According to IBM, we generate 2.5 quintillion bytes of data every day. By 2025, it is estimated that 463 exabytes of data will be created each day globally. These are mind blowing numbers and a lot of investment is required to process and store this data – the investment is worth it because, when harnessed correctly, this data can revolutionize industries. At Freestar, we process billions of data points every month.
This data is often referred to as “big data”. It is extremely large and complex data sets that are difficult to process using traditional data processing techniques. These data sets may include structured, semi-structured, or unstructured data, and can come from a variety of sources, such as social media, sensors, transactional data, and more.
Analyzing big data can reveal patterns, trends, and insights that would be difficult to discern using traditional data analysis methods. The insights gained from big data can be used to inform business decisions, improve customer experiences, and drive innovation in a variety of industries.
Today, I’m sharing with you how Freestar is using this technology and how we’re applying it to help our publishers maximize their ad revenue.
Let’s dive in.
How Freestar is Leveraging This Technology
Big data capabilities have been accelerated by the introduction of cloud computing. Cloud computing gives access to a variety of technologies that help organizations capture, store, process, analyze, and derive insights from massive amounts of data.
- Cloud Computing – Freestar has the advantage of being “cloud native” – we started life in the cloud, so have been able to harness its power from day 1. Cloud computing allows Freestar to process and analyze large volumes of data and extract value from it. The value we receive can tell what to optimize more, how to operate more efficiently and really where to focus our efforts. It gives our publishers a huge advantage because we’re able to enhance our ad operations which in turn, improves their bottom line.
- Machine Learning – Machine learning is a subfield of artificial intelligence that uses statistical techniques to enable computer systems to learn from data and make predictions or decisions. Our aim, at Freestar, is to personalize every ad request. We can accomplish that by using machine learning. Due to the volume of ad requests and the layers of data we have, it’s not something that can easily be done without this technology. Machine learning let’s us ask “what the most optimal way of serving this ad to this user?” We’re able to analyze data on a granular level to make ad experiences as customized and optimized as possible.
In this industry, you often come across the same questions. For example, “how do we optimize ads for iOS user in the United States?” That’s a relatively straightforward question where you take the answer and apply it to all iOS users in the United States. What big data and machine learning allows us to do is to split iOS users in the United States into many more facets (e.g. usage, time of day, etc.), so we can make far more granular decisions in real time and serve users an ad request catered to specifically to them. Using big data and machine learning, that straightforward question would have a different answer for every ad request.
Applying This Technology Today
Big data provides us with a wealth of information and insights that can be used to drive better decision-making, improve efficiency and productivity, enhance the customer experience, and gain a competitive advantage in the marketplace. Today, Freestar is applying cloud computing and machine learning to two ways:
- Dynamic Flooring – Our flooring technology utilizes artificial intelligence and machine learning to calculate the best performing floor price by website, ad-unit, geo, device type, day and hour for each impression delivered to maximize the ad request eCPM. We developed our dynamic flooring technology to help publishers more intelligently set floors in a first-price, header bidding world.
Freestar has a dedicated team of data scientists to analyze and monitor all the data points. Our dynamic flooring has been tried and tested on all types of inventory and it understands known and unknown inventory. The AI is smart enough to be able to predict the floor prices for new inventory it’s never seen before, based on all the inventory it’s seen in the past. - Ideal Ad Stack – An ideal ad stack varies depending on the needs of the publisher, but at Freestar, our ad stack has a goal of constant optimization. We’re always adding new dimensions to see how we can improve publishers’ ad revenues and personalize the ad experience. We look at every facet to ensure we’re optimizing each ad request – looking at timeouts, testing and running demand client-side or server-side, and considering turning specific demand off, etc.
We also look to keep our ad stack flexible to adapt to changes in the industry and constantly optimize to the most up-to-date technology and features for our publishers.
Key Takeaways
- Freestar processes billions of data points every month and leverages cloud computing and machine learning technologies to extract value from big data and personalize ad experiences for publishers.
- Cloud computing allows Freestar to process and analyze large volumes of data, while machine learning enables the personalization of ad requests by analyzing granular data on a per-user basis.
- Freestar applies big data technologies to dynamic flooring and an ideal ad stack. Dynamic flooring technology utilizes artificial intelligence and machine learning to calculate the best performing floor price for each impression delivered to maximize ad request eCPM. The ideal ad stack is constantly optimized by looking at every facet to ensure optimization of each ad request and adapting to changes in the industry.
Overall, big data can help businesses operate more efficiently, effectively, and profitably. That’s what Freestar is doing for our publishers. If you’re a publisher looking to leverage this technology, get started with maximizing your ad revenue today!
