There are two components to calculating the financial return.  One is the revenue component and one is the cost component.  On the revenue side we have planning and demand forecasting systems to try and predict the incremental revenue component associated with the products we are looking to buy.  But we don't have good systems to predict the true incremental cost to serve our customers.  That's a big problem.  Instead of figuring out each item's true cost we tend to treat every item the same from a supply chain perspective and apply a standard landed cost factor to all items based on looking at total supply chain costs as a proportion of total revenue.  The problem with this approach is that it grossly distorts the true costs of many items because this averaging hides the cost of more costly items and makes cost-efficient items look more expensive.  Think about how frozen items need freezer trucks which are more costly, or how some items like furniture sets may require additional labor in a DC, or how car exhaust systems may take up a disproportionate amount of DC space because their bulkiness means they go on the floor instead of in rack space.  So how can we go about better addressing the cost side of the equation?

We first need to establish a baseline history of our true or total cost to serve on every item.  This means putting in place a business intelligence system that can automatically and continuously capture granular item costs from all relevant systems in the supply chain - a supply chain control tower 2.0 if you will that incorporates cost into inventory visibility.  These costs can include purchase order cost, ocean freight, import duties, trade finance fees, local drayage, line haul cost to my DC, labor receiving, labor pick, pack and ship, DC overhead, outbound transportation, store receiving labor and inventory carrying cost.  The costs need to be apportioned out in appropriate ways - such as apportioning truck shipments by relative cube volume of each item on the truck, or apportioning a labor receiving activity by the number of cases received of each item.  In addition to apportioning costs by item they should be captured and apportioned along other useful dimensions of analysis - by location (factory, port, distribution center, store, etc.), by route, by supplier and by purchase order.  This allows much greater flexibility and efficiency when it comes to predicting future costs (for example, over a particular route or from a particular supplier).

We then need analytical tools to leverage that cost history to predict the total cost to serve for each sourcing and buying decision we face before committing to that decision.  These tools may need to mix analysis of actual cost history with known upcoming cost changes (like higher transportation costs, or labor rate increases or new regulations or import duty rates).  That means that purely statistical techniques like those used in demand forecasting cannot be used to effectively predict costs in most cases.

New types of business intelligence systems are needed to predict true cost to serve so that costs can be lowered and processes like sourcing and supplier negotiations made more effective.  Leading retailers and wholesalers in particular are investing in systems that can do just that.  They understand that having true cost-to-serve visibility is a strategic advantage and can enable better decision making throughout the supply chain, but especially on those critical, scary up-front decisions related to where should I source, who should I buy from and how should I route the goods through my supply chain.

These new types of business intelligence systems have several key characteristics:

• They can capture actual costs as they happen so that this data can be used to more effectively predict future costs.

• They have a way of predicting future costs even if there is no actual cost history to base it off of (for example, where it is a new type of item being sourced, or importing from a new part of the world).

• They embed powerful analytical tools that allow the data to be manipulated and displayed in a variety of different ways depending on the needs of the user.

• They keep a history of cost predictions and have the ability to compare predicted cost with actual cost when the goods move to see how good the predictions have been.  This is with an eye to tweaking the model to provide ever more accurate cost predictions.

• And finally, they are tightly interwoven with supply chain execution systems so that the cost predictions can be quickly and easily used in an operational context (for example, when considering what country to source a purchase order from).

The sooner companies put in place these types of business intelligence systems the quicker they can get a leg up on the competition by operating with a lower cost base.  By doing that they can reduce the fear factor associated with making the complicated decisions and focus on getting on with the business of making money.

Source: Manhattan Associates


Keywords: supply chain management IT, supply chain solutions, supply chain systems, logistics & supply chain, determining landed cost, Total Cost of Operation analysis