Most DSPs Are Marketing Fluff


fluff_smallADOTAS – They promise advertisers greater performance and increased returns for their online advertising campaigns. They claim to solve the problems associated with the traditional display media buying model through advanced analytics and intelligent real-time decisions.

They are demand side platforms (DSPs) and they have next-generation media buyers excited. Unfortunately, most DSPs are woefully limited in several important respects and most simply can’t deliver what they claim to.

In principle DSPs are designed to be a state of the art platform that give advertisers greater control over pricing, targeting, and managing their online media campaigns. DSPS give clients access to real-time bidding, advanced data acquisition, ad serving, reporting and analytics. The idea is that a single system can connect into multiple ad exchanges and can be accessed from a single users interface.

The potential benefits are instantly apparent: transparency, reach and optimisation across multiple media sources, data integration and deep analytics.

Much of the excitement around DSPs is due to media buyers seeing them as the vehicle for taking ad networks out of the equation and replicating their services at a lower cost. DSPs can achieve this.

But what many marketers do not appreciate is that though DSPs are doing something quite exciting — most are not doing it very well. Put simply, DSPs lack the advanced technology to derive the best results from complex raw data, they disregard the individuality of advertisers with a one-size-fits-all approach and they neglect the importance of creative customization when targeting users.

Most claim to have machine learning algorithms, where the algorithm evolves based on data. In order for a machine to learn and deliver optimal decisions, it requires scale of data. How many DSPs have the traction in order to have real scale of data? Some do, but very few.

Quite simply, the technology and engineering driving most DSPs is not advanced enough to process this information and translate this into the best, smartest real-time bidding decisions for media buyers. Most DSPs are not equipped to make the best and most profitable decisions on behalf of advertisers as they lack the scale of data and the intelligence behind those decisions are second-rate.

Without the ability to make the most profitable decisions for advertisers, the resulting ROI will be poor. If an advertiser doesn’t benefit from improved performance through a DSP, they might as well go back to a legacy ad server and buy media the old fashioned way.

2011 is the year when the tide goes out and we will all see who has been swimming in the display ecosystem with their trunks off. Those with cutting edge technology and truly defensible IP will swim. The rest will sink.


  1. “How many DSPs have the traction in order to have real scale of data? Some do, but very few.”

    Then why not just partner with the handful of DSP’s who do have the “real scale of data”?

    Off the top of my head, I can name two.

    Since the inventory available across each DSP is pretty much a wash, then (at least based on the logic of the article) centralizing your ad-exchange buys with one of these more advanced DSP’s should be deliver on this ‘promise of greater performance and increased returns.’

    Takeaway: Partner with a DSP that is integrated with an advanced Data Management Platform.

  2. I think you’ve missed the point. A DSP is a platform you operate not a product you buy. The core, expensive technology is cookie matching and the capability to bid in real time. Many companies including CPC remarketing firms are reliant on DSP technology to power their business but it’s how they use it that is the real magic not the infrastructure behind it.

  3. @James what is that, a shameless plug for Turn, the company that failed at being an ad-network, failed at being a dsp, and will fail at having a data management offering? Lets get real – these are completely individual stacks, and nobody wants to lock their data into a quasi ad-network/dsp that already engages in shady data practices. Lets get real.

  4. Well, I did say I can think of two, so I guess my post would also be a shameless plug for…?

    And I do agree that the “real magic” is how one uses this tool, although even that may be subject to change.

    It’d be interesting to have this same convo. come Q4 2011.

  5. Even the best performing technology is still a pretty suspect tool for advertising in terms of real ROI. Display is just comfortable for buyers and visible for screen shots, but hardly as effective as any pub/network/dsp will admit. It’s a volume game, hoping to find that target to click.
    DSP is definitely a buzzword and there are certainly some ‘marketing only’ companies out there. But in a highly questionable display world populated by unproven concepts (verification, aggregation, tracking and more) is one company in any position to claim to be ‘better’ than another? It’s like winning an argument at a pub, not losing an argument at a Mensa gathering.

  6. There seems to be a lot of assumptions being thrown around here. Has anyone here actually used a DSP. I can see no backup or “data” put forward here towards your arguments.
    Lets face it, a DSP / ad exchange provides more transparency (or presumably) for the advertiser this is already a positive thing. While these systems are still in there infancy they will quickly become an essential part of media spend. Real-time bidding allows a publisher to get the best rates for his inventory and gets the advertiser the inventory he wants for the price he wants – so everyone wins!


Please enter your comment!
Please enter your name here