graphic with photo of woman overlayed with text that reads Life in Adtech Senior Data Scientist

Life in Adtech: How Models and Algorithms Help Drive Advertising Performance

In our blog series “Life in Adtech” we’re showcasing the different ways in which StackAdapters are driving innovation in adtech, and creating a progressive tech culture. From using modeling and data to create advanced algorithms to building stylish and robust user interfaces, everyone at StackAdapt is contributing to evolving the StackAdapt platform and providing the best product and service to our users.

This month features Zeynep Akkalyoncu, a Senior Data Scientist at StackAdapt. In this post, Zeynep shares her journey into adtech, what her role entails, fun facts about data science, and how data science brings value to StackAdapt clients.

How did your journey into data science and adtech begin?

I’ve always been mathematically minded, but also fascinated by the humanities, especially linguistics and anthropology. As I began studying computer science, I found myself more and more drawn to data science because I liked how the field marries these two seemingly disparate interests. During my Master’s program at the University of Waterloo, I was fortunate enough to get a taste of data science in academic research, but I wanted to continue to apply these concepts to real-life problems. 

Joining StackAdapt as a Data Scientist gave me the chance to do just that by applying data science to the world of adtech. The learning curve was steep in the beginning because I was completely new to the field, but the endless support and patience from my team and other departments helped me to learn the ropes fast! 

What is the main focus of your role in data science?

Data science includes all the invisible (but indispensable!) machine learning processes that are behind the various features that we offer on the StackAdapt platform, such as page context AI campaigns and automatic optimization of bid prices. My goal as a Data Scientist is to build high-performing machine learning solutions in collaboration with Data Engineers who ensure the efficient flow of enormous amounts of data with robust ETL pipelines. I work on new machine learning algorithms and I maintain old ones, running historical backtesting to assess their value and writing code to deploy these models in production to serve our clients.

Data science works closely with different teams across StackAdapt to make sure we’re providing a great experience for our clients. Since we rely on large computing clusters and data infrastructure, we collaborate closely with our Engineering team. We also frequently work with the client-facing teams to understand how to better support the needs of our platform users. Our goal is to turn often ambiguous client or industry needs into StackAdapt platform features that are supported by machine learning.

What do you think is the biggest value that data science brings to StackAdapt’s clients?

Data science at StackAdapt builds on our client’s understanding of their audience, and the message of their campaign to automatically employ the most effective strategies for them. This requires a data-driven approach throughout the entire pipeline from pre-campaign planning and campaign performance, to post-campaign analysis. 

For example, StackAdapt makes it possible for clients to gauge the effectiveness of their strategies before running their campaigns, all with tools developed by the Data Science team. Much of the campaign execution, including bidding and pacing, is also supported by numerous machine learning services to ensure great targeting and placement of ads. Last but not least, data science provides insights on the campaign after it has ended to ensure that the client can run even more effective campaigns in the future.

How does developing models and algorithms help to maximize ROI and drive performance for StackAdapt clients?

The performance of StackAdapt campaigns is supported by a complicated network of machine learning services, so the individual effectiveness of underlying models and algorithms directly impacts campaign ROI. This is why the entire Data Science team strives to ensure the quality of both the respective services that we are working on and their interaction with other services. 

For example, bid prices are dynamically computed based on a sophisticated machine learning algorithm to acquire the best placement for an ad, which may be dependent on the ideal context as determined by our contextual algorithms. By incorporating domain knowledge and technical expertise, we work to build insightful and effective services that best cater to the individual needs of each client.

How are models, algorithms and AI creating innovation in the adtech industry? 

Applying the latest technological advancements in AI to the StackAdapt platform allows us to create a great experience for our clients and end-users. The models and algorithms that we develop as a team can capture information that could get lost in a traditional advertising setting. These models and algorithms process hyperdimensional data that often transcends the practical understanding of humans at a speed that would be impractical for us to replicate in most cases. 

Data science opens up endless possibilities for innovation in adtech. I think the most telling sign of an innovative adtech platform is the user-friendly functionalities of a platform that work with minimal interference. Sophisticated technical work behind the scenes means that the end user gets to have a seamless and magical experience. 

What is a fun fact about data science that other people might not know?

Given that most data science work happens behind the scenes, it can be easy to forget the human element of the field. While we do build sophisticated mathematical models and strive to implement efficient algorithms, the human experience is always the focal point of our work. Our goal is always to capture trends in human behaviour, produce insightful analysis about our habits, and then build products that enhance the human experience.

What is the most rewarding part of working on the Data Science team at StackAdapt? 

The data science team at StackAdapt has a culture of ownership that I really admire. We are encouraged to explore, experiment and pursue potentially high-impact projects that don’t have a definite roadmap. Many of our projects are born out of ideas brought up by members of the team during brainstorming sessions. Our work is guided by common values that we share, so that everyone can see the direct impact of their contributions toward the quality of the service that we provide and the experience of our clients. 

It’s all the more rewarding and enjoyable to work when you feel a personal stake in the performance of the company. I believe this works well at StackAdapt because we have a flat organizational structure and an open and collaborative environment. Thanks to this, I’m able to learn a lot from my teammates as we go through design documents, code reviews and experimental results together.

Interested in joining StackAdapt? See our open roles and apply, here.

In each new addition, we will be highlighting our teams and team members in our “Life In Adtech” series—stay tuned for more!

You may also like: