Go Big or Go Home

iStock 000052203088 Double 56ec1061c3f14

In today’s bigger is better world, where Supersize meals and 28-ounce Big Gulp sodas reign as king, it’s an easy leap for business leaders to think that the more data a supply chain has at its disposal, the more efficient and profitable the entire company will be. But data on its own is just a meaningless string of ones and zeros. The successful company is not the one with the most data but the one that makes the best use of the data it has.

That being said, it’s time for supply chain managers to harness the power of big data, if they haven’t begun already. In a recent look at data growth trends, market intelligence firm IDC projects by 2020 the digital data universe will reach 44 zettabytes (44 trillion gigabytes). Clearly it’s time for companies to go big or go home.

“Supply chains tend to be very data rich environments. Big data cannot only tell you where you are at now but potentially where you are going to be at in the future,” stresses Karin Bursa, executive vice president of Logility, a provider of supply chain management software solutions.

Big data represents an untapped gold mine of supply chain information that when leveraged well can help supply chains predict and prevent problems. Big data enables supply chains to analyze customer demand, transportation costs, facility data (including operating costs and capacity), raw materials suppliers, manufacturing locations, outsourcing and more.

But before companies can swim in the big data sea, they need to dip their toes in the water to gain a solid understanding of what big data really is and the ways it can be leveraged to the supply chain’s advantage.

Define the Data

A simple Google search using the words “big data” delivers a sea of results that in and of themselves might classify as big data, depending on the definition being used. In reality, big data means different things to different people, and it drives a bit of fear and uncertainty among those tasked with analyzing and applying it to business decisions. This uneasiness often translates into a failure to fully utilize the data a company has, notes Jeff Karrenbauer, president of Insight, a provider of decision support tools for supply chains.

But how much data is considered big?

Karrenbauer calls big data one of the biggest buzzwords in business today. And he says, if you ask people what it means, no one seems to know. “They’ll say it means a lot of it [data],” he says. “But there is no widespread agreement.”

According to Sundar Kamakshisundaram, global vice president of Products and Innovation for SAP Ariba, a firm offering software and information technology solutions for supply chain management, big data is a broad term that can and should be broken down into two primary categories: structured and unstructured data.

Structured data refers to any data that resides in a fixed field within a record or file. This includes data contained in relational databases and spreadsheets. Unstructured data is a generic label used to describe data that is not contained in a database or some other type of data structure. According to a study by independent research firm Aberdeen Group, upward of 83 percent of the data companies have is unstructured. It stands to reason then that companies able to harness this data can gain unique insights they may not access any other way.

Kamakshisundaram cites the following example to illustrate the differences between both. Structured data might be a company’s sales data for the year, which can then be broken down by salesperson, region and by product. This data can be used to create a pretty picture that shows the areas where the company is doing well, the products that are selling well, and which salespeople are making the most sales. But, with unstructured data, the company can dig a little deeper and examine the characteristics of the salespeople who are performing well, for example, so they can hire more people like them. “You have to look at personality information, proficiency tests, social media and so on,” he says.

Texts, tweets, blog posts, videos, images and more can serve up data that supply chains can use to their advantage. Consumers already do this; they tap into personal social media spaces such as Facebook and Twitter or peruse Amazon reviews to make smart buying choices.

The Centers for Disease Control and Prevention also provides a live example of how unstructured data may be used. The organization employs a system to comb social media and other online outlets for references to specific diseases, for instance the Zika virus. They rely on these spaces to help track outbreaks, identify the public’s information needs, and to make predictive decisions.

“Predictive analysis requires the use of data that is not part of the normal database,” Kamakshisundaram says. “In the supply chain we’re used to looking at transactional data, operational data and strategic data, but how do you combine this information with extraneous data like weather forecasts and trends? And then use all of it to make predictive decisions?”

Kamakshisundaram states companies must move beyond traditional data processing technology and add business intelligence tools to maximize data use. A tool that aggregates fundamental information about a supplier, then adds in less readily available data, such as liens, lawsuits and working capital information as well as intelligence derived from Twitter, GlassDoor or LinkedIn, can greatly aid decision making.

“Imagine what a benefit this would be if you were dealing with a supplier from another country, and didn’t have a lot of information on them,” he says. “Technology with sophisticated algorithms to crunch the data and predictive analysis tools, can help you make informed decisions and predictions. It’s almost like looking at a crystal ball.”

Boost the Benefits

 Combining gigantic streams of data with analytical tools provides the potential for supply chain innovation. It can improve forecasting accuracy, uncover demands and trends, identify risk, and more.

Analytics gives companies insight into broad trends. “Am I continuing the way I expected?” asks Bursa. It also looks at specific outcomes. “In other words when the trend and aggregate is not on plan, it helps me dive down in more detail, identify where the problems exist, and begin resolving the problems I have,” she adds.

Wayne Caccamo, chief marketing officer for Resilinc, a provider of supply chain risk and resilience intelligence and analytics solutions, provides the following example: If the new reports predict a major storm heading for a specific area of Taiwan, this intelligence would immediately connect back to which specific supplier sites are located in the region of forecasted impact, and which parts originate from there, connect it to the inventory levels and shipments and predict which products might have a raw material or component shortfall in the coming weeks.  This intelligence is a powerful tool allowing practitioners to act decisively and surgically to make sure supply of these materials are secured beforehand. “Connecting semi-structured data such as risk profile of a location and parts mapped to that location, to structured data such as bill of material and revenue, allows practitioners to predict revenue impact of losing a manufacturing site.  This is powerful intelligence that can allow executives to define which suppliers or sites are truly critical that their teams should spend time focusing on.”

By connecting structured data such as demand forecast to semi or unstructured data such as supplier capacity availability or utilization, a company can also manage capacity with suppliers proactively in order to support an extra surge in demand. “When it’s done proactively, the company can plan for scenarios where demand is unexpectedly different from forecast, and identify and address problems long before they manifest,” he says. “Time is money in this case.  The faster a company can predict a potential looming problem, the more time they have to work collaboratively with the supplier, or tweak the levers within their control to influence a favorable outcome.  Information applied in this way can convert a potential crisis into an opportunity for growing revenue, stealing market share from the competition and wield the supply chain as a competitive weapon.” 

Big data and predictive analytics can also impact Sales & Operations Planning (S&OP). This planning brings together a forecast and inventory policies, inventory and service goals, and supply production distribution. “By bringing lots of data together, you can look at multiple future scenarios and compare those,” says Bursa. “You’re not only looking at volume metrics but financial terms, budgetary information, sales plans, topline revenue and margin contributions, so that you can view the plan from many different angles.”

It can also help company’s respond quickly to unexpected events. For instance, if a supplier lies in a region at risk for typhoons, they can preplan their response should a natural disaster occur. “No matter how proactive you are, you cannot anticipate every kind of event or where it will take place,” Caccamo says. But by having a response in mind, companies can operate without disruption when disaster strikes.

Remember the ROI

Bursa once received a call from a customer needing help considering a proposal from an existing partner, who wanted to earn more of their business. The partner had offered a really good deal for them to source goods from Asia instead of Latin America.

“They used predictive data analysis to determine what would happen if they did this,” she says. “On the surface, it looked less expensive, 25 percent less than their current partner,” she says. “But they found they would have had six weeks more lead time, meaning they’d have to carry six weeks more inventory. Transport costs were also higher.”

The company ultimately decided against this option, and was able to make their decision within two hours through a detailed analysis. Bursa said the customer told her the decision would have taken them two weeks in spreadsheet form. She then asked the company about their confidence in the decision; they said had they done the assessment with spreadsheets their confidence level would have been 50 percent but with the software tool and predictive analysis, their confidence hit 90 percent.

Karrenbauer also suggests predictive analyses could help companies avoid outsourcing mistakes. “Outsourcing mistakes have occurred because people looked at it in insolation,” he says. When a company looks at manufacturing alone, and outsources to take advantage of $1 an hour labor, he says they are not looking at the big picture. Yes, the labor might be $1 an hour, but what are the costs of moving goods from an Asian country to the United States?

Push Past the Holdups

In “Deconstructing Supply Chain Analytics,” Gartner reports companies that do a better job of predicting future demand can trim 20 to 30 percent out of their inventory, while increasing fill rate by 3 to 7 percent. These kinds of results, according to the report, can deliver margin improvements of 1 to 2 percentage points.

Unfortunately, Karrenbauer says there is a battle of wills between humans and algorithms that is preventing companies from getting the biggest bang for their buck when it comes to big data.

“The prize keeps going up if you can convince people to look at the whole thing,” he says. “But because of internal politics it never goes anywhere.”

He explains the siloed structure of business units within a company makes it difficult to optimize supply chains. He adds that companies must thoroughly examine whether their objectives align with the incentives given to individual units and individual employees.

In one egregious example, Karrenbauer performed an indepth analysis for a vice president of distribution. The research found this VP could make changes that helped the distribution team hit their goals but also revealed a second set of changes that could deliver huge returns for the entire corporation. However, this measure would have pushed down the distribution unit’s results. “The VPs advice was to bury the second option,” he says. “He told me, ‘I’m going to make my budget.’ That is an example of silo management.”

Big data can deliver big results for those companies where all business units are united for the greater good. “You’re not supply chain people if you don’t look at the whole thing—you’re just borrowing the term,” he stresses. “When you have VPs for every function with separate financial objectives and bonuses tied to those objectives, then you have fiefdoms and people jealously guarding their fiefdoms.”