Regular readers of Logistics Viewpoints know that Adrian and I are fans of multi-tenant, network-based transportation management systems (aka software-as-a-service TMS). These solutions take the burden of connecting to carriers and cleansing EDI messages off the back of the shipper. They also make it easier for a shipper to add new carriers if they have a problem lane. And shippers can leverage the network to access benchmark analytics in a way that is not possible with a single-tenant solution.
But, ultimately, what matters is ROI.
For most shippers, the biggest payback from a TMS comes from mode selection, LTL-to-TL consolidation, and routing optimization. In large, complex transportation networks, all of the advantages associated with multi-tenant solutions pale in comparison to the savings associated with transportation optimization.
At the JDA FOCUS 2011 conference last week, a presentation by PepsiCo brought this point home for me. What PepsiCo is doing with JDA’s single-tenant TMS solution amazed me.
PepsiCo is running a single instance of the TMS across the Americas to support its food and beverage products. The TMS supports a complex distribution network of over 100 manufacturing locations, 600 warehouses, and 200 suppliers. The network includes inbound moves from growers and suppliers to the plants, outbound moves to distribution centers, repackers, copackers, bins that support vending machines, the company’s direct store delivery operation, and shipments to customer DCs. PepsiCo executes 1.5 million shipments per year via its private fleet, dedicated fleets, and common carriers. The company’s transportation spend exceeds $2.2 billion and all of this is managed from a centralized TMS.
The TMS handles constraints and decision logic based on delivery commitments, product, customer, capacity commitments, preset appointments, preferred carriers by business unit and/or lane, and other factors. PepsiCo generates continuous moves for its private fleet. In some cases the optimization may be regional (e.g., West Coast), in other cases national. In some cases the same truck can be used only for one business unit (e.g., Tropicana), in other cases the same truck can be used to carry goods from different business units (perhaps, Quaker Oats and Lays). Its drop down for optimization encompasses 20 criteria. In short, this was the most complex and holistic transportation optimization problem I have ever come across.
The presenter did not disclose the savings PepsiCo has achieved, but I suspect it is massive. Procter & Gamble, for example, has told financial analysts that better transportation optimization will save it $200 million annually.
Thus, what really struck me is that for a network like PepsiCo’s, if an optimization engine from one TMS vendor performs even one tenth of a percent better than an engine from a different vendor, the difference could be worth millions of dollars in savings. Then it is necessary to tune the engine to get the most out of it. That is probably worth even more.
A couple of weeks ago, the TMS team from Manhattan Associates came in to brief us. Manhattan also prides itself on being able to solve complex optimization problems involving inbound and outbound, private fleets, and common carriers. Manhattan made the point that for big, complex problems, it is willing to customize the optimization engine—i.e., put in special math to solve a peculiar problem – to produce better savings. At the time, I thought customization is usually the wrong way to go. However, after hearing about PepsiCo’s network, I have rethought this. Again, if customization improves the solve even one tenth of a percent, it would be worth it, even if the customization makes the implementation take longer and future upgrades more difficult. Multi-tenant solutions, of course, are never customized.
In conclusion, multi-tenant solutions have a lot going for them. But for companies with very complex networks and large transportation spends, it really is all about “the solve.”
(Note: JDA Software and Manhattan Associates are ARC clients).