Helena Cho, Director of B2B Marketing at Expedia Group, speaks with Diena Lee Mann, Founder and CEO of Spectio, in a video interview about her professional trajectory, her role in building prescriptive analytics toolkits, business challenges to make data-driven decisions, and delivering a successful data product.
Helena, could you tell us a bit about your career as a data and analytics leader?
Absolutely, I'd love to share about my journey in the travel industry, which I find incredibly rewarding. It’s been a journey of constant learning and collaboration, surrounded by some of the brightest minds in the industry.
My career began at Vrbo’s marketing analytics team, where we covered both sides of the marketplace - helping families and groups find perfect whole-home accommodations. This role provided me with a comprehensive understanding of consumer behavior and business-to-business interactions, setting a strong foundation for my future roles.
As I progressed, I had the opportunity to work across all travel brands at Expedia Group, like Expedia.com and Hotels.com, and later shifted towards the B2B side of the business. This shift broadened my perspective, allowing me to analyze and provide recommendations across various supply partners including lodging, air travel, and car rentals, as well as our B2B operations like Media Solutions.
Throughout my career, my teams and I have developed tools and systems that moved from simple descriptive analytics to advanced prescriptive analytics. For instance, we built a self-serve data product, crafted a marketing technology infrastructure, and engaged in sophisticated customer analytics such as cohort analysis and customer lifetime value projections. These efforts were designed to optimize our marketing channels and involved close collaboration with our machine learning team to create a Media Mix model.
With all of the technologies and talent that are out there in the field of data and analytics, why is it that businesses often struggle to make data-driven decisions with confidence?
The challenge of making confident data-driven decisions in business often comes down to several key issues, the first being the expectation of speed versus the reality of analytics processes. Decision-makers frequently need insights quickly to respond to market changes or internal demands. However, when they turn to their analytics teams, they often find that even seemingly simple queries require significant time to address—sometimes several days. This delay typically stems from the need to prioritize among multiple ongoing requests, or it might mean additional hours for the analytics team as they try to juggle new with existing projects. This misalignment between the speed of business needs and the pace at which data can be processed leads to frustration on both sides.
Another substantial hurdle is the lack of robust, user-friendly self-serve data platforms that allow leaders to access and manipulate data independently. While many organizations have started to develop these tools, they often don't offer the flexibility to easily pull the necessary dimensions and metrics specific to immediate business questions. Additionally, there's a common misconception that once a data product is built, it no longer requires attention. In reality, like any product facing customers, these data tools need continual iteration, improvements, and maintenance to address bugs and adapt to new requirements. This ongoing development is crucial to ensure they remain effective and can truly support data-driven decision-making with confidence.
Data as a product is a principle that organizations often struggle to put to practice. How have you developed the people and processes necessary to successfully deliver data products?
Let’s start with the processes: the crucial components that the data product needs include:
- Data loop: any part that needs to be implemented to the product/user experience without human decision should be automated
- Able to manipulate data to test end user's hypothesis - fixed format requires constant update requests (but also how do you support the endless permutation of metrics and dimensions?)
- Alert system that notifies user when abnormal result is observed
Now the people, first, you need clear decision making ownership, likely the executive who will benefit most from this data product. Next, do they really care about this data product? Based on your answer, your next best action will be different. Because if they realize that they are the decision maker and they care, you have the inertia that brings resources and timely decisions to ship the product. If they don’t care, it may not be a priority and may now is not the right time. How strongly do you feel about bringing this data product to fruition? You get to decide - you push for it vs. wait for another time.
Once you’ve got definitive decision make who cares about the project, then when building a data product for an end user, you need following skills: front-end UX design, back-end ETL build, and understanding the business model and priorities
Thank you, Helena, for joining me today. We need more leaders like you who stay close to the business, and are experts at building effective data and analytics organizations.
