- Customer satisfaction with reliability, downtime, and repair costs
- Inconsistency of demand (some parts have intermittent demand, while others may have high demand)
- Volume of spare parts actively managed and the risk of part obsolescence (caused by continuing increase of machine complexity and the introduction of new products and parts)
In addition, some organizations have entire programs focusing on repairing used parts. These organizations must anticipate which parts can be salvaged, how many times they can be repaired, and which parts cannot be saved.
For organizations that manage large fleets or provide machine maintenance, the challenges are even more extreme.
Demand forecasting types
In general, spare parts have intermittent demand; there may be demand at one point in time followed by a long period of no demand. This makes spare part demand difficult to predict, but there are several demand forecasting methods that can help:
- Croston’s method: This classic method deals specifically with intermittent demand. It is based on the Simple Exponential Smoothing method, but greatly increases its accuracy.
- Syntetos and Boylan Approximate (SBA) method: This method is similar to Croston’s but includes an added correction factor, which decreases errors.
- Artificial intelligence (AI): Advancements in AI have made possible new forecasting methods that leverage neural networks, for example.
- Monte Carlo simulations: These simulations predict when parts will fail over a defined period of time with different variables.
Correctly classifying the types of spare parts in your inventory is essential to forecasting. A high-level breakdown may include the following categories:
- Slow moving
Once all spare parts have been classified, organizations should do the following:
- Align demand forecast techniques to part classification.
- Determine key performance goals for inventory turn and customer service levels.
- Identify inventory control methods.
How part classification categories are defined, which methods are used for demand forecasting, and inventory levels may vary from industry to company. However, the key is agreeing upon classifications, metrics, goals, and the constant measurement of performance.
New product introductions
Predicting spare part levels for existing machines, engines, autos, etc. is hard enough — but what about new products? At a high level, inventory managers typically have two choices: predict anticipated inventory levels based on reliability models or simulations, or let the natural demand occur over time. Many organizations do not have advanced software solutions to run computationally intensive simulations, so they are left with reliability modeling or some level of estimating what parts will be needed.
Vanguard Software has helped some of the leading engine manufacturers, defense contractors, and aerospace companies with very complex spare parts forecasting projects. It is through this domain experience, combined with a market-leading platform, that Vanguard is positioned as an ideal partner for organizations looking to solve spare parts issues.
About Vanguard Software
Vanguard Software introduced its first product for decision support analysis in 1995. Today, companies across every major industry and more than 60 countries rely on the Vanguard Predictive Planning platform. Vanguard Software is based in Cary, North Carolina.