The Stream by River Logic

A Multi-Purpose Approach to Logistics Modeling

October 2, 2015 | By Eric Kelso

About a decade ago, I would occasionally watch the TV show Good Eats. I’m not a chef; I’m not even much of a cook. I can still burn microwaved popcorn. Truth is, I just enjoyed the science aspect of the show. You might be asking “What does a food show have to do with logistics, besides the obvious issues of transporting and storing food?”

The reason why I even remember that particular show is because the host, Alton Brown, impressed upon his viewers in nearly every segment one important rule he lives by: he never buys a gadget that only serves one, limited purpose. If something is designed for only one specific task, he reasoned, then chances are it will be used infrequently and simply take up space in the kitchen; that he was better off using a tool that can do not only that job, but enable him to do many other tasks whenever necessary. This was generally true even if his chosen multi-purpose tool was sometimes not quite as fast or easy-to-use as a single-purpose tool he could've used.

Logistics Modeling Software is Too Often a One-Trick Pony

This analogy especially applies to the software market for logistics modeling, which I’m loosely defining as a conglomeration of mostly supply chain network optimization with route planning, inventory optimization, truck loading, and many other niche logistics-focused applications.

As software development tools have become much better over the last few of decades, a proliferation of commercially available software applications now exist for solving a wide range of logistics issues. When available solutions either didn't meet requirements, the cost was too high, or companies just wanted more control, custom models were built for their internal use. Most such home-grown systems typically are coded in Excel or a 4th-generation language and can include rule-based, optimization and/or simulation methods.

Consequently, the net result is that most companies now have a virtual toolbox full of disparate software from different vendors, possibly running on different hardware platforms, feeding from and into different data systems that likely have different internal users making different (and possibly conflicting) decisions.

If Integrated Business Planning (IBP) is the ideal, if not readily achievable, is it truly better to continue relying on a suite of software applications for logistics decisions — each with their own hardware, data, training and other requirements? Is it possible to adopt a more general, multi-purpose approach to achieve approximately the same value?

Part-Time Use, Full-Time Cost

This seems like a simple question: would you normally spend a lot of money for software that, on average, you use about 10% of the time (sometimes even less)? Something that does one task well and looks nice but then sits on the (virtual) shelf for 10 or 11 months of the year adding no value?

Of course, most people would scoff and answer “not likely;” yet, this describes the typical usage for supply chain network optimization software. Most companies in need of such a tool cannot re-evaluate their supply distribution network every week, or even every month. It’s simply too disruptive for most businesses. Even if they wanted to, most businesses just cannot open or close a warehouse on a whim, as there are leases and 3PL contracts to be signed, employees to be hired, and many other details to be worked out.

For many companies, such decisions require a careful planning exercise that normally triggers a one-off project, which also might mean hiring outside consultants. Some companies will go years before re-evaluating their network again. For others, they might re-evaluate their network every quarter at most, but more likely every six months to two years.

Does this describe your experience? You leased some vendor’s supply chain network software, worked like crazy getting trained on it, collected data, ran countless scenarios, analyzed results, created reports, gave many presentations and finally made a recommendation.

And then you were done. The software went back on the shelf until who knows when; maybe next year, maybe never.

Full-Time Use, Full-Time Value

Good news. There's an alternative to the above scenario: use a 5th-generation, visual modeling language that incorporates most of the same necessary modeling capabilities found in the individual software packages. This includes the ability to model open/close decisions; FTL vs. LTL, multi-modal, mixed-pallet shipments; apply variable cost per shipment and fixed cost per trip; transfer pricing; VAT and other taxes; and hundreds of other constraints. And, with a visual solution and integration possibilities with external BI tools such as Tableau

River Logic’s flagship prescriptive analytics product Enterprise Optimizer® (EO) is one such program. EO is a modeling platform that can be used to solve for a large range of common business problems, including many logistics decisions. Various models and applications that have been successfully built and used in the last few years include:

    • Truck deliveries (route scheduling) for a large fast food chain in the Southwest US
    • Individual container movements for a shipping line throughout the Pacific Ocean
    • Truck movement of sludge yielded from a wastewater treatment plants in the UK
    • Truck loading for a large CPG company in North America
    • Multi-model coal shipping model in the UK
    • And many more…

This list doesn’t even begin to touch on the hundreds of models with a logistics component — but not the primary focus — of the business problem to be solved for.  Maybe logistics decisions were needed as part of S&OP planning process? Maybe logistics decisions were needed as part of a capital budgeting exercise?  Maybe it was a merger opportunity with a competitor? It could be any reason to add a logistics element to any existing or new model.

A Network Planning Example

As proof, in my next blog entry, I’m going to explain the model in the screenshots below.

The first screenshot displays a hypothetical distribution network for a large, multi-national corporation looking to consolidate freight movement into certain DCs geographically. Although I made up all the data, it is a working logistics model that only took me about a week to construct.

logistics-modeling-distribution-network

There are several supply chain network software packages on the market today that have a similar-looking user interface (UI) with similar-looking capabilities. Here's where the similarities typically stop, though: the second screenshot represents all the underlying financial flows based on the distribution network in the first screenshot.  These financial flows are driver-based (i.e., data driven) that are then used to construct a customized P&L, balance sheet, cash flow statement and other financial ratios.

logistics-modeling-distribution-network-and-financials

Yes, both screenshots are part of the same model.  By double-clicking on the "Financials" icon in the top of the first screenshot, it drills down to the financial flows in the second screenshot.  When the model is solved, the solution includes all the process flows as well as financial flows.  This is something that any logistics-targeted package is not going to provide.

More details to come next time…

Consumer Goods Case Study

You May Also Enjoy Reading
Could A Second Wave of COVID-19 Ruin Manufacturing for Good?…(Read More)
The Five Biggest Supply Chain Challenges of 2020…(Read More)
How to Start Using Prescriptive Analytics: Your 12-Step Guide…(Read More)
Topics

Posts by Topic

    Supply Chain Brief