Want to know whose demand-side platforms are tops? Forrester Research has a few suggestions.
The marketing and technology research firm’s Forrester Wave division has reviewed the top general-use DSPs in the crowded programmatic market and identified nine it believes are top performers.
Rated on their technical offering and business strategy, the nine are grouped into three categories: leaders, strong performers and contenders. Among the leaders: DataXu, The Trade Desk, AOL, Turn and AppNexus. In strong performers fall RocketFuel, MediaMath, and Google. And on the cusp between that group and contenders lies Audience Science.
Although the names on the list will be familiar to industry watchers, their respective rankings were a bit unexpected. Not a lot of buyers would predict that AppNexus’s bidder would rank above a dedicated demand-side player like MediaMath, for example.
Forrester initially looked at 22 platforms, but screened out most for not working across enough channels (minimum 10), not having enough active clients (minimum 300) or not being able to support enough scale (minimum 1 million queries per second). Notably absent from the list were major players like IgnitionOne, The Rubicon Project, and Criteo.
DataXu, a platform that doesn’t get a whole lot of PR, scored by far the highest in technical offering, while The Trade Desk (which has focused heavily on working with agency traders and for the most part ignored brands that are taking programmatic in-house) scored the highest on service strategy.
Another surprise was that Google’s DoubleClick Bid Manager — one of the most pervasive DSPs on the market — scored the lowest on technical capabilities, in part because it doesn’t yet have a functioning data management platform.
What the market leaders do differently
Forrester had a lot of love for DataXu, a smaller tech player with around 300 global employees, thanks to its data analysis chops and its commitment to becoming more than just a DSP.
DataXu was co-founded by a literal rocket scientist, Willard Simmons, who worked on planning a manned Mars mission during his PhD research at MIT. Simmons developed an algorithmic engine that could crunch a large number of variables, like how much water and fuel the crew would consume and how much their food waste would weigh.
Post-MIT, the algorithms Simmons developed became the core of DataXu’s media buying platform, Active Analytics. The system uses machine learning and predictive analytics to test ad investment patterns against projected and measured outcomes, and optimize media buying at both the tactical and strategy level. In its report, Forrester credited Active Analytics for DataXu’s high ranking, writing that by “trying to solve marketer challenges that go beyond digital media buying, DataXu has set itself apart as a more complete solution for marketers.”
Ed Montes, DataXu’s CRO, told Marketing that the company’s ambition is to be a lot more than a DSP. It’s currently working on extensive media mix modeling that bridges digital and conventional media channels, grinding its way through all of a brand’s advertising and sales data to figure out how much to spend on every channel, geographic region, and bidding strategy. Although Montes said the company’s not interested in automating buying beyond digital media, it can provide detailed recommendations on how budgets should be allocated to TV, print and OOH as well as search, display, email and all the rest.
“If you think about the broad spectrum of data that can be ingested by such a learning system, and the uses it could have, it could be developed to do things like price analysis of particular product against a particular geography,” Montes said. “As opposed to simply, ‘how much should you pay for an impression to get a better direct response metric.'”
Branching out of the DSP space seemed to be a theme among Forrester’s top performers. AOL and Turn both have their own natively integrated DMPs and analytics tools, and either already have (in Turn’s case) or are developing (in AOL’s) a digital marketing hub that brings together cross-channel planning, execution and reporting tools for buyers. MediaMath was similarly credited for its forward-thinking DMP, the only one of the pack to offer tagless data collection powered by Akamai.
Google was penalized for not having a native DMP in-market, though it has announced that it will launch one later this year. It also lost points for its dependence on other products within the DoubleClick suite, making it difficult to use as a standalone solution.
Forrester frowns on partner integrations
Several platforms lost points for relying on partner integrations to supply certain features, rather than developing their own native solutions. AppNexus and Turn were credited for having accessible APIs and building up ecosystems of natively integrated partners, covering everything from custom targeting features to fraud detection. But they were also criticized for not having native support to address many of these issues.
The report’s authors reasoned that too much reliance on external partners leaves the platform at the mercy of the market. Although an open and transparent ecosystem is desirable, the report said, relying on outsiders for core functions like data management and attribution won’t be sustainable, because it’s an added expense for the marketer and because the platform can’t guarantee consistent service from its partners. Forrester gave the example of Google’s attribution provider, Adometry, which decided to pull its integration with Turn and other third-party platforms, leaving them out in the cold.
Speaking to Marketing, Turn senior vice-president of marketing Paul Alfieri pushed back against that thinking, arguing that partner integrations are an essential component of providing clients with transparency and control. Turn has focused on integrating partner vendors, he said, because it wants clients to have the maximum range of choice and customization opportunities.
That’s especially true when it comes to third-party measurement and attribution, he said. “Marketers want the flexibility to choose, for example, a viewability provider. … At Clean Ads IO last week, there were lots of quotes from marketers saying that third party viewability and verification data is what we need.”
He added that the idea behind an open ecosystem is that the market will fill any gap in demand. Pointing back to the Adometry example, he said that Turn was able to find another attribution provider, Pixalate, before Google closed the doors.
