The term digital transformation means different things to different people. But one thing it means to a growing set of companies is running proof of concepts or pilots and learning from them. Historically, companies looking at a technology solution would put together a detailed business plan, and then only implement the technology if the forecasted financial returns exceeded their return on investment (ROI) target. But more companies are experimenting with digital technologies regardless of whether the ROI is there or not. As Diana Salguero – the IT Director for the Americas at Vopak – said in a presentation at ARC’s Forum about a solution they worked with, “a proof of concept never fails, you just learn from it.” In their case, the proof of concept of an IoT/energy planning solution in one department of the terminal failed, but they learned enough to go ahead and successfully pilot the solution across the terminal.
There are some technologies that are relatively easy to pilot. Some types of autonomous mobile robots (AMRs) can be deployed without integration to warehouse management systems. Rather, call buttons are used to summon the bots. Integration, of course, drives up costs. Further, a company can work with just one or two AMRs and see what happens. This is a low-cost experiment.
But in the supply chain planning (SCP) market, production pilots require integration to enterprise resource management or various supply chain systems, cleaning up dirty data, and the configuration efforts of an implementation team. This kind of pilot costs much more. For this kind of pilot, it is much harder to take the attitude that “Hey, if it works great. If it does not, oh well, I’ve learned something.” Many companies pilot SCP solutions, but they do no pilot them with the attitude that a failed pilot is acceptable because they have learned something.
This was a topic of conversation between me and Nil Durak, Chief Operating Officer, and Omer Bakkalbasi, Chief Innovation Officer, of Solvoyo, during a recent briefing. Solvoyo is an innovative provider of end-to-end planning, visibility, and analytics solutions. In their case, they are beginning to get their solution in front of digital transformation teams. And to underscore the spirit of digitalization, they call their pilots “test-and-learn.”
How realistic is this? They have just completed a test-and-learn with a very large consumer goods company. This pilot includes an end-to-end supply chain solution designed to generate concurrent (a plan across production, fulfillment, transportation, etc. that is generated centrally and executed locally), low touch plans in which, 85 plus percent of the time, the model-generated plan is executed as-is. In short, this is a project with a big scope, and it is consequently a big effort. Yet unlike traditional solutions, it did not cost the consumer goods company millions to pilot this and it went live in 6 months. This pilot is generating real plans that help to optimize this company’s business.
Now compare that to the potential savings. When I threw out the number “tens of millions” in potential savings, Ms. Durak and Mr. Bakkalbasi agreed with the number. And I did not throw that number out arbitrarily. For global multinationals pursuing a supply chain transformation, that is a conservative number. A recent benchmarking study shows that multi-echelon inventory optimization reduces a company’s inventory investment by 13 percent on average without sacrificing service. Furthermore, our internal ROI research shows transportation optimization generates freight savings on average of 8 percent of the freight spend, and so forth, across the full set of supply chain applications. With big companies, those savings end up in the tens or even hundreds of millions of dollars for implementations with a broad scope.
However, a test-and-learn should go both ways. A company can fund a pilot and demand that their existing rules for how the company conducts its business be respected. But a true digital test-and-learn would look to see what the unchained solution could generate in savings if it was not fettered by respecting constraints that reflect the traditional ways of working. In the case of this pilot with the consumer goods company, the constraints were to fulfill orders to the most important customer first, second most important customer next, and so forth until you get to much less important customers. It was only with these less important customers that the optimization engine could finally be unleashed to drive savings. In this case, Solvoyo estimates that the scope for optimization is limited to only 1 to 4 percent of the search space if the constraints were not removed. The company may be leaving a lot of savings on the table by continuing to use traditional rules for allocation. Optimization can be used to relax rules while not sacrificing service commitments for any high priority orders.
In marketing there is a well-known phrase stating that “half the money I spend on advertising is wasted; the trouble is I don’t know which half.” For smart companies, that has never been true. It is possible to run a test ad in one small region of the country and see what the lift is.
The same thing is true in supply chain management. For a multi-national company, it is possible to relax the constraints, see how service is impacted, and do it in just one country for one small set of products.
In conclusion, when it comes to digital transformations, there should be experimentation with both the technologies and the way those technologies are used. It can be difficult to take this approach when it impacts the way a company is doing business. But it is possible. And it is likely that this consumer goods company has a champion that will be willing to do these experiments.