Helping Hershey enjoy the sweet taste of success

Making people happy since 1894, the Hershey Company is one of the largest and most loved chocolate manufacturers in the world. 

Copy of Untitled Design (615 x 646 px) (1)
pattern pattern patterns
Site - Client Logo Carouse

What was Hershey’s challenge?

The Hershey Category Management team has always needed to process data from loads of retail data suppliers – each with a strict set of NDA allocations. Their team were up to speed with traditional data analytics, but the Azure Big Data toolset was unfamiliar ground for them.

To stay ahead of changing trends and the competition, Hershey had set their sights on Commercial Data Storage (CDS) to speed up the process of delivering insights to executives and data analytics to their field sales team.

Hershey started looking for the right technology partner to support their transformation project. Based on recommendations from third-party vendors, they turned to Advancing Analytics to build the foundation of their CDS platform based on Databricks technology. 

They wanted a framework-driven data architecture that could be maintained easily, while still managing within these strict information security boundaries.


What was our our solution?

Using our Hydr8 platform to deliver a best-practice Delta Lakehouse environment, we put the whole project on fast forward, Hydr8 soon enabled all users to take advantage of newly centralised data. 

  • After a short discovery phase using our tried and tested Analytics Platform process, we designed a platform that would support the development, deployment and management of Machine Learning models in production. 

  • Our approach was deployed on Microsoft’s Azure Cloud Platform. This platform takes advantage of democratising aspects of Azure Machine Learning to support both the casual and the power user. 

    Even an experienced data team faces a steep learning curve when moving from traditional Business Intelligence to a scaled cloud Lakehouse. Hershey chose our Applied Azure courses to help accelerate their team's learning.

  • Azure Databricks is used in conjunction with AML for scaled-out batch and real-time models. The platform is continuously deployed using Azure DevOps, to allow the Data Scientist to focus on model development. Models are managed in production using a combination of MLFlow and the Rendezvous approach to Machine Learning model deployment. 

What were the outcomes?

After replacing all their old spreadsheets with fully automated dashboards, Hershey automatically receives complete, accurate data feeds from their largest retail customers. The C-suite at Hershey is now able to embrace analytics and liaise with the data team to uncover insights that will keep Hershey ahead of the chasing pack.

At the start of every week, the senior leadership team receives an executive summary report in a Power BI dashboard. From there, they can quickly get a detailed overview of business performance. Executives used to receive 14 different spreadsheets with 20 tabs per spreadsheet. Now, they just use filters in the dashboard to bring up whatever insight they need. 

bottom-right-pattern-2 bottom-right-pattern-2

Hershey suddenly has access to newly available and game-changing data.
Their data team has been upskilled to make the most of a modern platform.
Data is now secured and managed in a way the technical team can own.
Our ongoing collaboration with Hershey has been the secret to speeding up a successful outcome.

What did Hershey think?


“By working with Advancing Analytics, we not only connected with the new retailer platform but also provided our business team with more value than they had had before. Until that point, we had been at a standstill with data with our largest retailer. The CDS platform, built on Databricks technology, sets us up for the future and has been transformational for our company. We’ve arrived at the next level of how to model data, serve it up and make it useable for the business team.”

- Jordan Donmoyer,

Manager, Customer Data Solutions, Hershey