Agency Demand Platforms

My friend Jay Sears (@jaysears) sent me an invite to the AdClub/ContextWeb event on Agency Demand Platforms a few weeks ago and was impressed to see the high caliber speakers but was bummed out that MDC/VMM was not asked to participate as well, as, my friends at Havas.

The event was packed – almost every seat taken in the beautiful NY Times Center and quickly got underway by a speech by Carl Fremont about the Digital Action program of the AdClub.  Digital Action is actually really cool – publishers donate inventory and the AdClub gets to sell it and keep the money to provide scholarships for their students.  We’re working on a deal which should hopefully provide some dollars to the AdClub so I was excited to see Carl speak about it.

Bant Breen (Initiative), Sloan Broderick (Mediacom Interaction), Greg Green (VivaKi), Matt Spiegel (Omnicom), Ross Sandler (RBC), and Wenda Harris Millard (MediaLink) all participated on the panel and had a healthy discussion that to industry outsiders may have been fascinating, but to industry insiders, was much of the same fodder we hear each day.  For those interested in reading the play by play of the event, AdExchanger has a fantastic writeup and I recommend checking it out.

I want to use this blog post to talk about a few of the issues/hurdles that affect Agency Demand Platforms and their roll-out to advertising agency holding companies such as IPG, WPP, Omnicom, Publicis, Havas, and MDC Partners.  Many of these thoughts come out at lunches or dinners I have with my peers at other demand platforms but I thought that publishing them online may spark a larger conversation with all industry constituents:  investors, agency execs, media folks, technologists, startups, PR, etc.

What is a Demand Platform?   This term is thrown around with increasing frequency but I’ve not seen a solid definition around it.  AdAge wrote a piece on all of our agency platforms that we are building and this comes closest to what I define as a Demand Platform:  a technology solution with a front-end interface where agencies (media, creative) have access to procurement, optimization, warehousing, and analytics.  All of us are building something within this vein but each of us has our own cupcake with different sprinkles.

I do not think that Demand Platforms have to be limited to single supply sources such as advertising exchanges.  It seemed as if the majority of the AdClub/ContextWeb event was based around exchanges, but supply sources can be various.  Currently, we are seeing typical ad networks, publisher exchanges (i.e. FimServe), reservation systems such as Apt, broad exchanges (e.g. RMX, AdX) and sites direct are able to plug into Demand Platforms.  Personally, I look forward to the day when inventory goes beyond OLA into search, mobile, print, radio, television, and OOH.

Hurdles

Agencies vs. Technology Companies: Agencies have always adopted technologies tactically, but if they are to roll-out a Demand Platform, to be successful, it must be strategic.  Most agencies are not staffed up with technologists (broad term for many different types of technology people) and most management in agencies are not trained in running technology companies.   Technology companies also invest in R&D and top talent, as what they build has scaling capabilities and they can recoup these dollars.  Agencies are a staffed differently and generally never hire beyond revenue.  Depending on the agency, holding company, and timing of the financial markets, agencies wanting to become a technology company may be a far dream (if building their own tech).

I predict most agency holding companies are going to work with strategic partners to roll-out their Demand Platforms as for the aforementioned reasons.  There are a few partners in the market today that are getting some nice publicity and traction such as MediaMath, Invite Media, Triggit, AdChemy, X+1, DataXu, Brilig, and others.  The question is whether to go exclusive or remain independent and this is a major question for many of us.   The partners above service the ad-serving & algorithmic optimization space but there is also data warehousing, data aggregation, and creative optimization that plays within here too and there is a whole host of other partners to deal with there.

Changing Media Culture:  Media planners and buyers have applied the human element to optimization for the past 15 years or so and are used to buying sites as a proxy for audience.  Most of our Demand Platforms focus on buying “audiences,” which is fundamentally different than purchasing sites in themselves, so re-educating the entire media planning and buying world about this new paradigm will have to occur.  Technology is going to strategically penetrate agencies if this happens and the education process is going to take a while and not only are we going to have to educate internally to our agency staff, but also, to our clients.  As always, there are going to be conservative clients/agencies and progressive which means that most people will fall in the middle.  I can’t stress enough how much of a barrier this is for the industry.  Changing culture is not easy and the visionaries will have to be in place in the agencies and clients must remain open minded, as the way they have purchased digital media in the past will change.

I am not stressing that media planners and buyers are going to be out of a job, as one can read the above paragraph and deduce that.  I hypothesize that we are going to have a something like a performance display group and an integrated team.  All IAB/OPA standard units will run through the performance group through a Demand Platform while the media/creative staff’s time will be freed up to create big brand experiences.   For publishers, read my “Goodbye Media Sales Execs” post.

Business Terms/Pricing:  I’m going to attack this dead-on as all too often, this question gets avoided.  In the USA, agencies typically do not take posession of media (unlike Europe) and sell back to their clients, but rather make a scaling commission on all of the media that is purchased on behalf of clients.   With Demand Platforms and the ability to purchase on Advertising Exchanges, there are similar models to SEM (search engine marketing) emerging:

  • Cost Plus – this model takes into account of the cost of the media plus a standard consistent markup for the “secret sauce” (optimization, data, research, analytics, etc)
  • Blended/Arb – this model is where the agency takes posession of the media and prices it back to the client at their discretion
  • Pass Through – media agencies will just take their typical media buying or success comission/compensation for OLA planning

Most of the players including VivaKi, Cadreon, B3, Adnetik, VMM do not publicly disclose their business models and that’s valid and fair.  I’d assume however that most models are based on one or multiples of the three scenarios above.  If there is another scenario that I’ve missed, I’d love to know about it.

Note:  as explained above, agencies in the USA cannot take positions on media and sell back to clients.  This is why seperate companies have been set up to provide church/state seperation.

Real vs. Near-Real vs. Non-Real-Time:  I don’t know how this is a hurdle, but I want to point out that saying “real-time” is the “cool” thing to say right now but there is almost no substance behind it to the trained eye.  There are less than a half dozen exchanges who can support real-time, and of those real-time exchanges, they are dwarfed in scale from the big guys.  There is no doubt that real-time is going to be interesting in the future, but as of July 14, 2009, this is purely an ego play.

The Agency Demand Platform Ecosystem

While working on the Agency Demand Platform project, an entire ecosystem is impacted and innovated upon.  The following business areas are affected by these new Platforms (depending on the particular platform of course):

  • Dynamic creative
  • Analytics (insight)
  • Data warehousing
  • Decisioning and processing
  • Visualization
  • Algorithmic development
  • Yield management
  • Ad operations (trafficking, pixel implementation)
  • Data privacy
  • Media planning & buying (communications planning)

Want to share your viewpoint?   Did I miss something?  Feel free to leave a comment or tweet me at @dherman76

Tagged as , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , + Categorized as Advertising & Marketing, Internet & Web X.0
  • Judy Franks
    I am interested in the contradictions between 'superior audience aggregation' offered by such platforms and the 'big brand experiences' you reference above. On the one hand, context doesn't apply(the site itself) and efficient audience accumulation drives the ecosystem. On the other, context and brand linkage fuse creative ideas and media platforms together into a customized experience. The co-existence of two potentially opposing points of view require careful consideration.
  • christianpbrink
    Just got around to reading it today.

    I think you're right that performance display becomes a radically separate category from 'big branding experiences' when the exchanges finally give you single-channel access to enough inventory that display campaigns can be automated. Teracent et al are trying to make that automation happen yesterday, but I think their fortunes are tied to the exchanges'.

    The fortunes of publishers too, maybe. At-scale display exchanges + campaign automation --> cut out the labor costs historically associated with running display successfully --> more buyers can afford display campaigns --> upward price pressure.
  • michaelkatz
    What you are inherently describing is effective supply chain management.

    To paraphrase your post, successful execution relies upon much more than just building a technology platform. It also encompasses the process innovation to create optimal work flows as well as the analytic support around the technology... I couldnt agree more.

    This is what I wrote about in May, http://www.mediapost.com/publications/?fa=Articles.showArticle&art_aid=106421

    It seems to me that there are a lot of people in the industry right now talking about the same thing and trying to tell a similar story. Building an inventory aggregation platform is one thing, actually being capable of successfully executing a business around it is a very different thing. Thats the part that nobody talks about. Its about the "how", not the "if"...

  • Thanks Darren. This is a much needed POV.

    I would only add that there are incredibly difficult run-time challenges as you begin to patch the platform ecosystem components together and create massively parallel distributed systems. The challenges increase exponentially if your platform is optimizing in realtime (under 250ms).
  • Lee F
    Wow. no comments yet? Darrren, Don't be so sad to not be on stage this week.

    Excellent insight and analysis!
  • Surprised as well, Lee. Not for lack of readers - certainly tons of traffic today.

    I've gotten private DM's on Twitter and a few emails regarding this posting. I'd love to continue the conversation.
blog comments powered by Disqus