HelloFresh

In a conversation with CIOReview, Ed Chen, the Associate Director of Growth Marketing Analytics at HelloFresh, illustrates on how the implementation of advanced technology can effectively enhance marketing drifts. Ed underscores  imminent predictive analytical trends, including the centralization of resources, provision of care outside the conventional settings, and the vital role of accepting change for superior outcomes. He encourages up-and-coming professionals to welcome innovative methodologies and stay adaptable to drive better results in the analytics sector.

Could you share a brief biography of your journey, including your current roles and responsibilities and the steps you took to attain your current position?

My journey in analytics began during my undergraduate days when I developed a deep interest in econometrics and statistics. Early on in my career, I worked in the insurance industry where I built models to understand risk and retention, advising treasury departments on insurance limits, reinsurance, and how to fund expected losses. Eventually, I pursued my MBA and worked in consulting, always focusing on analytics and helping clients make better decisions with data.

Throughout my analytics journey, I have mainly focused on the advisory role, guiding teams in setting up data-centric organizations and answering questions with data. I also spend some time understanding infrastructure and technology like pack stacks because it is crucial to know where you are to determine what is possible.

Now I work as an associate director in growth marketing analytics at HelloFresh, where I work closely with my business partner to answer all kinds of questions related to our budget for paid media. I advise where to spend, how to buy, and the right tradeoffs between channels to drive performance. It’s an exciting role, and I am thrilled to be able to apply my analytics expertise to drive growth for HelloFresh.

As someone interested in this field for a long time, I’m curious to hear your thoughts on how the analytics world has evolved. What are some prominent trends you’ve observed in the marketplace, and how do you see them impacting the future of analytics?

From my perspective, I would say the field of analytics has come a long way. When I first started, analytics was all about using formulas to calculate trend lines by hand. But with the evolution of computing power, things have gone up a notch Microsoft Excel became the go-to tool for building models, but people soon realized it had its limitations in its capabilities. People started exploring other options, like using Python and building their libraries to solve more complex problems such as forecasting, categorization, text mining, and image mining.

"Even if you have the programming and math knowledge from school, natural curiosity and context are things you can’t just learn. It’s important to stay curious and keep learning to stay ahead in this constantly evolving field"

And then there are the deep neural networks, which allow us to do machine learning that is incredibly accurate, but also incredibly opaque. So, while we can solve more complex problems, we sacrifice transparency in the process. But that is just one direction that analytics has gone in. There has also been a push to make analytics more accessible to the everyday business user with tools like Alteryx and DataRobot. These tools are essentially just GUI interfaces to Python or R-functions, but they allow inexperienced users to drag-and-drop functions into their workflows. While most of them are expensive, they make analytics approachable.

Then there’s reporting, which is at the edge of analytics. It’s not analytics per se, but it’s essential to understand your business performance before you start predicting or optimizing. There has been a lot of progress in this area as well. Tools can now create informative and visually appealing reports, even if the underlying data structure is not great.

All in all, the field of analytics has evolved since I started. There are more tools and techniques than ever before, which means people can solve more complex problems than we ever could have imagined. But with that complexity comes a tradeoff between accuracy and transparency, and it’s up to us to find the right balance for each problem we are going to solve.

Can you elaborate on how the recent changes in the analytics world have affected the enterprise world? What are some significant impacts or influences created due to these changes?

The analytics world has undergone significant changes. Democratization is now a major trend. Learning and utilizing advanced analytics tools and techniques is now more accessible to individuals, regardless of their background. High school students interested in analytics can grasp and apply sophisticated algorithms without necessarily going to college. With tools like DataRobot and Alteryx, even business users without a strong statistics background can run advanced models.

There are, however, challenges that are coming with the democratization trend. For instance, I’ve noticed that some young graduates who have learned many algorithms lack the context of when and why to use them. On the other hand, some business users may not fully comprehend the meaning and implications of the algorithms they use. This can lead to models being taken out of context and specialists spending more time trying to understand the business implications of their work.

It is essential to understand both the model aspects and the business context to mitigate this trend. Business users should work closely with analytics specialists to ensure they know the models they use and can apply them appropriately. A more sophisticated buyer is needed to ensure that the best models succeed and are used effectively in business.

What technology solutions are available and suitable for enterprises and teams to address the challenges of navigating the changes in the analytics world? In your opinion, what is the best way forward for organizations to tackle these challenges?

The best way for organizations to navigate such challenges in the analytics world is to have individuals with a deep understanding of business and mathematical concepts. While chat-based tools like ChatGPT are useful in getting insights, they have not evolved to a point where they can effectively contextualize data. This means human expertise is necessary to interpret and understand data correctly. 

I do believe that in the future, as machine learning models become more sophisticated and accurate, we will see a shift toward greater dependence on AI tools for data interpretation. I anticipate that within the next five to ten years, we will see a significant increase in the availability of tools that can provide the linguistic interpretability of data with accuracy and meaning. Till the time comes, it is imperative for organizations to have individuals on their teams who can effectively bridge the gap between data and business objectives.

Can you describe your leadership strategy and methodology for assisting clients and executing projects with precision and success?

As a data scientist, a significant part of my job involves addressing the queries and concerns of business strategists. When I started, it consumed more than 90 percent of my time, but over the years, I have adopted a strategy that helps me streamline the process. Suppose I receive a question more than once, especially from multiple sources. In that case, I begin setting up reporting systems that enable people to access the necessary insights without relying on me the most time.

Certain questions arise repeatedly, and I make it a point to understand why they matter to the business. This approach allows me to anticipate future needs and develop relevant models in advance so that I'm prepared with answers when the questions do come. My leadership style and my personal approach are rooted in customer-centricity. I believe that the needs of my stakeholders are paramount, and I do whatever it takes to fulfill their immediate and anticipated requirements, even if it means developing new models or exploring different approaches to data analysis.

Can you share your perspective on the future of analytics and the potential disruptions that may occur in the industry? How do you see the field of analytics evolving in the next few years?

One major shift we are seeing is a move from Frequentist to Bayesian models, where we explain things based on probabilistic models rather than deterministic models. This shift is possible due to the improved computing power, enabling us to use more sophisticated models.

Another emerging disruptive technology is chat-based interfaces, like DALL-E and ChatGPT, which can create realistic images based on text and provide plausible responses to queries. Although these tools are not full proof, they are still a significant leap forward from voice assistants. I believe this technology can potentially disrupt search channels as more people use mobile devices to do shopping and search. If analytics could be done through a chat interface, it would be more compatible with mobile devices and could revolutionize how we receive information.

Also, as business executives’ handheld devices, I can see the rise of front-end apps that receive information through mobile channels, with large and sophisticated models running in the background on servers. It will be easier for businesses to do analytics on mobile devices, leading to a more mobile-focused approach to data analysis. I believe these disruptive technologies will continue to shape the analytics world in the future, and we must stay ahead of the curve to stay competitive.

As an ending note, could you please share a piece of advice for aspiring professionals or peers in the field of analytics?

I always advise anyone looking to venture into the analytics field to keep the context in mind. Understanding the context of the problem you are trying to solve is key to success in this field. It can be the difference between being incredibly useful and requiring a lot of help. Even if you have the programming and math knowledge from school, natural curiosity and context are things you can’t just learn. It’s essential to always stay curious and keep learning to stay ahead in this constantly evolving field. Be willing to give back to the community by sharing knowledge and expertise to help you grow as an analytics professional.