The Death Of The Third-Party Ad Server

robgriffinddt“Data-Driven Thinking” is written by members of the media community and contains fresh ideas on the digital revolution in media.

Today’s column is written by Rob Griffin, executive vice president and global head of digital at Havas Media Group.

I have been in the business since late 1995 and have witnessed the evolution of the third-party ad server, which has run its course.

I remember hard coding ads to fixed search results pages and then extrapolating the largest traffic day of the month to report to the client the amount of “hits” they received. I also remember when one advertiser wanted to “third-party ad serve” its banners, my boss at the time said, “Who does that? No way!”

But fighting the future is futile. Technology’s progress is unstoppable, especially when a problem is being solved.

The whole concept of the third-party ad server centered on standardizing tracking for better accountability and optimization. Everyone counted impressions and clicks differently, so agencies and their clients did not want to rely on publisher provided numbers. Advertisers and agencies also wanted the ability to track post-click activity to prove digital’s value. The third-party ad server became our way to prove ourselves and we have all been handicapped, or maybe handcuffed, by that since.

DoubleClick largely won this race while Atlas bled into irrelevance. Since then, some specialty ad servers have emerged, such as MediaMind for rich media and Vindico for video. But now that Google owns DoubleClick and Facebook owns Atlas, we must consider the implications.

Since Facebook blocks DoubleClick, will Google block Atlas? If that happens, will Yahoo, Microsoft and AOL all need their own ad servers to block their competitors’ ad serving solutions? You could argue that they already do with their various demand-side platforms (DSP) and ad network solutions. Additionally, most of us use a collection of fragmented spot solutions – specialty platforms for social, search, content and retargeting.

Let’s be honest: DSPs would never have existed had the aforementioned ad-serving companies recognized the marketplace’s need and desire to tie buying and tracking directly to the inventory. So a handful of smart individuals started their own companies and changed our business forever, such that in the near future, it will all be managed via terminals (just ask AdFin).

Consider what attribution platforms, data-management platforms (DMPs), viewability and fraud technologies offer, plus the fact that most publishers now have standardized counting, and ask yourself: Is third-party ad serving dead? Do I need it? Has it become irrelevant and only still in play out of habit?

If site analytics with tag management enables retargeting, automated buys and dynamic optimizations, and if the DSP-DMP combo provides attribution, reporting, viewability and fraud technology, then what’s the role of the ad server?

The fraud and viewability platforms like Integral Ad Science, Double Verify, WhiteOps and Forensiq all audit publisher ad delivery, which used to be the purview of ad servers.

Is the ad server now just another tech commodity? Nothing stops the DSPs from rolling out their own formal ad-serving solutions. Most agencies already use platforms like CognitiveMatch to dynamically generate ads anyway, thus reducing the ad server to nothing more than the universal cookie by which I drive my attribution modeling.

Yet, why can’t I do that using another technology that’s not pegged to one publisher? If I use Atlas or DoubleClick as my universal tracking code, I am by default forking over data to Facebook and Google. Plus if each of the big houses has its own ad serving, I still need independent solutions to audit the auditor(s).

So I don’t need the ad server to audit my delivery anymore as I have better solutions. I also don’t need the ad server to deliver the ads anymore as I have better ways to do that, too. Then we have the whole possibility of a cookieless world where companies like Tapad are trying to solve for cross-device tracking. Are the ad servers in this conversation? They aren’t.

So I declare the third-party ad server dead. Or at least, a dead man walking.

Follow Havas Media Group (@HavasMedia) and AdExchanger (@adexchanger) on Twitter.

Enjoying this content?

Sign up to be an AdExchanger Member today and get unlimited access to articles like this, plus proprietary data and research, conference discounts, on-demand access to event content, and more!

Join Today!


  1. Noah Omri Levin

    I definitely see how this is possible from the perspective of a display buyer; however, when attempting to solve attribution problems for clients across paid media channels… clients that don’t have a DMP solution and aren’t planning on getting one in the near future… tools like DCM are still highly relevant and useful.

  2. As the former-head of revenue for a large publishers (Cheezburger), I begged the SSP’s to roll out an ad server. There still needs one platform to revenue optimize against all sources of demand for publishers and that to me is the ad server’s primary job. Right now the ad servers like DFP don’t yet do this while SSPs do seemingly a lot except traffic sold-by-the-publisher premium campaigns. I fully expect to see the SSPs take on DFP especially if they want to be full-stack ad-platforms.

    • Sanjeev Rao

      Hi Todd,

      You should consider looking into deal ID. It is a feature typically supported by SSPs and encapsulates the premium sold IO-style deals. Premium programmatic, not just remnant.

  3. Rob describes a fragmented technology landscape and simplifies the role of today’s modern ad server. Things have changed.

    Advertisers and agencies demand that ad servers support multi-channel, that they are dynamic enough to plug into the programmatic stack, with verification and audit capabilities to separate the buy side from the sell side. Bringing together data programmatic and traditional buying models: a crucial central system today where programmatic is still in its relative infancy. Globally, advertisers see a flurry of local market solutions, technologies that at the top of the food chain need a common reporting system, a common de-duplication methodology and delivery system. Dominant advertisers that have long term goals rely on ad servers as technologies that have already proven the test of time. The ad server may have matured but we will need to be blessed with long lives and healthy hearts to see their ultimate deprivation.

  4. Rob, good points all around. No doubt there will be consolidation as market matures…at the same time we are at the “tip of the spear” so to speak. New features and functionality that address particular pain points and changes in technology will continue to create opportunities for start-up niche players (is it a product or is it a feature?).

    The real elephant in this room is Google, who is playing both sides of the fence, i.e. agency technology partner and agency media vendor. Sir Martin acknowleged their frenemy but when will advertiser’s wake-up? DT.

  5. Rob,

    Great article. How do you see Nielsen OCR or the audience platforms like Axciom playing into this?

    Contribute to the trend away from third party ad servers, and/or move discrepancy discussion to audience from impression or clicks? Or not?

    I find it interesting that Facebook is lining up Atlas, Liverail and their partnership with Nielsen OCR in the video space which puts more pressure on DFA in how to stay relevant, I suspect.

  6. Okay so tell me if this is a stupid idea that has already been done a million times. I am thinking about developing an ad serving program that only works in between the advertiser and the publisher and you pay for it through a monthly subscription not through a percentage of the advertising fees. Right now it seems most publishers are paying about 40 percent of ad revenue to Google or whoever their net work is and thus the advertiser is paying about 40% more then what is acceptable to the publisher.

    Seems like a pretty large gap that can be filled by some simple 3rd party software that will insure accountability between the 2 parties and save or add a lot of revenue to one or both parties.

    Does this make sense?

  7. Gerbrand Petersen

    Hi Rob. A well written article, but I believe you have to put this in perspective. I certainly agree with you if you look at an advertiser’s point of view. However, third party ad servers are still of value today for smaller publishers who have almost zero knowledge of what’s going on inside such a server. New hosted services like are continuing to take a market share at the sell-side, especially for WordPress users who just want to sell or manage their advertising spaces. For those publishers with a small budget, the big players in the field are too expensive. I believe the third party (publisher) ad server is certainly not dead – there are always people who’re not technologically savvy and looking for a minimal service for a minimal price.