Reprinted from iMedia Connection

Overview:

DSPs are really just networks in disguise, they will make networks obsolete, and are a fad with no real staying power.  All false.  In just a few short months, the term demand-side platform (DSP) has become ubiquitous in the online advertising industry. It has become synonymous with all things real-time bid, exchange-sourced, display advertising — in many cases replacing the mainstay term “network” as the model of choice for advertisers. All kinds of media brokers are now scrambling to offer a DSP solution, relegating words like “network” and “optimizer” to the dustbin of display terminology. But as more platforms wade into the opportunity waters, it seems like an equal amount of fog is being injected into industry discussions. So I thought it would be an opportune time to lift some of that fog and expose several of the bigger DSP myths being perpetuated today.

Myth 1: DSPs are really just networks in disguise

False. There are some real differences between DSPs and networks, but recent trends have blurred those lines and given birth to this popular myth. At a fundamental level, traditional networks rely on a large stable of direct publisher relationships to deliver premium placements, easy reach, and ample scale, while owning the media risk and performance responsibilities. Many of these traditional networks live on today amid the DSP wave, successfully delivering campaigns along the way. But networks have begun to rely on exchanges as easy aggregation points for quick scale, and that is what started the move to DSPs. Then when the networks began to layer on automated optimization and advertiser-facing controls alongside their exchange-sourced media — either managed or self-service — they started to look like a lot like DSPs. This is how the whole DSP phenomenon began to accelerate. As networks began to rely heavily on exchange-sourced media while automating the trafficking process and exposing levers and knobs to the advertiser, some essentially became demand-side platforms. With the scale of the real-time bid exchanges and external facing controls, yesterday’s traditional ad network becomes today’s “hot” DSP. But there aren’t going to be as many DSPs as ad networks — read on for why.

Myth 2: DSPs will make networks obsolete

False. It will remain so until all available inventory is transacted on the exchanges. That said, DSPs are creating a “sofa bed” problem (neither a sofa, nor a bed) in the network business. At one end are networks — and publishers in some cases — delivering high-value inventory perhaps with custom or innovative creative. At the other end is the pure performance network that shoulders the media risk and runs cost-per-action performance campaigns. In this scenario, the exchanges are a fabulous source of cheap inventory for the performance networks. In the middle of this are the networks with a rapidly eroding publisher base. It is these networks that are in the greatest danger of obsolescence.

Myth 3: Using a DSP means you get brand-safe campaigns

False. It seems like brand safety is being treated like transparency in our industry. Everyone offers it in one form or another, but like transparency, brand safety means different things to different organizations. Just like transparency, everyone defines brand safety differently. There is one thing that separates effective brand safety from all the other forms of impression analysis and delivery verification out there today: preemptive filtering. Is your flavor of brand safety truly preemptive? Or is the “brand safety” you are getting really just post-impression analysis? Is it reactive?

Automatic brand safety is another of the industry myths perpetuated by networks rushing a DSP to market. Real brand safety, meaning a demonstrably safe environment for a brand advertiser to promote its message online, comes from truly preemptive brand safe filtering. But in practice, pre-impression analysis for relevance and safety is difficult to do. Evaluating a billion impressions a day for applicability takes a robust platform and intelligent integration with all of the real-time sources – and this requires investment in technology and people to do it right. But preemptive filtering pays huge dividends in efficiency and thus returns. With truly preemptive brand safety, you only buy what is safe and there is no waste. No wasted targeting, optimization, time, or ad-spend.

Myth 4: There are dozens of DSPs today, and there will soon be hundreds more

False. There is no shortage of organizations claiming to be a DSP. But you can count on one hand the number of solution providers able to deliver a truly successful DSP campaign (perhaps even minus a couple of fingers). In the heyday of networks, there were reportedly more than 300 ad networks out there, but few had a significant amount of their own proprietary technology. It is the large, technology-rich networks that have successfully transitioned to become real DSPs. The rest are scrambling to rely on relationships for their technology, as they had little or no technology of their own to expose when the DSP tsunami hit our industry’s shores. Because of the technology required, there will be far fewer successful DSPs than there were ad networks. Dozens? Maybe somewhere down the line. Hundreds? No way.

Myth 5: DSPs are a fad and will not have real staying power

False. Far from a fad, the DSP trend is about control, efficiency, and profitability. Advertisers never wanted to outsource all of their control; they did it for the quick scale and reach that networks provided. Now that the exchanges have matured and have exposed real-time bid APIs, control can move back to the advertisers, while retaining all of the scale and reach that networks can offer. Once we go down this road, there will be no turning back. The DSP trend is causing a fundamental change in the nature of how display media is used.

With any significant new trend, there is bound to be some confusion. Hopefully this clears the air around some of the more prevalent myths in our industry. There are sure to be more myths and many different viewpoints, but this is a start. What is certain is that the trend toward the self-service DSP model is going to have some staying power. It has already made an irrevocable impact on how display media is brokered and continues to drive upheaval. It was way back in 475 BC that the Greek philosopher Heraclitus said, “The only constant in life is change,” and those words are just as true today. The display advertising industry has gone through constant change since Hotwired ran the first banner ad in 1993, and the nimble organizations will be the ones that continue to flourish.

It seems like “brand safety” is getting to be like “transparency” in our industry; everyone offers it in one form or another but like transparency, brand safety means different things to different people.  And just like transparency, everyone defines brand safety differently.  There is one thing that separates effective brand safety from all the other forms of impression analysis and delivery verification out there today.  Preemptive.  Is your flavor of brand safety truly preemptive?  Or is the brand safety you are getting from your demand-side platform (DSP) provider post impression analysis?  Is it reactive?  This is another of the industry myths perpetuated by other solutions rushed to market.  Real brand safety, meaning a demonstrably safe environment for a brand advertiser to promote their message online, comes from truly preemptive brand safe filtering.  But in practice pre-impression analysis for applicability and safety is very difficult to do.  Evaluating a billion impressions a day for applicability takes a robust platform and intelligent integration with real-time sources.  But it pays huge dividends in efficiency.  With truly preemptive brand safety the pass back is meaningless.  With truly preemptive brand safety you only buy what is safe and there is no waste.  No wasted targeting, optimization, time, or ad spend.  Make sure your brand safety is preemptive before you put your message out there.

(Reprinted from ADOTAS) 2010 is already shaping up to be the year of real-time bidding (RTB) technologies and demand-side platforms (DSP). The availability of real-time bid access into the major ad exchanges is converging with an industry-wide preference for agency-administered ad-buying and ad-trafficking platforms.

The traditional ad network model achieved many things but has yet to deliver on the promise of truly cost-effective scale. The rules of engagement changed when downward-spiraling CPM prices forced publishers grudgingly onto the emerging exchanges to improve their yield and back-fill diminishing ad revenues. The networks had to follow their publishers and a new aggregated sourcing model emerged.

Real-time bidding is the hot feature this year and a staple of the effective DSP. Now the combined capabilities of RTB-enabled DSPs built on the exchange model are replacing traditional ad networks with a new buy-side network paradigm that is more nimble, more economical and more in touch with advertiser goals.

Control is subsequently moving closer to the advertiser, intermediation is being reduced and prices are arriving at a true market-driven equilibrium. Not only do advertisers have more control over targeting, performance and safety, but buying has become more centralized and access to inventory become more streamlined.

It is no secret that LucidMedia has one of the most robust contextual targeting engines in the marketplace, but what we have achieved in the DSP space over the last year has not been widely publicized. Late in 2008 we began engaging all of the large aggregators to co-develop RTB solutions. Next we developed an advanced, proprietary ad server to give us one of the most nimble systems available for trafficking campaigns.

We also built a unified inventory management system that could dispense with the complexity of hard-wiring campaigns to inventory sources. We included an automated optimization engine that can evaluate thousands of campaign facets in real-time, project performance trends, and govern campaign targeting based on advertiser goals.

In January 2009 we deployed the platform internally, effectively making it the industry’s  first production RTB-enabled demand-side platform. Named ADvisor DSP, LucidMedia has since executed hundreds of successful campaigns on the platform and is currently processing up to 45 billion impressions each month. This pioneering experience makes us uniquely qualified to access the necessary features of a successful RTB-enabled DSP initiative.

There are certain universal features that agencies need for any successful DSP initiative. All buy-side platforms must offer multisource RTB integration and scale, an advanced ad server, page-level contextual analysis for targeting and brand-safe filtering, audience profiling for retargeting, universal frequency capping, detailed performance roll-up reporting with discrepancy management and reconciliation, flexible and intelligent bidding strategies, the ability to leverage third-party targeting data and a managed service deployment approach.

Advanced Ad Server: A good DSP starts with an advanced ad server. Core ad serving capabilities like frequency capping, day parting and targeting are minimum requirements. The savvy agency should also look for full-featured holistic campaign management features like cross-aggregator universal frequency capping, pre- and post- impression auditing, preemptive brand-safe filtering, page-level impression verification and hyper-segmented targeting based on content, demographics, behaviors, site, location, and time of day. Discrepancy management and reconciliation are also critical aspects of the built-in ad server.

Integrated RTB: Properly integrated real-time bidding is not standardized or modularized. There is still a great deal of heavy lifting development needed to bring on each inventory source, scale and balance the volume, and bid effectively. A good DSP needs to have these problems already solved.

Intelligent Bidding: Smart bidding strategies are also critical to an agency’s bottom line. The DSP model promises that agencies can claim a larger slice of the ad spend dollar, moving it further and further away from the networks. But without intelligent and flexible bidding, that slice can be lost. Look for detailed graphing of bid price and win ratios and the tools to quickly adjust bidding per source. Look for RTB solutions efficient enough to drive bid costs below the $0.001 threshold.

Retargeting: Another aspect of the successful DSP is the ability to roll up audience and link users to performance for retargeting purposes. Reaching the right audience with the right message at the right time is a core tenet of advertising and although this applies to all mediums, display always reaches its users through a proxy device.

The right platform goes beyond simply rolling up interpolated audience facets and instead segments users into actionable profiles using third party data based on their propensity toward a desired action.

Managed Services: The concept of an agency-side buying and management platform relies heavily on managed services at the inception of any in-house DSP program. The current transitional period is favoring the managed service approach to demand-side platforms as agencies step into the traditional ad network role. Managed services allow the transition and knowledge transfer to happen in the most effective manner.

Optimization: Automated optimization is a differentiator for the demand-side platforms. Historically optimization was the mystical secret sauce of the more technical ad networks but it has become a required attribute of the full-featured DSP. Look for optimization that can juggle thousands of campaign targeting facets, project outcomes, and model performance scenarios prior to launch.

Inventory Sources: Sources and scalability are key factors for every agency to consider. The right platform needs to plug into all the large repositories as well as the more niche aggregators plus the all supply-side optimizers. This kind unprecedented impression potential, scale and broad reach are requirements for large direct response campaigns, major corporate branding promotions and scaled niche segments of very specific demographic traits.

These are the major features as well as ad network staples like brand-safe filtering, transparent reporting down to the page level and flexible targeting to content, demographics, and behaviors that an agency should look for when selecting a DSP.

The right demand-side platform allows agencies to easily audit the networks and exchanges on their media plans, efficiently acquire page-level contextually and demographically targeted inventory, ensure brand safety across all sources and most importantly enforce a universal frequency cap. The new breed of DSP with integrated RTB gives agencies pre-impression filtering, post-impression auditing, and allows agencies to cherry-pick the most effective impressions in real time and then feed their campaigns with an optimization engine that automatically maximizes return on spend.

This is reprinted from “Data-Driven Thinking”, a column on AdExchanger written by members of the media community and containing fresh ideas on the digital revolution in media.

Demand Side Platforms (DSP) are hot! I can tell by the huge agency interest, and even more eager venture capitalists anxious to get in on the latest craze. Traditional ad networks and newfangled technology platforms are declaring themselves to be DSPs. Others who did much of the evangelical spadework for DSPs appear to be stung by the sudden rush. There is now an attempt to define a “true DSP”. At this stage, a “true DSP” as defined by a list of features serves little purpose and is as much a disservice to the industry, as it is disconnected from reality. In fact, many of the current DSP competitors–those with the most significant solutions already in-market–are successfully violating that definition of a “true DSP” to the benefit of their agencies partners.

The truth is that a “fully self-service DSP” would be far too disruptive to most agencies at this early stage. There are far too many levers, knobs and buttons in a DSP robust enough to deliver the optimum cross-section of pacing, performance, and price for an agency to take on today. They range from mundane tasks like dealing with objectionable impressions and buys from non real-time sources to more arcane optimization tasks, RTB source integration, bidding strategies, discrepancy management, and post-campaign reconciliation.

As my own company has learned by providing DSP services to agencies over the past year, agencies are still not appropriately staffed to be full-fledged buy-side networks yet. Media buyers are already over worked and stretched to the limits and are looking for a DSP to do more with less. That means automating many tasks but also off-loading just as many (if not more). A managed service that facilitates knowledge transfer and leads to a semi- self-service approach is far more realistic today. The reality of the situation is simply not that black and white! At the end of the day what an agency really wants is to know what works for their clients and how to repeat those outcomes in the future. At least that’s been our takeaway. They want the data behind the performance, the audience segments that engage with their message effectively, and the easy-to-pull levers that will let them do it again on the next campaign. They want no more and no less.

It is also important to recognize that many of the most important DSP offerings are coming from the networks where the technology has matured and has been thoroughly field tested. Indeed, my company falls into this category. The networks already have the relationships and business models in place to support the early adopters. Real-time bidding is an excellent example of this necessity. To get multi-source RTB going at scale going takes a significant amount of relationship building and technical heavy lifting and many of the networks have already invested heavily in this. Discounting the network players with a single stroke, and leaving it all up to the agencies, is doing the industry a massive disservice.

There are many other aspects of a “true DSP” that were missing from the recent list as well. Our experience has been that no DSP should leave home without staples like universal frequency capping across exchanges and publishers, objectionable content filtering for brand safety, some sort of automated optimization for CTR, CPA, or eCPC and the ability to plug in others, an advanced and robust ad server capable of propagating campaign changes in minutes, seamless targeting of audiences with native and 3rd party data, and granular content targeting independent of site or section.

Only time and buyer requirements will define the “true DSP” and decide if networks and DSPs can truly coexist. The industry needs time to shake it all. The sector is too young to define the full product category or for one player to define it completely. I will conclude with a heretical prognostication–the next generation of ad network will be a hybrid DSP solution with a service layer.

We recently did a Q&A with AdExchanger.com about LucidMedia and our work with the Ad Exchange community.

Q: What trends are you seeing in online display advertising?

A: Trend #1 is performance.  All the trends we are seeing are towards performance-based advertising and increasing return to “do more with less”.  Advertisers are looking for new ways to increase their return with a flat budget but all the same growth pressure.  They are turning more and more to direct response, CPC, and CPA solutions to move the media risks as far away as possible and guarantee the backend performance.  Apparently our fully transparent platform with multiple optimization facets really resonates with clients in this tough new economy.

Trend #2 is exchanges.  We are seeing a dramatic shift from networks to exchanges.  The exchange model has matured at the same time the network model has had its reputation tarnished.  The real-time bidding and open, transparent nature of the free market exchange model has finally distilled to the top as the solution of choice when efficiency and return become as important as reach and scale.

Q: Is it fair to say that LucidMedia is the contextual solution for Right Media Exchange?

A:  Yes and no.  Yes, we are a contextual engine on Yahoo’s Right Media Exchange but that is not our core business, it’s just one of the many irons we have in the fire these days.  And our work on RMX is a two way street.  We provide contextualization services to publishers and advertisers there but we can also purchase media there.  We have many of these types of media arrangements because it is how we provide the scale, efficiency and performance that are so appealing to our advertisers.  It took years to do but we are now able to pass those benefits on to our advertisers and our agency partners.  That is a key to the current rapid growth and momentum I mentioned in the first question.

Q: Any plans to work with Yahoo!’s APT Platform?

A: We are watching APT closely, as many are, because it holds a great deal of potential scale.  Obviously we’d like to continue our close and positive relationship with Yahoo and Right Media and integrate with APT when it is fully operational.

Q: Can you take us through the process of how LucidMedia’s deal with Right Media works from a technological perspective?  For example, tell us how it all works for a single ad impression on a publisher’s site.

A:  Right Media has established 63 standard categories that correspond to a 4 digit contextual flag.  RMX publishers have to deploy new ad tags and then advertisers can target their inventory more effectively.  LucidMedia’s ad tags are distributed to RMX publishers instead of YieldManager tags.  Then, during ad call, the LucidMedia tag analyzes the context of the page in real-time to determine the most appropriate categories.  We then augment the YieldManager tag with additional Query String Codes that represent the most appropriate Right Media Exchange categories.  All advertisers have to do is use Right Media’s Query String Targeting capability when they setup a campaign in YieldManager and they can access the highly targeted contextual inventory right away.  Your readers can go online to find out more about it too.

Q: Can Lucid Media’s ClickSense technology contextualize social media other than, say, the “social media” category?

A:  Sure.  Our technology does not care about self-declared categories.  Rather we deal with exactly what the content is about and categorize it accordingly.  Our solution scrapes the page in real-time to determine precisely what each inventory page is all about.  What we find is that the majority of content is improperly categorized at a high level.  We find social media can be all over the map category-wise.  Technology, Arts and Entertainment, Automotive, Sports, Pets, Family and Parenting, Health, and so on.  So we tag media at the page level for what it is about and not what it is supposed to be about.  And our performance proves out the approach.  We’ve worked with advertisers who have studied this and their findings are always the same: that context is a true predictor of intent.

Q: How is the contextual engine for LucidMedia different than the competitors such as AdSense, Contextweb, Kontera and others?

A:  Although we all have a contextual solution in common, our focus is on providing a broad platform that encompasses a range of advertiser and agency services beyond just running media.  In fact, clients can actually utilize our platform as a compliment to some of the other contextual providers.  In terms of technology, our approach is a bit different as well.  At the core of our platform is a patented semantic solution that includes advertising industry-focused taxonomies which let us target content at a very granular level.

Q: In the future, can you see a network providing the essential services that an advertiser requires and, thereby, disintermediating media agencies?

A:  I think that is unlikely to happen.  For example, our focus is on providing a media management platform to agencies that allows them to outperform their competitors and pass on new levels of efficiency, transparency, and safety to their advertising clients.  So we are not trying to act like an agency, we’re trying to put tools in their hands that make us a crucial part of the value chain.  We feel our position between the agencies (and the advertisers) and the networks (and publishers) is the best place to do business.  This allows us to act as a media buying platform for the exchanges–and even the other networks out there–and provide our data as the new currency of performance.  It’s a very exciting place to be right now!

Q: How is LucidMedia ensuring brand safety?  And, how does LucidMedia provide access to the Long Tail?

A:  Brand safety is rising in importance these days.  Our Verified Inventory Platform (VIP) leverages our deep contextual technology to find the right content to meet the ever-tightening advertiser performance goals.  But we can also block content in the same way when they indicate something is inappropriate for their brand.  Because we evaluate the true meaning of each page for categorization, we get the by-product of knowing exactly what topics are on each page.  We have compiled an extensive list inappropriate topics that we call our Objectionable Content Filters.  With these filters enabled we can make sure that their brand won’t appear on pages about hate or pornography or even things like natural disasters or war.  And these are all customizable by the advertiser because what is inappropriate to one may be desirable content to another.

As for accessing the Long Tail for our advertisers, we deployed the concept of Media Classes within our platform to take advantage of the Long Tail.  We not only categorize the page content accurately, we also categorize the type of source it is found on like news sites, social networks, blogs, enthusiast forums, gaming sites, wikis and webmail portals.  We also categorize the sometimes undesirable media classes like peer-to-peer file sharing sites so advertisers can not only target specific media classes, they can also filter against certain classes of media if they want.  This opens all kinds of doors for our advertisers.

Q:  In your opinion, what will be some of the key drivers which will allow ad exchanges to progress from a remnant-only to a premium and remnant model?

A:  The key drivers will be their openness, a clear value proposition to the publishers, and their ability to support real-time bidding.  The exchange platforms have to be easily extensible so everyone can play.  And we’ll need the publisher side optimizers to keep advancing as well.  They play a key roll that the exchanges are not filling today and they exert a great deal of pull on the publishers drawing them to the exchange model.  I expect to see a lot of interesting changes in the next few years and LucidMedia plans to be right in the middle of it all adding value to the agencies and advertisers.

I was reading Mike On Ads’ multi-part series on Ad Exchanges and I got to wondering what forces came together to create the current ad exchange phenomenon? There’s no denying the emerging ad exchanges are replacing the old yield management solutions out there. They are aggregating the supply to drive new market efficiencies and a new level of transparency in the non-premium marketplace. One of the early factors motivating this was the proliferation of ad networks which have been growing at a staggering rate. ThinkEquity Partners recently reported that there were over 300 ad networks in 2007 which means the number doubled in less than two years. They went on to report that the non-premium market will grow at 28% annually from $2.2B today to $7.6B in 2011 so it’s no wonder the number and the types of ad networks exploded. Vertical, contextual, behavioral, demographic, re-targeting, geographic, site specific, there is no shortage of ad networks out there now. Because of this explosion the mantra of the crowded, long-tail, remnant world of non-premium advertising has become differentiate or die. Why? Because there was (and still is) pressure from all sides to stand out. Pressure from eroding gross margins, strain from publisher recruitment, competition over inventory, negative stigmas about duplication and a lack of transparency, and a fast and furious industry roll-up. The growing revenue base and customer demand needed another solution and the ad exchange was the logical evolutionary path. Ad exchanges are streamlining the process with a whole new level of efficiency that the ad networks tried to deliver but lost along the way. The ad exchange is basically an ad server ecosystem through which advertisers, publishers and networks all manage their advertising business. They do this together and in an open, platform agnostic way that allows market dynamics to work their magic. So now the ad networks are feeling pressure from the exchanges too driving an ever increasing need to differentiate themselves. Think Right Media and the DoubleClick Advertising Exchange as examples. Advertisers and agencies rely on ad networks for the efficiency, reach, and optimization they bring to the table and are willing to give up some editorial control for it. But ad networks tried to control the whole process through proprietary means. This opened the door for exchanges to step in because they simplified and unified the trafficking process on an open platform that was transparent to the process. And there we have it, transparency is the final piece to the puzzle that unlocked the exchange phenomenon. Transparency takes the duplication out and removes the waste. We’ve all heard the timeless advertising adage, “Half the money I spend on advertising is wasted; the trouble is I don’t know which half.” Well transparency addresses that problem. So the ad networks out there will continue to differentiate themselves if they want to survive. The winners will be the exchange-friendly networks who can deliver the same transparency that enabled the exchange phenomenon in the first place.

Dave Morgan, TACODA’s founder and Executive Vice President of Global Advertising Strategy for AOL, just hit the nail on the head with his blog commentary on the stigma surrounding ad networks. He points out that the ad network brought a pork bellies commodity attitude to online advertising in the lean, post-bubble days when any revenue was good revenue. He goes on to point out that this is changing and bravely outlines a few solid reasons why. What drove this home for me was his conclusion that networks are going to have to continue to deliver solid value, with high-quality ads and rates, to begin to offer something that publishers will never be able to deliver on their own because of their focus on the content. He highlights the need for networks to deliver effectiveness over efficiency. Bravo. This is really the crux of the situation and a very real opportunity for ad networks to step up to the plate. And this is exactly where LucidMedia is strongest, in bringing increased effectiveness and the proof report down to the impression-level to prove it instead of just new efficiencies through scale, technology and focus. We saw this void back in 2004 when we launched ClickSense, our robust contextual engine, and have recently applied it to lifting the effectiveness of our customer’s campaigns in the LucidMedia Network. This new effectiveness is inherent in our deep categorization which puts publisher’s content into our 13,000+ industry categories to allow for far more granular targeting than has ever been available in the past. When publishers can sell their inventory with 6 levels of categorization to pinpoint the meaning of their content there is a very real lift in conversions. This assumes all the underpinnings are in place like engaging creatives and streamlined landing pages but when it all comes together there is a tangible return. Hopefully other ad networks will do the same and bring real increases in effectiveness to the table in the future and begin to erase the negative connotations around the ad networks.

Google recently announced a new service that Web publishers can use to manage their online ad sales and serve up ads to consumers. This news is extra interesting now that Google has DoubleClick/DART legally in their stable. I am waiting to see how much transparency they will be turning over to the advertisers and how much overlapping functionality they will have with their other technology acquisitions. I bet they parade this out with much fanfare as a beginner solution with plenty of easy links to upgrade and the upgrade solution is DART. It really comes down to how “advertiser-centric” you can be when you are talking about ad serving. What this industry doesn’t need is just another ad network. What this industry does need is to lose the opacity that has plagued it since the earliest days. Advertisers need brand safety today and brand safety depends on transparency. You can only rest assured that your brand is being represented properly in the Wild-Wild-West that is online advertising when you know where every single impression is going. Not to mention detailed categorization. If anyone can provide the industry with useful contextualization it’s Google but I doubt they will. I want to see some proof reports showing how well they target and every impression URL from this new offering before I make the call whether or not this is a good thing for our industry.

Adotas recently posted an interesting, albeit oddly written article on the value of the ad exchange. The real story seems to be the value of the vertical ad networks–with which I cannot agree more–but their analogy between mutual fund managers and ad exchanges is a fascinating one. They equate the ad exchange to a specialized mutual fund manager who can add real value to your portfolio and that is very true for the new breed of ad exchanges out there today. The scenario buy that they walk through truly highlights the value that an ad exchange can bring to a media planner. They point out that when an advertiser wants to maximize ROAS they need their media buyer (their agency folk) to create a balanced “portfolio” mixing direct investment in large targeted sites with the specialized engagement of the vertical ad networks and the broad reach from the many low cost remnant networks. The share of voice and audience is highest on the expensive premium sites, highly relevant on the vertical ad networks (vertical search sites for instance) and yet there is massive reach to be had across the broad remnant networks. It is the mix that is critical to getting the job done. This is how you nail the ROAS for a big branding campaign where audience engagement is a critical factor–and this is just what the exchanges are good at delivering. They have the mix all ready to go, you just need your media planner, your mutual fund manager, to recommend and execute that perfectly balanced portfolio and your investment will pay off. What an exciting time for interactive media!

Turn Networks has been making a lot of noise in the press recently with some additional capital and the launch of their fully automated ad network. I perused their literature and demo and really like some of the ideas they are taking to market. I especially like the fully automated nature of their targeting based on multiple factors including a mix of contextual site analysis and past behavioral. I also have to give them credit for what looks like a good pricing model for everyone. I would be very interested to see some data on how well their targeting works through their automated approach. All too often I see ad networks talk about contextual targeting and providing relevant ads but in the end I usually find out they are using just caveman-like site or network level targeting to get the job done. That means most networks will sell you categories of automotive related content or health related content but the determination of what content is about which subject comes from looking at the publisher’s overall theme. Cars.com is automotive content and Health.com is content about health. That sort of thing. The reality is that much of the content on these sites is not about cars or health. It can be about almost anything especially when it includes socially-driven content from things like forums or community features. The only way to truly target for relevance is to look at each URL and analyze the content you find there and not lump it into a broad, site-level grouping. Only then do you find that your ads really start performing and your KPIs go through the roof. In a recent case study we found that eCPC rates went up 76% over the typical run-of-site campaign when thorough contextual targeting was employed. That’s how you really lift your ROAS.