The Importance of Real-Time Data Processing in Combatting Ad Tech Signal Loss

Sponsored Content

Personalization, targeting and retargeting, attribution. In an online landscape where third-party cookies were king, ad tech companies felt pretty confident saying they had these concepts well in hand. 

The problem was, they never really did. Third-party cookies were always just a slice of the identity pie when it came to omnichannel advertising experiences, and an increasingly mangled slice at that. We should all thank Google for the collective panic attack it induced within the ad tech community when it announced its intention to fully deprecate third-party cookies on Chrome. Sure, that intention ended up being about as solid as Grandma’s Jell-O mold after spending a few hours on the Thanksgiving dinner table. But it was the much-needed kick in the pants many ad tech players needed to finally build their solutions using newer data sources and identity graphs.

Third-party cookies might now be sticking around on Chrome, but everybody has already acknowledged that the emperor isn’t wearing any clothes. Third-party cookies don’t cut it in an omnichannel world. They haven’t for a long time. So—moving on.

Here's the rub: If ad tech firms are going to finally let go of deterministic signals like cookies (not to mention IDFA and other increasingly irrelevant signals), they must embrace probabilistic models to accomplish the same tasks. That takes data. TONS of data, coming from everywhere they can get their hands on it. First-party data, second-party data, third-party data—heck, throw in some zero-party data. They need it all. 

Tons of data requires databases that can handle it—and in real time. When you get right down to it, that’s no easy task. The right database partner becomes everything. 

The future leaders of ad tech will be the ones that take their obligation to real-time data processing, at scale, seriously. To understand what that really takes—and what the right partner can bring to the table—we spoke with Daniel Landsman, Head of North America Ad Tech and Martech at Aerospike. Here’s what ad tech leaders need to know about the current state of real-time data. 

Let's kick things off: Why are we talking to Aerospike about real-time data? What makes Aerospike unique in the world of databases and why should ad tech execs care about you?

At Aerospike, we're all about speed, scale, and reliability. You can think of us as the Ferrari of the database world. Our NoSQL database is designed to handle billions of transactions with the lowest latency.

Speed and scale—music to any ad tech leader’s ears. So, what's the secret sauce behind Aerospike's performance?

Our secret lies in our Hybrid Memory Architecture™. (We call it “HMA” for short.) We combine the speed of DRAM with the persistence of flash storage; it’s basically the best of both worlds. 

High level, without getting too technical, we only require indexes to be memory as pointers to the data, and then we keep the data on disk. So, it ends up making infrastructure more affordable at scale when handling large workloads with high throughput and low latency. We also have intelligent clients that shard the data and maximize hardware we run on. 

Ari Paparo often jokes about databases slowing him down. How would Aerospike change his mind?

Ari’s a smart guy, and he likely could figure it out if he went to our documentation page. But, in short, here’s where the rubber hits the road: In ad tech, you have to process high-velocity and high-veracity data in an extremely short time window. That is where Aerospike shines.  

We specialize in high-throughput, low-latency (i.e., sub ms) workloads doing millions of QPS that are mission critical for all the ad tech companies we power. 

Ari would see our technology is well positioned—but in all honesty, he probably already knows since Beeswax is a client. Just look at our long list of customers that have moved to Aerospike for their ad tech workloads, citing our reliability. We often hear, “Aerospike just works.”

Speaking of real-time, why is it such a game-changer in today's ad tech space?

In industries like ad tech, where decisions are made in microseconds, real-time data processing isn't just a luxury—it's a necessity. Whether it's detecting fraud the moment it happens or delivering hyper-personalized content instantly, real-time capabilities are becoming a requirement—not just for ad tech, but for many other industries as well. 

AdTechGod would probably say, "Data is king, but speed is godly." How does Aerospike empower companies to rule in both domains?

Having scale or speed is nice, but having both is better. In ad tech, it’s essential. We operationalize data at speed and scale for our customers, allowing them to make decisions more quickly with higher volumes of data. We do this through the aforementioned technologies. 

Let's get into some real-world applications. Can you share a success story that showcases Aerospike's impact?

Ever heard of a little French company called Criteo? They use Aerospike. In fact, they were so efficient that they were able to decrease their hardware footprint by 75%, from 3,200 servers to 800, shut a data center down, become more carbon efficient, and run over 200M QPS on Aerospike.  

That's impressive. But not every ad tech company runs at Criteo levels. Is adopting a high-performance database like Aerospike a complex and costly endeavor?

Not at all. In fact, moving to Aerospike results in a lower total cost of ownership, and the move is pretty straight-forward. We’ve been doing this for over 14 years now, so we have the process down. We have a team of solution architects who both help do the modeling and even a team that runs the database for you, if that is what you want. Additionally, we have migration packages that make it really simple as well. I've seen it, time and again, where clients move to Aerospike and save tens of millions of dollars against our main competitive set. 

If you had to sum up Aerospike's value proposition in a tweet, what would it be?

“We enable companies to operationalize their data at massive scale faster than any other vendor on the planet. We were purpose-built for high throughput/ultra low latency multi-terabyte workloads with predictable performance and unparalleled total cost of ownership.” #SubMSLatency #CostEfficient #KVStore #SystemOfRecord

Nice hashtag game! Before we wrap up, what's on the horizon for Aerospike? Any exciting developments we should keep an eye out for?

We're continually evolving and supporting many different data models—from our origins as a key-value store, to document data, to graph and vector—all using the same underlying database platform. We continue to invest in cloud-native technologies to offer even more flexibility and scalability. Our focus is on making real-time data processing accessible to businesses of all sizes, enabling them to innovate without limitations.

OK, last question—if you could give one piece of advice to ad tech companies struggling with data challenges, what would it be?

Get better engineers…

But, in all honesty, you need to have a quality infrastructure in place to support your data stack. Additionally, if and where relevant, do things in real time to increase personalization to enrich the customer experience.

Wise words. Thanks for joining us and shedding light on how Aerospike is pushing the boundaries of what's possible with real-time data in ad tech. 

About Aerospike

Aerospike is the real-time database built for infinite scale, speed, and savings. Its customers are ready for what’s next with the lowest latency and the highest throughput data platform. Cloud and AI-forward, Aerospoke empowers leading organizations like Adobe, Airtel, Criteo, DBS Bank, Experian, PayPal, Snap, and Sony Interactive Entertainment. Headquartered in Mountain View, California, its offices include London, Bangalore, and Tel Aviv.

Learn more at Aerospike.com