Can the future be forecast?

The amount of complexity within the planning and scheduling of supply chains has enabled software suppliers to promote ‘solutions’ that provide answers to planning questions to four decimal places. But is this required?

Think of yourself; you know more about you than anyone else. How far into the future (by week) can you forecast your circumstances with any conviction – three months at best. After that it is absolute guesswork concerning your health, relationships, work, money (mortgage or rent increases) and housing (availability, fire and flood risk). It is the same with forecasting the sales of products.

However, the requirement of planning is not absolute accuracy, but direction, with an indication about the likely amount of change. Are sales likely to increase or decrease; more capacity required or less, more or less purchases of critical materials and so on.

Given the complex nature of supply chains from Tier ‘n’ suppliers at mines and farms through to the customer’s customers, weekly sales forecasts by SKU should only be acceptable for a short horizon. After that, detailed forecasts just become more wrong. A more believable structure is where forecasts are done by SKU by week, for about 8 weeks from the end of the scheduled products ‘freeze’ period , then by product group (using a standard measurement across product lines) for each 3 months out to 12 months, and by half year for the balance of the planning horizon.

Segment to manage the present

But before planning the future, the present must be managed. And for supply chains that commences with the management of finished goods inventory, based primarily on the Variability of sales for each product. An approach is required that identifies SKUs with a more stable pattern of sales and which requires lower buffers of inventory and capacity. This enables the deployment of resources to manage the SKUs that are more unpredictable and variable.

To measure sales variability and to assess the predictability of a sales pattern or how well it can be forecast, use the Coefficient of Variation (CoV). The lower the CoV class, the more stable the sales and therefore its suitability for forecasting. SKUs with a similar CoV class are placed into segments that have appropriate management criteria, which provides about six product flows, or outbound supply chains. This was discussed in a previous blogpost and the topic will be revisited.

Forecasts and inventory control actions provide the ‘one plan’ view of future sales in businesses making and selling fast-moving consumer goods (FMCG) and consumer packaged goods (CPG) – products with a ‘use-by date. Until ‘outside-in’ data collection and analysis capabilities improve, forecasts are unlikely to focus only on end-user demands So currently there is a mix of data which can be relevant to a sales forecast or action, depending on the CoV class or segment. Examples of the data to be selected are:

Inside-out

• internal data concerning orders from Sales and incoming orders and their delivery to customers; list prices and discount criteria; seasonality of product sales, public holiday calendar; product promotions calendar, and other relevant characteristics

Outside-in

• data on consumer purchases, either from retailers’ point-of-sale equipment or sold by retail sales data companies
• macroeconomic information that informs consumer behaviour and trends. Includes: government provided economic data e.g. GDP, CPI, rates of unemployment and inflation; retail sales; housing sales and building approvals. Reports published by firms that sell economic forecasts
• external data and information that provides indications of consumer sentiment and likely demand. Includes: geopolitical events and forecasts; natural disasters that may affect supply; competitor actions; analysis of references in social media to identified industries, companies and products; analysis of searches on the company’s web site; weather forecasts

To ensure that S&OP meetings do not become focused on individual products, nor ‘bogged down’ by excessive data, aggregate data must be used through the process, based on ‘product families’. A ‘product family’ is identified on the need to use capacity at a ‘bottleneck’ resource. The volume by period for a ‘product family’ indicates when capacity at a ‘bottleneck’ is required to balance the expected demand. This is the required outcome from planning the operations of a business, using the Sales & Operations Planning (S&OP) process.

Product families may also reflect the brands or models of major customers, such as when supplying parts to automotive assemblers. The number of ‘product families’ varies between 5 and 15, but preferably in a range of 6–12. The sales forecasts for each ‘product family’ in S&OP is built from the SKU sales forecasts and actions.

The measurement of output for SKUs is expressed as a standard unit of measure e.g. tonnes, shippers (the carton size shipped to customers), litres, hours, pallets etc. S&OP reviews are conducted using ‘product family’ and the standard unit of measure, which provides a consistent communication base across all functions. For each product line (SKU), the Bill of Materials (BOM) within the ERP system is structured to ‘roll-up’ into the SKU’s ‘product family’.

How your organisation’s supply chains are planned must have equal importance to scheduling the production, physical movement and storage of finished products and inbound items. But for Planning to be successful requires that complexity is removed from the process.

Segment to remove complexity

To achieve the objective of removing complexity, Planning commences with segmenting finished goods inventory to reflect the sales patterns of SKUs. For the S&OP process, SKUs are segmented into ‘product families’. To assist Procurement for sourcing, inbound items are segmented into categories and sub-categories (called Category Management), which identifies the inbound supply chains. As customers and suppliers are not seen to be equal, they to must be segmented, as this differentiates the level of service provided to customers or expected from suppliers.

Segmenting enables improved Planning, because the planned activity is better suited to each segment, rather than having only a single response for the supply chain element.

The aim of the Supply Chains group is to provide Availability of products for customers. However, Availability is influenced by Uncertainty in your organisation’s sales and supply markets and at customers, internally within your organisation, at contractors and suppliers. Uncertainty adds to Complexity in your business. To segment products and items, customers and suppliers enables a more focussed management of each segment and therefore the potential for reduced uncertainty and therefore risks.