| We are told that the constant increase in computer | | | | levels (availability) for the system. Such an analysis |
| processing power (one version of Moore’s Law | | | | can tell you the first, best spare component to buy |
| states that computing power/dollar doubles every 2 | | | | (and where to put it in a network) to improve the |
| years) is changing the way the world, and supply | | | | 'native' system availability, and then the second, and |
| chains, operate. This article looks at some of the | | | | third spare, until the spares budget is exhausted. |
| ways the Internet and other computer enabled | | | | The traditional item by item analytical approach, |
| technologies have created 'best practices'. | | | | where only the failure rate, repair time and gut feel |
| The question for you is whether or not you are | | | | are considered, delivers an unrepeatable, biased, |
| keeping up! | | | | sub-optimal spares package for a given budget. |
| 1. Forecasting and Demand Management | | | | If you operate capital intensive equipment and do not |
| Many companies need to forecast customer demand | | | | use system based, marginal spares analysis, you are |
| because their lead-time to supply is longer than their | | | | not keeping up. |
| customers will typically wait for products. The most | | | | 5. Supply Chain Financing |
| common method is to forecast at some aggregate | | | | Many companies have optimised their supply chain |
| level (product family or geography) and then figure | | | | networks and inventory deployments, adopted lean |
| out what to send to each store or regional | | | | manufacturing and streamlined their product |
| warehouse by essentially making another forecast; | | | | development cycles. The final frontier for these |
| such as demand this year will be in the same | | | | leading companies is supply chain financial engineering. |
| proportions as last year. This is done because it is, | | | | Buyers and suppliers are traditionally in conflict as |
| 'impossible', to generate and review a forecast for | | | | each seeks to optimise their payment terms – |
| potentially millions of item/location combinations. | | | | buyers try to extend terms and suppliers trying to |
| Best practice is to let a machine make a statistical | | | | reduce the time between invoice and payment to |
| forecast for each item in each location and then | | | | match their inventory turn. |
| aggregate this demand for management review and | | | | Best practice lies in a lender, usually with a relationship |
| adjustment (by product family or geography). The | | | | with the buyer, to leverage the arbitrage between |
| forecast adjustment can be applied automatically to | | | | the cost of debt to the buying firm and the cost of |
| each item / location in the proportions based on the | | | | debt to (typically) the smaller supplier. |
| original forecasts. This practice can be enhanced by | | | | Payables financing removes the conflict by allowing |
| so called 'tournament' forecasting where the | | | | the supplier to extend terms (say 90 days instead of |
| computer program tries out a whole range of | | | | 60), with the buyer company’s lender funding |
| possible forecast models and picks the best model on | | | | the invoice as early as the day of issue. |
| the basis of the lowest, unbiased forecast error. The | | | | In common with traditional factoring or invoice |
| tournament is re-run every time the underlying | | | | discounting arrangements, the supplier receives a |
| demand data is updated (typically weekly or | | | | percentage of the due payment up front. However, |
| monthly). | | | | with a supplier finance approach, the process is |
| If you are not using item/location tournament | | | | initiated by the buyer through its own bank and, |
| forecasting, you are not keeping up. | | | | thanks to the buyer's stronger credit rating, the |
| 2. Inventory Optimization | | | | terms are likely to be more favourable than the |
| A very common method of managing inventories is | | | | terms on offer to the supplier through a local bank. |
| to apply a policy or rule, typically based on some | | | | This can be structured as non-recourse funds by |
| segmentation analysis (e.g. ABC), to hold a certain | | | | banks that are experienced in these reverse |
| number of weeks of historical average demand – | | | | factoring solutions. |
| for example, 4 weeks of cycle stock and 2 weeks | | | | The lender’s exposure is to the buying company, |
| of safety stock. Unfortunately, rules-based | | | | even though the seller receives the advance |
| approaches tend to be ‘one size fits’ all. This | | | | payment. The lender can achieve a margin which is |
| means, by definition, that the rule will deliver the right | | | | higher than its normal return achieved on money |
| amount of inventory for some items, too much | | | | loaned to the buyer, but below the cost of funds to |
| inventory for other items and too little inventory to | | | | the seller, since it is lending against the credit profile |
| meet service levels for other items. As a result, we | | | | of the buyer. The seller benefits from a shorter cash |
| get inventory imbalances that result it excessive | | | | to cash cycle and may be able to pass some of this |
| inventory costs, impeded cash flow and poor and/or | | | | benefit on to the buyer as a lower price. The buyer |
| inconsistent service levels all at the same time. In | | | | benefits from the positive cash flow effect and the |
| addition, rules-based approaches are only sensitive to | | | | potentially lower price – improving profitability, |
| changes in demand. So what, you ask? Well, this | | | | key ratios and growth opportunities. The lender |
| means they’re relatively static and not linked to | | | | increases its volume of business, improves the margin |
| other important factors, such as service level and | | | | obtainable on the buyer risk and strengthens the |
| forecast accuracy. For example, if you want to | | | | relationship with the buyer. |
| increase your service levels, you have to estimate | | | | The key to making such a system viable is process |
| (i.e. best guess) what change in your inventory rules | | | | automation and real-time visibility of invoice data |
| will deliver this. If you invest in a forecasting system | | | | enabling all parties to track each invoice, its advance |
| and improve your forecast accuracy, a ‘weeks | | | | payment and final settlement – reducing risk and |
| supply’ approach won’t reward you with | | | | cost for the three participants. |
| reduced safety stocks. Here again, you have to | | | | If you are a large buyer of goods or services, and |
| figure out what the impact is and change the rule | | | | are not using automated supply chain financing, you |
| yourself. This gets particularly difficult when you have | | | | are not keeping up. |
| a significant number of items and stocking locations. | | | | 6. Optimal Supply Chain Network Design |
| Fancy doing this for 10,000 products every week! | | | | When one company takes over another, everyone |
| But there is a solution. | | | | talks about 'synergy benefits' and 'economies of |
| Best Practice is to have a service level goal for | | | | scale', which may be true in terms of market share |
| demand satisfaction off the shelf and then calculate | | | | and product portfolios, but they can be slow and |
| the necessary inventory parameters (i.e. order | | | | difficult to achieve in the supply chains. The integrator |
| quantities and safety stocks), taking into account all | | | | usually starts with at least two of everything – |
| the relevant variables: | | | | warehouse networks, transport and supply contracts, |
| Now, if we increase our service levels, our inventory | | | | information systems and organisational cultures. The |
| parameters for all items adjust automatically because | | | | usual approach is to attempt to 'cherry pick' |
| there is a direct link – and our customers are | | | | whatever appears best (or cheapest) at the time |
| happy because we’re servicing them consistently | | | | without necessarily understanding the impact on |
| to target. If we improve our forecast accuracy, our | | | | supply chain risks, inventory levels or customer |
| investment in safety stock will be adjusted | | | | service. Even without a merger situation, |
| accordingly, and we’re rewarded with better | | | | understanding what the growth path should be, at a |
| inventory turns and case flow. The result is the right | | | | strategic level for manufacturing and distribution |
| mix of inventory for each and every time in every | | | | capacity at home and abroad is often painfully |
| location, and the benefits are achieving ‘best in | | | | reactive. Roughly 80% of the cost of a supply chain |
| class’ inventory turn rates and customer service | | | | is fixed at the design stage – the rest are just |
| levels at the same time. | | | | rate negotiations. |
| If you are using rules based approaches to manage | | | | Best Practice is to use integrated graphical and |
| inventory, you are not keeping up. | | | | mathematical models to design the supply chain at |
| 3. Optimisation - Schedules/ rosters / work | | | | the strategic level. The graphical network maps, |
| assignments | | | | usually linked to geographic data (such as road |
| The spreadsheet is the most common tool applied to | | | | network time/distance matrices), allow the |
| the problem of resource scheduling, be it school class | | | | non-mathematically inclined to visualise the existing or |
| rooms, warehouse order picking, operating theatres, | | | | proposed network. The optimisation solvers attached |
| factories, television studios, airports or delivery | | | | to the model allow billions of combinations of facility |
| trucks. The planner in all these businesses is trying to | | | | locations, transport links, capacities, inventories and |
| find an optimal solution, but will generally settle for | | | | customer allocations to tested and a cost or profit |
| the first 'satisficing' solution they encounter based on | | | | optimal solution defined. |
| some internal heuristic of what constitutes a good | | | | If you are not using mathematical optimisation for |
| roster, schedule, route or assignment. | | | | strategic network design, you are not keeping up. |
| Best Practice is to apply computer algorithms to | | | | 7. Sales and Operations Planning |
| solving the problem and generate an optimal or near | | | | Some businesses behave as if the 'good old days' are |
| optimal solution. Recent advances in both computing | | | | still with us. They either have plenty of idle capacity |
| power and applied mathematics have made | | | | that they can flex to meet any customer demand or |
| extremely large problems amenable to direct | | | | are comfortable holding plenty of inventory |
| computation as well as lowering the threshold where | | | | somewhere in the supply chain – and their |
| such methods become economically viable for smaller | | | | shareholders are probably unhappy. These businesses |
| organisations. Computer algorithms can often throw | | | | often pay a bonus to a salesman who can sell above |
| up options a human planner would never consider and | | | | budget, but cane an operations manager who cannot |
| produce solutions that fair, unbiased, more profitable | | | | keep the pipeline full. |
| and quicker to produce. | | | | To these businesses the budget is a once a year |
| If you employ more than one dedicated planner, who | | | | guess at the future (last year plus 5%) and tracking |
| is doing it all 'by hand' in a spreadsheet, you are not | | | | is based on hindsight – what happened last |
| keeping up. | | | | month and where we are year to date. Sales has |
| 4. Spare Parts for Capital Intensive Industries | | | | three sets of numbers – budget, actual and what |
| Keeping capital intensive equipment operating is key | | | | might have been if Operations had delivered. |
| to delivering profitability in industries such as mining, | | | | Operations has the budget and what was planned, as |
| power generation, electronic networks and | | | | well as what needed to be pushed into the schedule |
| commercial airlines. Various maintenance philosophies | | | | to accommodate 'urgent' orders, expedited raw |
| can be applied to keep the down-time to a minimum | | | | materials and the new customer that just popped up. |
| involving scheduled pre-emptive replacement, | | | | Accounting has the budget, variances to budget and |
| condition monitoring or waiting for a breakdown and | | | | straight line revenue forecasts. |
| then responding rapidly. | | | | Best Practice is to run the business using Sales and |
| Determining the spare parts to be held to support a | | | | Operations Planning, on one set of numbers, driven |
| rapid return to service (assuming an on condition or | | | | from a forecast for every item sold. On a monthly |
| on break-down methodology where component | | | | cycle, forecasts of future demand are run for the |
| removal events are randomly distributed) is a 'black | | | | coming 18 months, adjusted with market intelligence |
| art' in most industries, and often left to the plant | | | | and new product forecasts and locked as the official |
| maintenance engineer and advice from the equipment | | | | Demand Plan for all purposes. This Demand drives the |
| supplier, who may have an interest in selling spares. | | | | Operations plan – sourcing, capacity planning, |
| Best practice, as often seen in military aviation or | | | | labour and materials planning, stock levels and |
| NASA, is data intensive since records must be kept | | | | deployment at a rough cut level. The approved |
| of each component failure or removal to provide the | | | | Supply Plan then feeds a forward view of costs and |
| basis of a sparing analysis. This kind of record | | | | revenues that is approved by the CEO and measured |
| keeping is mandatory in civil aviation but is increasingly | | | | against the budget. The budget is effectively the |
| found in modern computer based maintenance | | | | S&OP Plan at the start of the financial year. Top |
| management systems applied to production plants or | | | | down adjustments imposed by 'head office' need to |
| data networks. Other factors, such as Essentiality | | | | be reflected in the real plans with real actions |
| (related to redundancy and failure modes in the | | | | attached. |
| machine design), item cost, cost of down-time, failure | | | | Sales is measured and rewarded on the accuracy of |
| rates (often expressed as the mean time between | | | | the Demand Plan. Operations is held accountable for |
| failures) and total repair cycle time (if the component | | | | executing the agreed Supply Plan. The business has |
| can be repaired and returned to the serviceable | | | | an agreed basis for long-term and near term planning |
| spares pool) can be combined in a marginal analysis of | | | | and unit costs go down since operations gets a |
| each incremental spare part/location combination. | | | | forward view to optimise capacity. |
| The resulting cumulative inventory investment curve | | | | If you’re not running your business on one set |
| represents an 'efficient frontier' - a line of optimal | | | | of numbers, not using integrated Sales and |
| spare part inventories that deliver maximum service | | | | Operations Planning, you are not keeping up. |