SEARCH searchBY TOPIC
right_division Green SCM Distribution
Bookmark us
sitemap
SCDigest Logo

About the Author

Karin L. Bursa
Vice President of Marketing
Logility


Karin L. Bursa is vice president of marketing at Logility, a provider of collaborative supply chain management solutions. Ms. Bursa has 25 years of experience in the development, support and marketing of software solutions to improve and automate enterprise-wide operations. You can follow her industry insights at www.logility.com/blog. For more information, please visit www.logility.com.



Supply Chain Comment

By Karin L. Bursa, Vice President of Marketing, Logility

December 6, 2012



Demand Planning at the Sub-SKU Level

Demand Aggregation and Disaggregation are Keys to Creating the Best Possible Forecast at all Levels of Granularity


Demand planning, sourcing, supply and production planning all perform better when demand is disaggregated from a high-level forecast down to accurate, granular, sub-SKU forecasts as early as possible in the product life cycle. For most companies, sub-SKU forecasting (the task of translating high-level forecasts into specific quantities by size, color, configuration, etc.) is an arduous task likely managed with spreadsheets or systems that are limited to 2, maybe 3, attributes.

Bursa Says:

start
Flexibly defining profiles from existing product demand data is the start, but scalability is essential.
close
What Do You Say?
Click Here to Send Us Your Comments
feedback
Click Here to See Reader Feedback

Our Guest Experts

Nimble Supply Chain: Visibility and Agility

Supply Chain Comment: Common Sales and Operations Planning Myths

Supply Chain Comment:Regulatory Compliance in a Changing Environment

Supply Chain Comment: Planning Optimized - A Journey of Excellence

Supply Chain Comment: e-Commerce is Just Commerce

The Battle for Supply Chain Talent

Demand aggregation and disaggregation are keys to creating the best possible forecast at all levels of granularity. Each is required to reconcile corporate (strategic) plans with operational (tactical) plans. In a demand aggregation hierarchy, the lower levels represent demand for sub-components, while the higher, executive-friendly levels summarize demand by product family, group, region, etc.

High-level demand planning is more accurate and is the focus for most planners. It is much easier to predict how many SUVs will be sold in North America this year versus how many white Toyota Highlanders with a sunroof will be sold in Boston over Memorial Day weekend. The forecasting problem is exacerbated at granular levels due to lower demand, and multiple product and market attributes such as finish, style, color, size, gender, region, speed, power, material type, trim level, configuration, etc. For example, a ceiling fan manufacturer may start with a base model that has several configurations for finish, color, speed, blade length, and lighting kit. Generally, the most accurate forecasts are achieved by disaggregating high-level demand down to tactical levels.

The Demand-to-Supply translation is crucial. While the demand plan should be unconstrained, the handoff to supply planners and sourcing teams can be hurt by the fact these groups have less knowledge of, and exposure to, the market. Ideally, demand planning and supply-side execution should be fully informed and free of conflict. In the real world, too often strategic planning and day-to-day tactical execution work on very different levels. Closing the gap between high-level forecasting and granular supply planning, sourcing, and production orders has been nearly impossible for most organizations, giving supply teams little insight regarding, for instance, how to best adjust order quantities to meet vendor minimums. These decisions are much easier when made in the context of a finer-grained demand plan.

One way to achieve a more granular demand plan is through proportional profile planning which allows demand planners to work at a level of aggregation that matches their business requirements. SC Digest’s Dan

Gilmore recently reviewed this approach in his video Supply Chain Digest’s Cool New Product of the Month. Planners are also able to allocate the forecast more accurately across product attributes and options at lower levels of aggregation before submitting to the supply-side team. Because the number of products under management proliferates exponentially into attribute-based sub-SKUs, proportional profile planning must handle an essentially unlimited number of configurations.

Proportional profiling addresses a common, but hidden, flaw in the demand planning process: product family or category-level forecasts that are accurate enough for monthly S&OP meetings can be significantly skewed at the SKU/location level. Giving demand planners the ability to forecast at the sub-SKU level lets them hand off more complete and accurate forecasts to the supply-side team for sourcing, production, and replenishment planning. Taking the burden of ad hoc SKU-level forecasting off the shoulders of the supply team vastly improves procurement decisions, vendor negotiations, and managing against vendor minimums.

Flexibly defining profiles from existing product demand data is the start, but scalability is essential. The profiling function must scale horizontally across thousands or tens of thousands of products, with the ability to select and collect demand histories, organize and apply profiles, and manage a library of them over time. In addition, the system must scale vertically through three, four, or more attribute levels, for instance:

 
• style / region / color / size
• recipe / flavor / container size
• model / capability level / feature configuration
• price range / brand / configuration

Profile creation must be flexible enough to handle literally any hierarchical stack of SKUs defined by any attribute. Most profiles extend to only three or four tiers, but the capability to handle deeper levels of granularity must be available.

Proportional profiles allow planners to work at a level of aggregation that is more predictable and matches their business requirements, then use product attributes and options to allocate the forecast across lower levels of aggregation.

Agree or Disagree with Our Expert's Perspective? Let Us Know Your Thoughts at the Feedback section below.

Recent Feedback

 

No Feedback on this article yet

 

 
.