What is Greenfield Analysis
A Greenfield Analysis (GFA) can be summarized as the analysis of distribution options that do not currently exist in the supply chain. A network study typically begins with the establishment of a baseline - the current supply and distribution networks. Underpinning this, historical data is often used to establish transportation rates for inbound and outbound lanes along with capacity constraints and other facility-specific attributes.
The aim of network design is to make decisions regarding the future and to improve the current state of the network, so GFAs are often done in order to gauge what an optimal distribution network might look like or to determine the next best distribution center(s) given the current state of the network. GFAs can be conducted, at least in principle, by determining rates for a large set of candidate locations, but a Center of Gravity (CoG) study with a broader scope might be employed with the goal of minimizing total travel distance.
Where do other methods go wrong?
There are problems with conducting a GFA or CoG study in this manner, since often there is a significant disconnect between the reality of a supply chain and a Greenfield study, adding to the disconnect that inherently exists between what constitutes a model and reality. The problems associated with typical GFAs and CoG studies, in our experience, can be summarized below:
- Inconsistency of both inbound and outbound rates with respect to historical or distance-based rates (which themselves are not necessarily reflective of future rates)
- Limited scope with respect to the actual locations considered (specific candidate locations vs. the theoretical best location).
- Use of an “as the crow flies” distances vs. actual driving distance
- Lack of incorporation of distribution center characteristics such as capacities or costs.
To expand, often it is impossible for candidate location rates or purely distance-based rates to correspond with the historical rates used in the baseline. In an ideal world, these rates would all belong to the same "family" of rates, that is, these rates would be from the same dataset to ensure consistency. One could, to address this, formulate all rates in the model as a function of distance since transportation rates will in general be proportional to distance, directly lending itself to a CoG study; however, you’ll find that distance can be a tricky metric to go by, especially if actual driving distance isn’t considered.
As an example, the driving distance between Milwaukee, WI and Detroit, MI is 380 mi while the flying distance is 250 mi, such that the driving distance differs from the flying distance by a factor of 1.5, and this relative difference could vary greatly for an arbitrary pair of cities across the world. This problem of distance is further complicated by the fact that a “true” Greenfield study is not limited to any one location and instead, in theory, extends to any location on the Earth, such that we have something of an infinite space problem.
These issues compound even more when there is a need to consider upstream inbound costs or capacity limitations, resulting in a situation where we might, as an arbitrary (but somewhat generous) example, have an ~80% accuracy with respect to rates, locations, costs, and capacities, amounting to a ~40% accuracy on the whole for a model of a future state since uncertainty tends to be multiplicative.
A “true” Greenfield analysis, therefore, would aim to address the problems are commonly associated with the present GFA and CoG methodologies.
Best 5 Greenfield with Inbound
The Logility Solution
Decision-making is best done under a consistent framework, and Logility Network Optimization offers a more accurate Greenfield solution by leveraging its proprietary reference database which generalizes the model to be able to include candidate locations not present in the current supply chain with a much greater degree of accuracy than other methodologies. Logility's Greenfield solution can also easily factor in common location considerations such as lease and labor rates and additionally allows users to define their own cost, time, and capacity attributes to be considered when selecting a Greenfield location.
Due to uncertainty in data on the whole, network optimization is most useful when assessing the relative degree of change from one network state to another, such that rates that come from the same family of rates are much more useful than rates calculated via distance or based strictly on historical data.
To summarize, a “true” Greenfield analysis must have the following:
- Consistent inbound and outbound rates
- Broad scope of locations considered
- Actual distribution center characteristics incorporated
Read more:
True Greenfield--Answering the Service Time Conundrum