On this episode of Wonderful Work, we chat about intelligent automation and data culture with Jereme Pineda, Senior Product Manager for data optimization at Zalando. Jereme has enjoyed a long and diverse career in analytics and data and has worked for famous companies such as Basware, Sievo, and Microsoft.
How Jereme started his career in technology
Before graduating, Jereme worked in the National Statistics Office in the Philippines and was responsible for handling birth certificate records. "I started as a data entry specialist at a time when there was no machine learning, so I had to do it manually. After graduating, I went to a company called America Online. They outsourced their technical support operations to the Philippines, and I was one of the first team leaders there."
Fast forward to 2008, he migrated to Singapore and got an excellent job as an analyst for Nokia, which triggered his move to Finland. Jereme worked with 32 million contacts per year, giving him a lot of data to analyze. Later, he moved to the UK to join Nokia's British team before ultimately settling in Finland. "If I didn't come here for Nokia, I would've come here for heavy metal music."
The golden rule for businesses working with data maturity
"You need to have a very clear use case or problem you want to solve with data," says Jereme. Many organizations jump on the bandwagon and unload all their data from legacy systems. They invest in a nice data platform and hire engineers and data scientists, only to find they've got nothing to show for it.
"Firstly, you must start with a problem that you understand how to solve and understand how data can solve it. Secondly, leaders must understand the unique advantages of data and what that means for the organization. What’s more, having good quality data is key to success, as well as a knack for data.”
Understanding the difference between "garbage" data and high-quality data sets
"Bad data comes from the source. That bad source could be processes you're capturing digitally or events your applications are generating," Jereme explains.
According to Jereme, many enterprises use applications that were built without analytics or data as a use case. "When you design those systems, they spit out data that won't be optimal for data analysis and can't create machine learning datasets.
Right now, many companies are going through the ordeal of unloading data and only see it after it's pumped into a data warehouse."
Cleaning up dirty data
To solve this issue, teams must look at their systems and really make data a "first-class citizen" in engineering teams. "To transform as an organization, teams really should ensure their applications produce excellent data sets," Jereme explains.
The data needs of enterprise businesses
Jereme has been responsible for designing and implementing many analytics platforms, which have been used by thousands of global organizations. As the designer and architect of analytics, he knows all about enterprises and their data needs.
"Enterprise businesses operate in a paradigm," Says Jereme. "100 employees means working with 1000 daily decisions. Empowering those people with data and insights has a compounding effect on organizational transformation."
"In the past, many companies made a decision based on a gut feeling," Jereme continues. “Often, they don't have enough data or insights to aid their decisions."
However, things are changing, and many companies are adopting a data-driven mindset.
A data-driven future
An increasing number of companies now embrace a data-driven approach. "People are using very simple technologies like spreadsheet applications, which is a start," Jereme explains.
This is a big step in terms of looking forward and embracing the idea that a company should transform and use data as a fundamental workflow tool.
Rather than raw data, companies should use processed data that is safe for human consumption. "Creating a strategy around data and thinking about what kind of decisions have the most impact is the best approach, especially in the context of analyses," says Jereme.
Why Excel is still king of the analytics world
Back in the '90s, Jereme saw Excel as a magical tool. "Me and some of my friends put our ages into a Windows 95 Excel sheet and visualized the data as a bar chart," Jereme remembers fondly.
"Most of us, including the generation right now. have our first experience with data through a spreadsheet application. We built some nice analytics dashboards with spreadsheet applications."
While interviewing customers, Jereme found that people still wanted to extract data the old-fashioned way, despite having impressive new dashboards to work with. "People seem more comfortable analyzing data with familiar tools," says Jereme. "I'm in a high-tech company, but I am always mindful that people get the most from working with simple things. What's more, spreadsheets are not stagnating. Many are now infused with new AI technologies."
What enterprise leaders should know about data engineering
"Enterprise leaders need to extract data and put it in a repository efficiently," Jereme explains. Doing this makes data useful for data scientists to analyze.
The cornerstone of a strong data culture
The definition of a strong data culture varies. "start-up companies that start in the cloud tend to have a very strong data culture. However, the situation is slightly different for enterprise companies," Jereme explains.
"A strong data culture comes from hiring leaders with a knack for data and a proven track record in using data. During the interview stage, ask candidates to describe the data they've used in the previous role. If they can't answer the question, then search for other candidates".
How companies can use data for innovation
"Companies must understand the value of data," Jereme explains. It's important to produce data that supports analysis or even create machine learning applications to work hand in hand with analysts and data scientists."
"When your data is in bad shape, you'll need to clean it with advanced machine learning, which is an ineffective approach. You need to fix data at the source. Starting with quality data helps you achieve better results faster."
According to Jereme, data science or artificial intelligence projects normally involve cleaning data, which accounts for around 80% of the work. Only around 20% or even 10% is about the writing of the machine learning algorithm."
Jereme's tips for developing a data-driven mindset in a traditional enterprise
"Look at successful use cases of companies that have used data to transform their operations and then emulate them," says Jereme.
To underscore his point, Jereme draws upon his experience at Zalando. "The way they introduce and transform data is phenomenal. They define problems to solve them and prioritize their customers before they think about the technology they will use."
Jereme's final thoughts
"Becoming a data-driven company is not hard. However, the endeavor requires teamwork, excellent leadership, and seamless collaboration. When you want to solve a problem, start by considering your customers and stakeholders," Jereme concludes.
The Wonderful Work podcast
If you'd like to learn more about Jereme and his views on data culture, check out Jereme and Lari's full chat on the Wonderful Work podcast.
