The State of Supply Chain Control Towers: Insights from Indago Supply Chain Research Community

In June 2023, we asked members of our Indago supply chain research community — who are all supply chain and logistics executives from manufacturing, retail, and distribution companies — if they are using control towers in their supply chain operations. Surprisingly, given how long this technology has been available, only 40% of the survey respondents said they have implemented one or more control towers, and another 16% said that they plan to implement one in the next 12 months. Why is this percentage so low? What are the obstacles to adoption? And how does data quality fit into this discussion? Those are the key questions I discussed with Pervinder Johar, Chief Executive Officer at Blume Global, on a recent episode of Talking Logistics

A Fragmented Approach

Most of our Indago member respondents who have implemented a control tower are using multiple ones, each one focused on different operations. I asked Pervinder why this is a common approach. Pervinder comments that it is a result of supply chain complexity and the development of the SCOR model years ago that broke down supply chains by Plan, Source, Make, and Deliver (and later Reverse was added). This created organizational silos along these disciplines that had their own processes and technology. And this continues today with different functional groups having their own control towers. This has been invigorated by software companies developing their own control tower solutions for their areas of expertise. “There is a lot of unlearning that needs to be done,” says Pervinder.

The problem with this fragmented approach is that the organizational and technology silos, with their associated metrics and goals, tend to optimize their own little corner of the supply chain but may actually de-optimize other parts of the supply chain and the end-to-end process.

Pervinder also points out that control tower fragmentation often comes about because of the cycles of rising and falling demand and supply. He gives an example in this short clip:

During the pandemic, for example, the rapid rise in e-commerce created a sharp focus on transportation and home delivery. Now there is a renewed focus on demand and supply chain risk. “What is needed is a comprehensive approach that covers supply and demand across supply chain cycles,” he says.

Decision Support vs. Decision Making

We asked our Indago members if the control towers they are using are mostly “decision-support” or “decision-making” solutions. Currently, 80% of the respondents use their control towers for decision-support, with only 20% saying their control towers make automated decisions on what actions to take and execute them. I asked Pervinder if he  believes control towers will become more “decision-making” moving forward.

Pervinder notes that control towers started 20 years ago as dashboards that provided users with visibility to data to help them make better decisions. That continues today for many companies. He mentions that the emergence of generative AI technology and applications such as ChatGPT have led to C-Suite discussions about AI strategies and may be the tipping point for control towers to transition from being decision-support systems to becoming decision-making ones. While the technology may not have been available for this type of automation previously, it is available today.

Dealing with Imperfect Data

One of the biggest inhibitors to supply chain technology success has always been the quality of the data coming into the systems. The survey shows this problem was both a reason some companies have not implemented control towers yet and an issue for those who had. I asked Pervinder for his thoughts on this.

Pervinder states that gathering data and data quality issues have always been an issue, but that new technology can actually generate high quality data. 

“No matter what you do, there are always going to be data quality issues or companies who won’t share their data,” he says. “What you need is technology that can deal with imperfect data and technology that can generate data such as sensors and machine learning. Also, data that isn’t used will quickly get out of date. Rather than trying to resolve data quality issues up front, therefore, use technology that will generate data and correct data over time to improve results.”

The Major Takeaways

The Indago survey covered several other topics related to control towers, such the availability and use of simulation and predictive/prescriptive capabilities. I encourage you to watch the full episode for Pervinder’s insights and advice on that topic and more, and also download the Indago research report for additional insights and commentary. Then post a comment and share your perspective and experience on this topic.