From Exploration to Activation: How HARMAN Makes Smart Decisions In Real Time
HARMAN is all about a people-first experience. Whether at home, in the car, or in the office, everyone reading this has been touched by a HARMAN product or service at some point. You may have a HARMAN digital cockpit in your car, a HARMAN Kardon home theater in your living room, or JBL earbuds for your daily commute. Touring musicians use our sound and lighting equipment in concert, and many of your favorite albums and singles were recorded with our pro audio gear. And many enterprises take advantage of our digital transformation solutions to achieve better business outcomes using emerging tech such as generative AI, cloud, and advanced analytics.
Founded in 1953, HARMAN International today is a $10 billion company with 34,000 employees. An independent subsidiary of Samsung, the company has three main businesses: Automotive, Lifestyle Audio, and our Digital Transformation Solutions group.
Our goal is to make life easier and connect people to what they love, wherever they are. We do that through optimizing production and getting our products and services into the hands of our customers faster. And the best way to do that is by making the most of our data.
We Spent Too Much Time Ingesting Data
Our data analytics team is a smaller organization within the HARMAN IT group. While Samsung drives our mass data governance and analytics strategies, we manage analytics for the entire company, including ingestion, maintenance, and governance. Our team received official status around 2018, and our goal is to enable and empower internal users to utilize data and analytics to advance our business goals.
I joined HARMAN in 2019 and am currently our Senior Director, Data and Analytics. When I started working here, various departments were using QlikView, but they were developing applications in silos. Individual teams were building apps that were specific to their operations, and we didn’t scale them across the company.
We also used various analytics tools, including Salesforce, Tableau, and Einstein, and there wasn’t a single analytics standard to ensure continuity and consistency across departments. The development of analytics apps in silos meant business users spent a lot of time finding, ingesting, and validating the data going into these tools. I would say users spent 80% of their time ingesting data and 20% analyzing it. To get business benefits, we wanted to flip that ratio.
My predecessor had already decided to transition to a single instance of Qlik Sense, creating an opportunity to enact an enterprise-wide platform and consolidate data analytics usage and development. So, we formed a centralized data and analytics team to affect the transition. Qlik Sense was the perfect solution. It was scalable, and we could leverage our previous QlikView knowledge to make the most out of the platform. We didn’t build a better mousetrap; we reset the one we had—this time with better bait.
Using Qlik for the Right Things
We don’t use Qlik for everything—we use it for the right things. We start by asking a lot of questions. What is the sources of this data? Should we bring it into Qlik, or does it make sense to use the internal analytics functions of the platform or system of origin? We also ask whether developing a Qlik app aligns with HARMAN overall business strategy and what business benefits are expected in the short and long term.
Some of our use cases include sales, inventory, and supply chain analyses. We leverage Qlik across our operations, from sales to legal and finance, and it is having a tremendous impact. Of course, to reap the most benefit, we ingest data that is granular enough to provide full visibility into our operations. When you can drill down to the transaction level, it increases trust in the data and, in our experience, allows us to generate more meaningful insights.
Profitability is a business benefit that we often influence, and one of the best use cases that we have created was a financial application. The profitability app was used by one of our business units to turn itself around, going from being in the red to becoming profitable within a year. Our data and analytics team focused on developing an app that made profitability data readily available, generating insights that helped turn that business unit around faster. They went from monthly reports to daily analysis, helping identify errors in the system as well as opportunities for growth.
A Qlik App for Supply Chain Management
Our supply chain app has also been especially impactful. In addition to pandemic-related shortages of critical components such as computer chips, labor shortages and manufacturing capacity issues have also impacted our automotive, lifestyle, and services divisions. We need real-time supply chain data to ensure we can provide products to our end users, be they consumers purchasing Bluetooth speakers or car manufacturers needing touch displays for onboard entertainment and navigation systems.
To move quickly and give us options when we encounter supply chain constraints, we developed our Demand Cube app in Qlik. It provides full visibility into our operations and parts supply and helps establish communication channels between all the involved parties earlier in the procurement process.
If we plan production for the next two years and match it against our suppliers’ capacity to deliver components, we can weather shortages by changing product specifications and sourcing from alternative vendors. At the same time, it helps our suppliers know how much of their production they must allocate to meet our requirements. This level of clarity and visibility helps us understand component changes and suppliers’ capacities, so we can successfully negotiate the procurement of needed parts at the best possible price, especially in crisis situations.
We Enable and Empower Citizen Developers
Our data and analytics team strives to develop pertinent applications for the entire company, but we also seek to enable and empower citizen developers. We have given 300 people outside our team access to training and Qlik so they can develop applications for their respective business units. We provide access to secure and reliable data sets, control permissions at the server level, and have strict policies about moving their applications into production.
Once their applications are approved and put into production, the data and analytics team takes over and maintains them. This pre-and post-production governance process helps ensure that citizen-developed apps comply with our guidelines and standards, access reliable data sets, and are fully supported. We’ve taken enablement further than most companies because we can trust our people to securely develop the apps they need to use our data. We have significantly increased the potential reward by reducing the risk of granting them Qlik access.
We’re doing more because we only manipulate data once and securely share it across use cases instead of having individual users and business users ingesting it separately, needlessly duplicating labor and increasing the potential for errors.
Piloting a ChatGPT Integration for Deeper Insight
We saw ChatGPT as an enabling technology that we could harness to further accelerate data to insight. With that in mind, we did a pilot embedding GPT functionality into Qlik dashboards. The first solution uses VizLib’s library of third-party visualizations.
The idea being, since the data is already in Qlik, we should be able to use Generative Chat to answer unstructured queries about the data. This was so successful that we developed a private integration with generative AI, which we're piloting for broader use. This generative capability has really gotten the conversation around AI/ML moving at a faster pace. Previously, people would ask, "What is AI/ML?" Now, the question is, "When and how can we use AI/ML?" It's definitely an area of growth that we are acting on and investigating, especially now with Qlik’s Auto ML codeless automation, which allows us to run experiments and probe deeper, extracting further value from the company’s data sets.
We’re also working closely with our Digital Transformation Solutions group to integrate generative AI and Qlik, pushing the platform’s boundaries by moving beyond descriptive analytics and empowering people to gather contextual insights from our data using natural language questions—even if they’re not data scientists. Instead of simply visualizing data, they can ask what it might tell us about the impact on our operations six months or a year down the line.
After we ask the questions, ChatGPT does the heavy lifting. We route natural language queries and our data set to ChatGPT through the OpenAI API and then receive an answer. The process is transparent to end users because the front end is a Qlik dashboard, and they don’t have to write SQL. We are making major strides ad are putting solutions into the market in Q4 2023.
Qlik Has Made Us Adaptable
Analytics can power large-scale changes at any organization, but don’t try to boil the ocean by trying to solve every problem at once. Ask your analytics team to tackle the specific business cases with the greatest possible benefit. Invest energy in the applications yielding the strongest ROI and work with an analytics enthusiast to channel the energy to learn. The development that citizen developers do, if governed, can yield incredible results. This approach will allow leaders to simultaneously address enterprise-wide issues and business unit needs.
Ultimately, embracing analytics leads to greater adaptability. Whether managing your supply chain, optimizing production, or maximizing profitability, you can’t afford to be reactive. Qlik has given HARMAN visibility into our operations and those of our suppliers. We can see what’s happening now and what’s coming down the line, so we can strategize and plan more effectively.