Over the course of a product life-cycle there are at least six decision points where big bets must be made despite very poor forecast data. These are: long term capacity planning, tooling and capacity management, NPI supply management, A-parts supply management, EOL supply management, and spare parts management. Prior to a product launch, millions are invested in capacity and inventory to meet a forecast that will be off by 30% to 80%. This may result in 5% decreases in margins or significant losses in market share. In volatile industries, the long leadtimes associated with custom parts can turn lean pipelines into quarters or even years of inventory, or can result in customer backlogs of months, when the market demands it in days or weeks. Finally, in some industries, 105% to 120% of the profit is made in the first 3 to 6 months after launch, only to have the 5% to 20% given back to EOL costs and spare parts support.
When it comes to managing risk and flexibility in your supply chain, it is all about reducing the time it takes to position assets, such as capacity or inventory, and then maximizing the revenue earned on those assets. Of course, in the absence of considerable supply and demand uncertainty, the time pressure on the supply chain would reduce considerably, and the risk of having invested too much or too little would all but disappear. Unfortunately, most business tools and approaches take a limited view of the uncertainty problem; for example, relying only on the point forecast as a measure of the demand, and rules of thumb to offset the effects of uncertainty. From leader to laggard we have found that this rule-of-thumb approach significantly underperforms in most business scenarios and even amplifies the effects of forecast error. More bluntly, point forecasts and rules-of-thumb have cost companies billions in terms of lost market-share, expedited fees and inventory carrying costs, write-downs, and write-offs.
As a result, efforts are well underway to develop a more rigorous and comprehensive framework for quantifying and managing the effects of uncertainty. We will explore how this framework can be applied in all three stages of the product life-cycle. In the context of the product life-cycle, we will refer to six types of problems: 1) long term capacity planning, 2) tooling and capacity management, 3) NPI supply management, 4) A-parts supply management, 5) EOL supply management, and 6) spare parts management.
As was discussed in the articles, “What is SRFM” and “Understanding The True Cost of Sourcing,” the Supply Risk and Flexibility Management (SRFM) framework focuses on risk-adjusted Total Sourcing Cost metrics, quantifying the performance of supply agreements (contracts) against a range forecast. The range forecast captures not only the low and high scenarios, but also the dynamic nature in which it might oscillate between the high and low scenarios. The approach also emphasizes risk-metrics, not just a static average or the “if-everything-goes-according-to-plan” projection. For example, most VMI/SMI programs projected zero inventories for the buyer on average or by plan, but resulted in considerable inventory liabilities when the forecast melted. A primary purpose in creating forward-looking risk metrics is to identify and then mitigate exactly these types of exposures.
Let’s turn to a few examples. In the last 3 years, several automotive manufacturers introduced a sunroof integrated into an all glass roof. Clearly, the adoption of this new roof was highly uncertain; first, it was more expensive than the traditional sunroof, and second, many customers may still prefer the traditional sunroof. In the auto industry, it is commonplace for the buyer to pay for the production specific machines, tooling, fixtures, and gauges. Hence, the buyer is faced with making an investment that will determine the capacity level long before the adoption of the option is known. Therefore, the impact of the investment decision, in conjunction with the company’s aggressive policies regarding customer backlog, led to a critical trade-off between price risk and availability risk.
Relying on a rule of thumb to cover the plan plus a standard percentage, the buyers consistently found themselves over-investing in capacity. In this case, the benefit of applying SRFM was threefold. First, in receiving a range forecast instead of a point forecast, the buyers knew what range of outcomes they would need to cover. Second, knowing the range of demand that they would likely need to cover, the buyer can evaluate a range of strategies, factoring in the initial investment plus the cost and time to expand capacity. Third, by quantifying the performance of these different strategies, the business objectives could be met at the lowest cost and risk. Below is a sample of the output from the analysis. The first alternative, 82k, corresponds to the rule-of-thumb, the abbreviation OT represents overtime and Exp represents capacity expansion. Of course, the capacity expansion required a considerable leadtime.
Without delving into all of the details, the take-away from the tables and charts are that by decreasing the capacity investment, prices would have reduced by 6% and 9% in the low scenarios, 5% and 6% in the medium scenarios, and 4% and 0% in the high scenarios. Additionally (the colors correspond to percentiles), there is roughly a 15% chance that if the 62k option is selected, the capacity expansion will be required. Additionally, when the shortages occur, they almost always are less than 5%, with only a few periods seeing a small likelihood of reaching 10%. Given that consumers of this brand are willing to wait some amount of time for their vehicle, this backlog was consistent with retaining most of those customers.
Certainly the challenges with new product introductions are not limited to the auto industry. Apple experienced shortages with certain colors of the I-Mac and both releases of I-Pod. The PC manufacturers struggle with these issues as they enter into broader consumer electronics. The same can be said of almost every high-profile launch of a perfume, cologne, or other cosmetic line where successful ad campaigns can triple demand and failed launches can result in tens of millions being written off. And finally, the toy industry is riddled with $100 million dollar misses on the Tickle-Me-Elmo (not enough capacity) and the latest generation of Star Wars action figures (too much inventory).
Post launch, the focus turns to managing A-Part spend. Typically these parts are high-value, long leadtime, exposed to allocated capacity, or available from limited sources. The primary purpose of SRFM in this context is to balance the trade-offs between availability and liabilities while continuing to hit price targets. It goes without saying that long leadtimes are impediments to flexibility. Unfortunately, business cycles can create problems out of even short leadtime components. Towards the end of 2000, many components were on allocation, prices were increasing, and shortages were rampant. In a matter of months, capacity utilization dropped to as low as 30% in some sectors, and months of inventory become quarters if not years of material. The write-offs and write-downs were well documented in the hundreds of millions. Today we already see this cycle repeating itself. For example, fabs (bare boards) went from rock-bottom prices to constrained supply in last six months of the year.
Consider the case of one electronics capital equipment manufacturer. The equipment is highly configurable, coming in over 10,000 possible configurations. Fortunately for most of their products, most of the supply challenge revolves around just 20 part numbers that account for nearly 70% of the cost on the BOM. Here the application of SRFM is twofold; first, to negotiate flexibility terms commensurate with the uncertainly levels in this high volatile industry, second, to monitor the ongoing supply position to proactively identify bottlenecks and to ensure balance across the commodities. Below is an example of the supply position report used by the capital equipment manufacturer.
There are several noteworthy elements to this report. First, the report consists of forward-looking projections, providing a management level overview of the state of the supply over the upcoming months (the next year in this example – more likely the next quarter or two). The analyst can also provide month-by-month drill downs, in case the average performance over the quarter or year does not reveal significant exposures in any particular month. Second, it reports both average performance as well as risk metrics, as they are defined by the decision maker. As in the VMI/SMI example described in the introduction, the Y Channel satisfies the average inventory constraint (less than 40 days), but fails the risk-inventory constraint (less than 120 days). Third, the report highlights the material in violation of any of the management goals. Again, this report is just exemplary. In the actual implementation, additional metrics were reported, such as purchase level recommendations, cash outflow, and maximum supportable ship plans.
Finally we turn our attention to the end of the product cycle. Perhaps the biggest source of uncertainty here is the timing of the transition. Swift changes in the marketplace may force an unplanned obsolescence, leaving the pipeline stuffed with rapidly depreciating material. Then there are other products that hang on long after their planned termination, either because their successor is late to the market, or the customers are simply unwilling to let them go. Complicating the situation is the fact that suppliers may also be exiting the market, driving up the part costs at a minimum, but potentially also driving volume commitments to sustain the aging technology or life-time buys. Just as we saw in the A-Part supply management discussion, contracts would be negotiated to balance the trade-offs between price, availability, and liabilities, and the Supply Position Report would support the ongoing management of these risks.
History has shown that wherever assets meet uncertainty, the risk of multi-million dollar misses, or even stock-price altering events, is real. This happens at stages of the product life-cycle. The right processes, tools and frameworks for managing these risks can generate huge savings, as well as protect the income statements and balance sheets from violent swings. An upcoming Parallax View article, “Making SRFM Happen,” outlines the necessary steps to develop the processes, tools and framework. At the heart of these steps are the ability to capture the uncertainty that you are trying to manage, and ability to project the performance of your initiatives against this uncertainty. As always, the right set of metrics will ensure that you are asking and answering the right set of questions.