nature. Guessing the level of demand is no match for actually knowing the level of demand. True, algorithms employed in forecasting supply chains are infinitely superior to the crude techniques employed even a decade ago. Leading retailers like Tesco, for example, can factor into their forecasts variables such as the weather, seasonality and applicable bank holidays - one of a range of fine-tuning techniques that has enabled them to halve their inventory stockholdings in recent years to one week's worth of sales. But there's nothing that's quite so easy to forecast as an order that a business has already received, especially for businesses where the hapless management must not only forecast the aggregate size of the demand, but a complex array of sizes, shapes, colours, and so on.

Hence the sweeping transformation of industry's approach to the whole inventory management problem - an ongoing revolution that began in the early '80s. The solution: make to order, not to forecast. It's a powerful technique and one that businesses are increasingly incorporating within not just their approach to inventory management, but entire manufacturing and commercial strategies. The purpose-built factory of one of the winners of the 1997 Management Today Best Factory Awards, Bertrand Faure (UK) Ltd, is deliberately located midway between Honda's Swindon car assembly plant and Rover's Cowley plant in order to do precisely that. Every 34 minutes the factory despatches 54 cars' worth of front and rear seats to Honda, loaded in the order in which they will be required. When the seats leave the factory, the cars onto which they will be fitted are already going down Honda's assembly line.

Seemingly obvious (at least in hindsight), it's difficult today to understand how radical a departure 'make to order' appeared when businesses first began seriously adopting it in the early '80s through initiatives such as just-in-time and kanban-based procurement. 'The supply chain back then simply wasn't agile enough,' recalls Kumar Bhattacharyya, director of Warwick Manufacturing Centre, 'and it took the Japanese coming in to shake things up.' Even today, he suggests, industries touted as apparent exemplars of inventory excellence have in reality supplier bases that substantially lack the responsiveness to the demands of their customers. The result? Plaudits for companies that do manage to make to order - but gargantuan inventory holdings in their suppliers' warehouses in order to make such responsiveness possible.

Perhaps surprisingly, inventory is often to be found in the customer's goods inwards warehouse, rather than the supplier's finished goods warehouse.

One way for a business to 'buffer' demand without damaging the ability to respond to variations in the order book is to require suppliers to carry the risk, through an approach known as 'consignment stock'. This has been panned by critics such as Brian Scholey, a principal consultant at KPMG, as 'a substitute for management'. 'The lazy man's route to inventory reduction', it seemingly wipes inventory off the books without doing anything much to eliminate the costs of inventory from the supply chain.

Consignment stock does have clear virtues, however: Dell Computer Corporation, for example, has achieved an enviable inventory management performance by making a virtue out of the fact that the components that make up the company's products must be sourced on relatively long lead times from suppliers around the globe or, in the case of monitors and speakers, need not be incorporated into the product during manufacture, but can be united with it en route to the customer (see box).

But inventory still physically exists and must be financed, stored, tracked, insured, moved, lost and (ultimately) be declared obsolete if fashions or technologies move on and the goods on the shelves can't be shifted.

John Rowe, Sainsbury's director of logistics, reckons that a quarter of the food retailer's £1.5 billion supply chain management budget 'is spent confirming that inventory is where it ought to be, and that it is moving as it should'. And even after making substantial inroads into inventory levels, Graham Booth, supply chain development director at rival Tesco, reckons to take a further £100 million out of its supply chain cost base over the next four years principally through further enhancements to its reordering system, category management and better handling techniques at its distribution centres and stores. Like other businesses, it is currently embracing a fashionable retail-based inventory management initiative known as Efficient Consumer Response (ECR) that aims to see retailers and their supply chains work more closely together through a variety of processes: cross-docking (to eliminate warehouses); sharing sales data gathered at the checkout; and transmitting orders electronically.

Meanwhile, explains Ian Walker, head of the UK supply chain group at management consultants CSC, manufacturing industry is beginning to embrace a new initiative of its own: strategic procurement. This aims to condense the demands for components that arise when manufacturing businesses buy similar products for different purposes, under different part numbers and from different suppliers. An early implementation at National Power's engineering spares, notes Walker, found that some identical items were held under 14 different codes; around 50,000 items had had no movement for three years, and 350,000 items had no apparent users. By rationalising this, and separating the demand for easy-to-predict maintenance spares from the demand for impossible-to-predict breakdown spares, savings of £100 million were identified.

Others are eagerly awaiting savings of their own. At the £3.5 billion global defence business, GEC-Marconi, for example, group procurement director Russ Armitage points out that with 30 sites in the UK and 10 in the US (one of which had 160 different numbering systems and an overall external procurement budget of £1.4 billion) the potential for savings is enormous.

'Before, we'd been trying to work things out by hand,' he says, referring to the recent deployment of a specialist software package to assist with the exercise. 'We had the philosophy, but lacked the tools.' It's an observation that sums up many of the approaches to tackling inventory management.



Dell Computer Corporation's inventory levels remain enviably low: the combined figure for work-in-progress and finished goods stands at just 11 days' worth of sales. Since 1983, stockholdings have been eschewed in favour of a make-to-order policy. 'For us, build-to-order means something quite precise,' says Phil Hubble, vice president of European operations. 'It means built to the exact specification of the final consumer's requirement.' By contrast, most of the machines bought from a high-street retailer or an industry major such as Compaq or IBM are built to forecast. Each day, Dell's sales subsidiaries across Europe take orders from the customer and transmit them to the factory for manufacture. The following day - or at latest the day after - the machines in question are trundling down the production lines in the Limerick plant that serves the European market.

Between 30%-40% of the plant's raw materials are procured on this basis.

Departures from anticipated demand patterns are rarely a problem, says Hubble: 'typically it's make a little more of these components and rather fewer of those, and the overall quantity remains the same.'A further 45% of material requirements comes from suppliers' 'hubs': stockholdings of components which are called off to order once the mix of customer orders is known.

The logic is simple: sized like a television set, there's no point shipping items like monitors to Limerick, to ship them again on to the final customer.

So Dell stores such items as close as possible to the final consumer, and ships the computers via a hub at which they can be united with the monitors to which they are to be coupled.

The stockholding at the Liverpool hub which serves the UK market amounts to some 20 days' worth of inventory. Crucially, the ownership of the stock lies with the supplier, and the monitors and speakers for each computer only go into Dell's inventory for around half a day before being invoiced to the customer. Similar hubs are located close to Limerick for high-value items such as hard drives and chips which have to be installed inside the computer.

As the price of memory chips and other components continues to fall, the company's exposure is limited to little more than the computers on its assembly lines. Similarly with the periodic quality problems that dog the industry, limited stock means limited risk.

Find this article useful?

Get more great articles like this in your inbox every lunchtime