Forrester this week released a report on the demand-side platform (DSP) landscape amid the roiling waters of ad tech. The firm sought to assess and score 11 of the most prominent DSPs: Adform, Adobe, AOL, AppNexus, DataXu, Google, MediaMath, Rocket Fuel, The Trade Desk, Turn, and Viant. Overall, it didn't find much differentiation among DSPs and perhaps, that's a reason for customers to pare vendors.
There have been major ripples in the landscape as M&A activity heats up—for example, Turn was acquired by Singel (Amobee), The Trade Desk is now a public company, and AppNexus has filed paperwork signaling its intent to go public. And that’s just the tip of the spear.
Forrester’s 2017 DSP Wave ranking cited six DSPs as “leaders”: MediaMath, DataXu, Trade Desk, Turn, Adobe, and Rocket Fuel. Interestingly, Google, AppNexus, and AOL, three giants, were acknowledged in the middle as “leaders” and “strong performers.” Time Inc.-owned Viant and Adelphic were cited as “contenders."
In order to qualify for a Forrester DSP Wave ranking, platforms are required to have a $100 million run rate (revenue) in their main markets, employ more than 300 people, and return bids in 100 milliseconds or less 98% of the time.
RTBlog asked Richard Joyce, Forrester senior analyst, for more insight on the findings:
RTBlog: In the report, you say there are challenges for authenticated customer data where omnichannel platforms are concerned. How do ad-tech vendors, each with various strengths and weaknesses, address these challenges?
Richard Joyce: Authenticated customer data is a challenging issue. We think it is fundamental to omnichannel marketing. This type of data allows vendors to do omnichannel more effectively— tailoring messages to known users, understanding channel and device interplay, and [providing] better probabilistic modeling for targeting and measurement.
We have seen vendors do two things: Pool this type of data in co-ops, and go acquire or license this data. There isn't any vendor that would disagree that having more of this type of data can make its platform more powerful.
As far as non-standard system access, the obvious reason that this was included was because omnichannel assumes access to all paid channels.
Organizationally, a platform that makes planning, buying, and management of paid channels within a single platform, means more collaboration and holistic digital advertising. In our analysis, it was specific to platforms like Facebook, Instagram, Pinterest, Snapchat, etc. -- accessible via ad API.
In reality, as we predict marketers will focus much more heavily on quality over quantity in the future, access to traditional systems that are still plagued by poor quality inventory, like open exchanges, won't be as important.
Also, "standard" marketplaces are evolving. Header bidding is one way that is changing, but also increased interest in buy types like programmatic guaranteed and private marketplaces are shaping the future of these platforms.
And, as we see channels like TV become more programmatic, it’s pretty clear that this type of inventory won't be purchased the same way that display and online video has been purchased on the open exchange.
RTBlog: The sector is consolidating quickly. What’s your take?
Joyce: Yes, we will see [more] "merging". Even after we kicked the evaluation off, we saw evidence of this in Amobee's acquisition of Turn. This topic wasn’t part of the evaluation, but you can see the holes that Amobee fills for omnichannel.
Ideally, omnichannel requires solid identity management and recognition across devices (which I’m not saying must be deterministic or probabilistic), access to all paid channels, and machine intelligence that can optimize paid brand experiences, keeping channels agnostic and informing media and creative with customer context:device, location, content, "moment," etc.
Efficiency will always be an underlying aspect of programmatic, but there needs to be a transition to understanding effectiveness.
RTBlog: The findings, overall, indicate there's not much difference between the DSPs. If this is the case, how can decision-makers evaluate which ones to work with? Which are providing real value, etc.?
There's a huge pall over the sector in terms of adding real value and problems with transparency around costs, fees, and performance.
Joyce: It's really unfortunate that transparency is still a major concern for the majority of marketers. But it's partially their fault as well. Not in terms of education, because over the past few years marketers have gotten super-smart about programmatic.
But from an incentive standpoint, marketers are still about scale and efficiency. In a model where these are factors of success, there is a lot of room for opaque buying and pricing practices.
But all of the current offerings and features aside, marketers make decisions that are specific to their business, and they should treat ad tech as a partner. So evaluating these vendors should be answering two questions: "Does the tech, and the team surrounding the tech, help me accomplish what I need to?" and "Does the tech and the team surrounding the tech, make me and my team better?"
True partnerships will eliminate transparency issues around costs, fees, and performance. But only changes in incentives will address issues of transparency around media and data.
RTBlog: You mention that DataXu offers "marketer measurement problems outside of digital advertising, which sets them apart from competitors." Is this, in part, an argument for open source systems/tech? Should ad tech work toward more open source tech? Why or why not? Is this a guarantee of strength?
Joyce: I don't like to advocate for one or the other, because in reality, marketers use both for their own reasons. But advertising is a marketing strategy. So, I do believe that these types of systems need to work together to help marketers solve their problems.