Create a free Manufacturing.net account to continue

Demand Signals At The Speed Of Light

When it comes to getting the right amount of the right products to market, we need to get the right demand signals to make our critical manufacturing and inventory management decisions. For better or worse, we depend on many aspects of these demands signals: where they come from (our transactional systems, marketing activities, customer research), how complete they are, how accurate they are and how fast we get them.

When it comes to getting the right amount of the right products to market, we need to get the right demand signals to make our critical manufacturing and inventory management decisions. For better or worse, we depend on many aspects of these demands signals: where they come from (our transactional systems, marketing activities, customer research), how complete they are, how accurate they are and how fast we get them.

It seems to me that we always have to sacrifice one dimension for the other. If you want your planning data fast, then you sacrifice completeness or accuracy. If you want your data accurate, then you’re going to get it too late.

What if these signals came at the speed of light? What if you could know ahead of time? What if you could predict what the future would be?

As I see it, there are three important areas for improving the speed and accuracy of demand signals in order to make critical merchandising, inventory, and manufacturing decisions: Using all β€” including the newest β€” information sources; being able to tap into the information; and learning how to predict through analytics.

1. Using All Information Sources

Savvy demand planners are looking to new sources to understand the marketplace. Within the last few years, consumers have been putting their attitudes, preferences and opinions everywhere on the web. E-mail, social networking, blogs, and texting β€” a virtual ocean of demand data is sitting out there for the picking. The marketplace has raised its hand and is telling us what it wants. But can we hear it?

The technology is out there to do this, and it is transformational. I’ve seen the emergence of social and contextual analytics, and its application to supply chain issues is evolving rapidly. The question for supply chain leaders should be β€œShould I be an early mover in this space, or should I let my competitors take the lead?” The answer might not be easy, but supply chain leaders should at least know what is out there and start getting the facts on these new technologies.

2. Tapping into the Information

The second area is accessing the data we already have. In our super-instrumented supply chain, the data we collect is massive and deep. One beleaguered CIO recently confessed to me β€œI have incredible stores of supply chain data. I’m just afraid somebody is going to ask me for it!”

I almost wanted to laugh or cringe, but I know how serious an issue this is for so many companies. This particular client was quite sophisticated in gathering data from the supply chain, having sensors embedded all the way into its corporate customers’ own operations to measure supply and consumption instantaneously, i.e., at the speed of light. The problem was that they could only use the data operationally to do things like place orders. They couldn’t get it into a format where they could analyze it, integrate it into their demand planning and begin getting ahead of their customers in predicting demand.

3. Predicting through Analytics

Lastly, the best supply chain leaders put the new data and the access to it together with predictive analytics. As the name suggests, they predict the future. The instrumentation (e.g., sensors in our supply chain) enables us to get the demand signals at the speed of light. Predictive analytics takes those demand signals and enables us to get ahead of actual consumption, and improves our ability to balance inventory and reduce working capital.

Consider what we are seeing at smart consumer products companies. They start by connecting to retailers to provide instantaneous feeds of consumption information. We've built "demand signal repositories" to accommodate a few of these collaborative partnerships. We've then used these demand signals to develop "freshness algorithms."

For consumer product companies that have a product with a finite shelf life, we've found that they can normally predict within only 30 percent accuracy which items will go stale. We recently completed a project in which we used information to build an artificial intelligence algorithm that reduced this to 9 percent and significantly improved the supply chain from that consumer product company to the retailer.

The challenge to supply chain leaders is to develop a keen and deep understanding of these new capabilities. Only when you understand them, see them in action and define the business value for your organization, can you truly start to leap β€” at the speed of light β€” into the future.

Glenn Steinberg has helped dozens of the world’s largest brands solve their complex supply chain challenges. For more information, visit www-935.ibm.com/services/us/gbs/bus/html/bcs_index.html.

More