Advances in Supply Chain Data Analytics and Management

Line Art

Never mind that it costs money, nor that it up-ends familiar business practices to update supply chain data analytics. Organizations worldwide continue to replace manual, spreadsheet-based forecasting processes with integrated planning systems. Unlike people-populated spreadsheet models, cloud planning applications integrate directly with data sources and apply advanced algorithms to mine and measure copious abstracts. The result is a revelation of trends, patterns and other features of interest that people on their own can simply not divine. This is the realm of advanced analytics.

It’s an exciting revolution, and becoming more so as advances in cloud connectivity and big data accessibility expand our known (and perhaps unknown) universe. But herein is a lesson. Before you go out and spend big money chasing the data-driven dream, take careful stock of what you want to accomplish, the data you already have available, and how much of it you’re currently not utilizing.

You may find there is no provable need to spend $1 million on a massive two-year ERP implementation. Rather, with the right people on board, and a solid understanding of existing data, you’re quite likely to generate far greater financial returns by first defining your desired processes, and then investing in best-of-breed statistical forecasting — way less time and money.ff

In their book, “Mining Your Own Business,” researchers Jeff Deal and Gerhard Pilcher make a cogent case for understanding the predictive power of data, and (given the right tools) for taking all the steps necessary to capitalize on data that is already available, whether in-house or from publicly available sources. They also stress the dire need to build consensus and commitment before any serious supply chain data analytics undertaking.

In their own words: “… pursuing a data analytics initiative without proper planning and organizational buy-in is like purchasing an expensive piece of home exercise equipment without sufficient commitment. The equipment may seem exciting at first, but without a dedicated regimen, it will soon end up sitting idle in the basement or serving as a clothes rack in the corner of the basement.”

Needles in Haystacks? or Hidden in Plain Sight

Again, look at what is available. Chapter 1 of “Mining Your Own Business” begins with a story about the astonishment of an Assistant Inspector General (AIG) at a U.S. federal agency. The AIG was floored when he learned that by tapping just one public data source (the Federal Procurement Data System), Deal and Gerhard were able to quite accurately predict which of his agency’s contractors ran the highest risks of waste, fraud, and abuse.

Turns out, several of the contractors the co-authors identified at the high end of the risk spectrum were already under active investigation by the agency. How did they do it? Creatively, Deal and Pilcher wandered through available data, and thought about how it could be used to project risk. Once they homed in on the right data source, they applied the mining and processing power of advanced analytics to reveal the needles in the haystacks, in this case the relatively shady actors among a noisy population of relatively legit.

First Homes, Then Bicycles

The discovery made by Deal and Pilcher dovetails with an example of our own, at least in so far as the proper use of available data is concerned.

About a decade ago, a top mass-market U.S. bicycle maker came to Vanguard Software with the aim of improving product demand forecasts, supply planning, and operating margins. The maker had apparently been struggling with grossly inaccurate sales forecasts, and thus costly inefficiency in everything that followed: production scheduling, resource allocation, stocking levels, you name it.

The maker had been relying on somewhat on internal sales history to guesstimate demand, using spreadsheets to aggregate gut forecasts from sales team heads. But they hadn’t updated this process to an enterprise grade approach, one that would derive forecast inputs statistically (among other things). Nor had they tapped some very available sources of data, such as new and existing home sales, tracked by multiple public and private organizations.

Through careful analysis, and the application of advanced modeling and simulation techniques, the team at Vanguard was able to show a compelling correlation (albeit with a lag) between home purchases and bicycle sales within specific geographic regions. The data essentially provided an additional regression model that improved bike demand forecast accuracy (the same relationship might have applied to swing sets).

Start Small, Yield Big

So, if you’re convinced that supply chain data analytics, together with advanced modeling and simulation techniques, will improve forecast accuracy and enterprise planning, you’re right. But remember not to put the cart before the horse. If you want to become a data-driven organization, start with the fundamentals. Assemble the right people to discuss the costs and benefits of implementing a new system, process, and culture. Discuss current pain points, possible solutions, key performance objectives, and competing pathways to optimum net benefits for the enterprise as a whole.

From there, take you’re your cross-functional team findings to the top. Management buy-in is essential to developing a culture of commitment to data-driven decision making. Do this before dropping serious coin on any major implementation. You might make greater financial gain with a much smaller investment in a forecast-focused cloud plugin.