overcoming supply chain issues with graph-database tech

By Harry Powell, head of industry solutions at TigerGraph

How loudly do we all scream for ice cream? The average American eats four gallons of it each year, according to the US Census Bureau. In all, ice-cream makers in the United States churn out more than 1.3 billion gallons a year.

While enjoying their double-scoop cone or sundae with sprinkles on a warm summer day, hardly anyone thinks about the behind-the-scenes processes and supply chain that went into producing them. But that effort is multi-layered and complex, and, for one of the world’s largest ice-cream producers, it had become anything but a treat.

To understand why, one must understand a little about vanilla. You may know that vanilla is often ranked as the world’s most popular flavor. You may not know that vanilla is the second most expensive spice, costing more $600 per kilogram (about 2.2 pounds).

The orchids from which commercial vanilla flavoring is derived grow in only a few corners of the globe. Madagascar produces 75%, with the rest coming from Papua New Guinea, India, and Uganda.

The ice-cream producer sources vanilla for its factories around the world from many suppliers in various countries, at different times of the year, sometimes buying through middlemen, sometimes direct. Prices tend to vary widely, not just for the vanilla but also for packaging and transport. Quality also can differ, requiring the company to constantly stay on top of who has the best product at a given time.

So, while a certain brand’s vanilla may seem to the customer like, well, just vanilla, it may be made using a variety of sourced ingredients and formulas.

Trying to manage all this disparate data in its vanilla supply chain was giving the ice-cream maker indigestion. While all of the data was stored in a data lake, each country and business function had its own siloed data sets and schemas. That lack of coordination hampered the company from connecting the data to compare across regions and functions. Instead, it had to collect data abstracts from the multiple data systems of record and then combine them into a single big table using a complex, inflexible, and error-prone process.

The resulting table was huge, difficult to use, and hard for team members to interpret. The entire operation also was too slow—like ice cream, the usefulness of constantly changing data can melt quickly.

By putting a set of blinders on the ice-cream producer’s attempts to get a handle on its vanilla purchasing across regions, these technological hurdles also were a competitive disadvantage, because even a 1% saving in the cost of buying vanilla can shave tens of millions of dollars from the company’s costs.

Then the company decided to change its data-management recipe. It added graph technology, which enables the merging of data silos (like the ice-cream maker had into one enterprise-wide data set).With that connected view, the company could apply graph analytics, which enables complex inquiries in near real time to gain new understandings of the many dependencies in the supply chain.

With this approach, the company could ask questions like: Is there an alternative source of vanilla that has similar characteristics but costs less? How can we mix vanilla from different sources to get a consistent ingredient? Are we being charged different prices in different regions by the same supplier? How will sales be affected if one supplier is unavailable?

The ice-cream producer’s experience vividly demonstrates graph technology’s ability to connect meaningful data from a relationship perspective, at scale and with speed, and unlock critical insights about how people or entities interact with a business.

Because graphs are purpose-built for storing and processing connections, they inherently establish the relationships within data. All an organization needs to do is visualize and navigate through a graph to analyze previously hidden relationships and tackle sophisticated problems with nuance. Graph technology supports both transactions and analytics, so companies can bring massive processing power into the database.

Instead of hunting for facts in sprawling databases, the ice-cream producer can now explore data’s natural shape and see the relationships that are revealed when many data points within databases are simultaneously connected to multiple other data points and analyzed in a more holistic manner.

The ice-cream producer isn’t alone in its embrace of graph. According to Gartner research, graph technologies will be used in 80% of data and analytics innovations, up from 10% in 2021.

Judging by the ice-cream maker’s account, the results of doing so are pretty sweet.