All DSPs are Not Created Equal

[Watch the video below (or here) for a deep dive on this issue with Bruce and Turn’s Ethan Lubka]

There have been rumblings in the industry recently that demand-side platforms (DSPs) are basically commodities. The argument is that though they are costly and complicated to create, once viable, the primary value of such platforms lies in their ability to access exchanges.

That is a bit like saying that all cars are commodities in the sense that they get you from Point A to Point B. However, we all know there’s a difference between a Fiat and a Ferrari.

Likewise, when paired with a robust data management platform, a DSP is part of what I call the “What Works Machine,” a mechanism that solves the biggest problems that marketers have. The What Works Machine combines data and targeting and helps marketers market better.

The Algorithms Really Matter

The commoditization argument assumes that all DSPs are more or less alike. If we go back to 1999, there was a similar charge related to search engines. Of course, Google was much better than AltaVista and Lycos. They built a better mousetrap, one that was based on its PageRank algorithm. The Google machine learned from the user clicks that landed on the search results it spat out — this machine got smarter, better, faster over time and the rest is history.

Having the best technologists who build better algorithms really matters. At Turn, we’re technologists and we believe deeply in this notion: that better computer science can yield an algorithm that’s better. We can build a better mousetrap, and machine learning is at its core. Today, the best DSPs also employ machine learning.

The reality is that every DSP is very different. Every DSP is built from different bones on different tech stacks with different things in mind. And some are powerful enough to process a very high number of queries per second (QPS).

That’s an important factor because if your QPS is higher, that means you’re responding to as many bid requests and bid opportunities as possible to ensure that marketers get the right price and the right metrics.

To create a DSP that’s able to do all that you need a large, talented engineering team, solid tech and great leadership. That’s a lot harder to execute than some might assume, as Facebook discovered with its Atlas platform. As Facebook acknowledged, its homegrown DSP often delivered ads to bots rather than humans.

Some marketers, assuming that all DSPs are on equal footing, go for the one with the lowest pricing, not realizing that those savings come at the cost of missing key metrics and going without the support of top-notch industry people.

Turning On the What Works Machine

French Statesman Georges Clemenceau once said, “War is too serious a matter to be left to the military.” In the ad world, the parallel is that ad technology decisions are too important to be left solely to anyone outside of the brand: the onus is on marketers to investigate the technology on their own. In particular, marketers need to understand the infrastructure, architecture and decisioning time frames that their computing hardware calls for.

Ideally, the CMOs of today are tech savvy enough to see the benefits of one technical solution versus another. Certainly the CMOs of tomorrow will need to be tech savvy. The differences can be stark. In January, Google’s machine learning system beat the world’s top player of the game of Go, an Eastern-based test of strategy and intuition that had defeated every other machine learning system. The most striking thing about that victory was that the world’s best (human) Go player, Fan Hui, remarked that “I’ve never seen a human play this move. So beautiful.” Machines in this sense can be teachers to even the deepest experts.

We are getting to a similar point with marketing – when machines can make moves and decisions that humans would never imagine. That is actually going to empower marketers and make them smarter rather than take anything away from them.

The complexity of such solutions are really unfathomable. A data management platform allows you to organize data, a DSP lets you activate media against that data and a data-mining platform gives you insights from that data. When marketers are fully deployed across all three, there’s a virtuous circle in which these all inform each other and the system improves itself.

That is the conceit of a What Works Machine, that the marketer lets the machine loose to apply machine learning to process data and make decisions in real-time in ways that humans simply cannot. This may sound audacious, but the top technologists in the industry will be able to build it. Like the search algorithm that Google built, it’s a machine that learns and gets smarter as it gains experience (consumes data). In 2016, tools are more than just tools, and if you believe in the power of technology, choosing the right DSP really matters.

About Amobee

Founded in 2005, Amobee is an advertising platform that understands how people consume content. Our goal is to optimize outcomes for advertisers and media companies, while providing a better consumer experience. Through our platform, we help customers further their audience development, optimize their cross channel performance across all TV, connected TV, and digital media, and drive new customer growth through detailed analytics and reporting. Amobee is a wholly owned subsidiary of Tremor International, a collection of brands built to unite creativity, data and technology across the open internet.

If you’re curious to learn more, watch the on-demand demo or take a deep dive into our Research & Insights section where you can find recent webinars on-demand, media plan insights & activation templates, and more data-driven content. If you’re ready to take the next step into a sustainable, consumer-first advertising future, contact us today.