Tag Archives: dsp

Self Service Still Requires Service

In the ad tech space, you basically are selling against two service levels:

  • Self Service
  • Managed Service

Over the past five or so years, these have become en-vogue since most companies in ad tech sell platforms of some kind.

As a buyer (or user of the platform), you basically have a decision to make.  Do you want to operate the platform yourself or do you want to rely on the company selling you the platform to handle operations.  The former is self service and the latter is managed service, fairly self explanatory.

Over the past 4-5 years in the DSP space, the majority of the marketplace was managed service.  This is for a host of reasons including lack of talent to operate platforms, simplicity around billing, and expected performance.

I am witnessing the tide shifting.  Buyers of platforms are shifting to self service operations and building on top of and around these platforms.

A bit of advice to companies selling “self service.”    You probably want to sell “self service” because it makes your investors happy as it’s higher margin (less staff to service) and you want a tech platform multiple, not a service multiple.   Companies who have biased towards this have severely lacked in support, even to the self service users.

You need to support your users of your platforms, whether or not they are self service.  Over service them.  Because if you do not, someone else will.  Your technology will be a commodity in future years.  Your service will stand it apart.

* Reading:  If you want to read a good book about service, read Setting the Table by Danny Meyer, the owner of some fabulous restaurants including Gammercy Tavern, Tabla, and plenty of others.  Service transcends industries, so don’t fear that the book is not about ad tech.

Mobile Display Networks & DSPs

At both The Media Kitchen and Ventures, we’ve been watching the mobile display networks and DSP space evolve.  At The Media Kitchen, we’ve been testing and running with about a dozen mobile display companies and at Ventures, we’ve been talking to many of the mobile DSP’s over the past year or so trying to figure out how the market will evolve.

A buddy of mine, Ryan Griffin of Digitas had a great quote in Crains this week talking about how mobile display networks are unfolding very similar to how web desktop display networks did in the late 90s and early 2000s.

“A year and a half ago, we were at the stage of where we were in desktop [Web advertising] in 1998,” said Ryan Griffin, group director of media and mobile at Digitas. “It’s now feeling like 2004.”

I’ve been following that party line as well as things are very, very similar.

If we look to the early desktop display networks to predict the future of the mobile display folks, we know that there will be plenty of mobile players, some early exits and some mobile display companies remaining strong independent companies (potentially even going public themselves).  It’s played out this way… Quattro, Admob, Greystripe, and others have already been acquired by strategics and there are plenty of independent mobile companies still in the mix including but not limited to Gradient X, SessionM, Fiksu, mdotm, StrikeAd, inMobi, Adelphic, and dozens of others.

Some open questions I have for the ecosystem:

  1. Will desktop based DSP’s or networks merge or acquire (or be acquired by) mobile DSPs or networks?
  2. Will desktop based DSP’s build their own mobile components? Similar to DataXu.
  3. How are current mobile networks and DSP’s tackling the measurement and attribution issue?  The UDID ban has really put a dent into the measurement space so wondering who will tackle this.

#3 above is important.  The reason why it’s so important is because if you are a vendor on a media plan and can measure/track your performance, then it’s much easier to substantiate yourself on a plan and have a higher chance of getting extra media budget.

Areas that I’d stay away from in the mobile media ecosystem:

  1. Rich Media – if we look at desktop display, rich media CPM’s used to be in the $1-3/cpm range but now, we’re down in the $0.20-0.80/cpm range. The CPM’s have fallen dramatically and I expect to see this within mobile.
  2. Network – if the DSPs succeed and mobile inventory flows to exchanges, then networks have a much smaller role in the evolution of his new media marketplace.

Just some thoughts on the mobile media ecosystem.  Would love to hear yours, leave them in the comments.


The 87.5% Category According to Luma – Lots of Acquisitions

I’ve spent a bunch of time with ad servers in my life.  It all started when I was installing phpAdsNew for my brother‘s website, Exotic Car Network, creating ad zones across Student.com in OpenAdStream, a web property that about 7 of us ran in the late 90s, working on the team to create a self-service ad buying platform for eBay, creating a proprietary ad server for the in-game advertising marketplace called Radial, founded MDC‘s trading desk practice using BidManager and TerminalOne, and finally, using DART, Atlas, and MediaMind at the agency that I’m currently at.  Now looking back at it, I’ve centered much of my career around served and tracked media.

So as you can see from above, I spend quite a bit of time with media technologies.

I believe they will play a large role in the future of advertising and I will continue to play in this field over the coming decades.  There’s been an increased amount of coverage in this space which was once reserved for the back floors of the premier industry showcase, AdTech.  Terry Kawaja is bringing some light humor and some fantastic charts, John Ebbert is creating a mini-media empire (well, not an empire), Brian Morrissey is resurging an old newsletter back to the top, and there’s been a handful of acquisitions lately including Admeld (Google) and MediaMind (DG Fast Channel) totaling close to $1 billion.

According to the display Lumascape, the only category with 87.5% of companies acquired, yes, 87.5%,  is the ad server category.  Crazy when you think about it.  Atlas, DART, MediaMind, Pictela, PointRoll, MediaPlex, etc have all been acquired.  There are some independents in the market today such as OpenX and insurgent AdZerk, but the majority have already been acquired.  I predict that the category is still ripe for innovation and will continue to see many new players enter the space.  I was a personal shareholder of MediaMind and it was one of the larger positions I’ve held.

I continue to think that the best 3rd party ad server is exactly that – a 3rd party ad server that is not biased towards any media.  Almost exactly a year ago, I wrote a piece titled, “Insurgent:  How to take down Atlas and DART.”  I continue to dislike Google’s positioning in the marketplace as overnight they theoretically could shut off access to their inventory for non-DART users (could, not should), Microsoft has an ad serving system built in the 1990s and still feels like it, and this left an opening for a major ad serving player to come in with an independent stance, thus MediaMind gained traction.  I love this stance as mentioned above and it needs to continue.

A few days ago, I wrote a post about attribution and the growing advertising operations line-item.  If you haven’t read it, you should.

If I personally was to start a company tomorrow, I’d probably create the next 3rd party ad serving system built for the future of all media (able to serve site-direct placements, social media and RTB) and include the opportunity for biddable, rich media, video, and full reporting & analytics.  I believe no ad serving system delivers superior reporting and analytics so this is an area that I’d specifically make sure I’d nail.

I think this is an area for massive innovation because the vision that the industry hasn’t recognized the full vision for the future… I believe that all media will be served, tracked, and optimized across all channels.  Television, print, radio, and out of home will all in some way or another be served, tracked and optimized.  This obviously cannot happen overnight as there are quite a few barriers and obstacles to go thru, but the opportunity is huge.  There is a reason why 87.5% of the companies in the ad serving segment have been acquired.

Note, I don’t think you need to start from scratch.  If you could raise some money, you can start by acquiring several of the pieces.  There are quite a few DSPs who could use an exit right now.   There are even some large independent ad servers who would be interested.  A roll-up strategy would be interesting and something that could come together nicely.

One of the bigger parts here is that you need to service the advertiser (or marketer).  This needs to be written on all company walls.   Mandated thru corporate handbooks.  This is similar to how SSP’s service the publisher.  You cannot be all things to everyone and when this happens, decisions sometimes become very tough to make as you try to please too many constituents.  Think about all the decisions Google (and Admeld) had to make about where to flow it’s ad dollars – to Google-run sites or it’s 3rd party network of sites.  Stick to one core area of focus and innovate within it.

This area gets continually dinged because there isn’t a ton of money to be made.  $0.04 average CPM and 200 billion monthly impressions net out to about $8MM in monthly ad serving fees (~$100MM/yr).  That’s a nice company but think of how many are larger.  If you create a robust reporting & analytics infrastructure, ad verification, workflow solutions, etc – you can charge a premium.  I believe you can.  Create a premium bundle of services to execute within the Ad Ops space, and sell them as one package.  There are buyers.

The space is only going to heat up further.  Continued innovation, a lot more media technology thinking, and investment will raise the industry forward.  I know I want to be a part of it!

Data Driven Platforms: Search & Display

Using data to make better business decisions is nothing new but since 2008, it’s been a very hot topic.  It’s been covered in almost every issue of Advertising Age, a whole new web destination now exists: AdExchanger, and agencies spun up new trading desks that are essentially high powered SWAT teams that combine rich data-sets with media for exceptional results.

Google has built a $154 billion company based on the use of lots of data to make the right decisions for it’s advertisers.  The data sets that Google are using are based off of search queries, which one can argue is one of the most powerful data sets that exist as it’s pure “hand raising.”  Over the past ten years, the big search engines have integrated with search engine marketing platforms such as Marin and Kenshoo to provide access to that marketers and agencies can use these tools to make better decisions which should improve performance and create workflow efficiencies (amongst many other reasons).

The trend we are going to see in 2H2010 and certainly in FY2011 is the emergence of these tools within the display, video, and mobile world and combining SEM with them.  While I can’t speak for any one tool directly due to confidentiality reasons, we are going to see many of these once SEM-only players move upstream to capture additional ad dollars and to use all-data (search+display+video+mobile, etc) to make better decisions.

As illustrated below, agencies such as Efficient Frontier are moving into this space as well, as well as, tools/platforms are integrating into biddable display sources.  These are not the only companies moving into the space but are illustrative of the trend.

What does this mean for the standalone display side platforms?  For the standalone SEM platforms?

An operational hurdle that will have to be addressed within the media agency world is that search and display is generally bought from two separate groups so either a) these groups will need to be combined or b) we need to provide clear roles and rules for each group.  I think option (a) is a much better choice as I’m all for integration.

SEM Platforms

Short vs. Long Term & Incentives

I’m taking que from Seth Godin and keeping my posts shorter and more succinct and hopefully this will spur more frequent updates.  Let me know your thoughts if you get a chance.

Anyway, I’ve been thinking about the issues around short vs. long term vision and incentives and the interplay between them.  Let me start out by stating my position:  I think we have really screwed this up.

What’s best for the business is relative to the time frame that you are speaking/inferring.  Employees may make really solid short term bets and sacrifice the long-term vs. passing on a great short term oppty to focus on the long term.

What spurs this?  Without any scientific proof or really any research, I’d have to assume that “incentives” play into this.

Let me play this out with a real world example that hits home to many readers of this blog.

At “work,” I’ve created the evolution of the data, targeting and media industry with the belief of demand side platforms and their role out within the big marketing & services holding companies.

In order for DSPs to work across biddable inventory, publishers have to put their inventory into an “exchange” of sorts (I use exchange loosely).  Because many of the exchanges are part of media companies that are also representing inventory and have direct sales team, there is inherent sales conflict.  This sales conflict makes the long-term vision of the industry harder to attain because people go for short-term gains.  The rationale?  I’m assuming that the sales staff of these organizations are incentivized to beat quotas; and looking-out for the long term does not put food on the table today.

So how do we fix long vs. short-term thinking and re-align incentives throughout entire organizations?

Right now, it’s screwed up.

New White Paper: Getting Real

Some of my friends over at DeSilva+Phillips a boutique investment bank (also known as MediaBankers) have just released a white paper (links to the PDF) entitled Getting Real.  I am always skeptical about investment bank led white papers (unless they are presentations by Mary Meeker which I love) but these guys pulled off a pretty solid document.

The goal of the paper was to talk about the marketplace of ad exchanges, RTB (real time bidding), and the future of online advertising.  I would say that the paper does a good job of setting up the history and current ecosystem that surrounds ad exchanges but lacks anything tangible about the future of online advertising other than saying that many media channels will become digital so a digital infrastructure today is required to service them, which I’m a big proponent of.

There is about $430MM invested in this ecosystem over the past 3 years which has attracted lots of attention from the press, marketing services companies, and of course, investors.  Figuring a venture capital fund wants to see some sort of liquidity within a funds lifespan (~10 years), the next 5-7 years are going to potentially be very “acquisitive.”  Thus probably the impetus for DeSilva+Phillips to release this document to establish themselves as leaders in this space.

Some of the main highlights of the paper:

  • A nice breakdown of what “ad exchanges” actually are including the great Online Advertising Ecosystem graphic from Matthew S. Goldstein that has been circling around some inner circles.
  • Discussions around Right Media’s decision to focus only on premium inventory
  • A solid definition around DSPs without over-hyping them
  • Accenture, which is not normally considered a player in this space is potentially moving in and could open up a new business within the space (has AdChemy partnership, owns a search bid management platform)
  • “Real-Time Bidding is the glue for melding display-ad marketing and search marketing”
  • Talk about SSP (supply-side platforms)
  • Yahoo vs. Google and the ultimate threat of Google (it is scary)
  • Neutrality around ad exchanges and how it’s not a long or even mid-term option (though I’d argue that is what this industry needs)
  • Adding two more acronyms to the vocabulary:  DBO (Dynamic Media Buying Optimization) and ARM (Audience Relationship Management)

You can download the paper here.