Targeting Is the Ad Network “Killer App”


target1.jpgADOTAS EXCLUSIVE — In looking at the bevy of online advertising networks in the marketplace today, it seems increasingly evident that targeting is rapidly overtaking inventory quality among ad networks as the one aspect that their value hinges on and the one that truly differentiates them. Targeting arguably has become the ad network “killer app.” According to E-consultancy 2007 Online Ad Network Buyers Guide, targeting was only 1% behind inventory quality as the single most important differentiator. That was 2007 and the war has escalated considerably since then. Now I’d argue that the single most important factor and point of differentiation within the 400+ ad networks in the marketplace is their targeting.

Inventory quality is still an ultra-critical factor when evaluating ad networks but it has become more like a commodity. Everyone has great sites and can whomp up a spectacular site list with all the visual shock and awe that anyone can imagine. But targeting — good, precise, accurate, performance-driving targeting, takes real technology, which is actually a pretty rare commodity among the ad networks. Most ad networks are made up of people and publisher relationships. That’s how they scale. Add more sites and ad more sales people and the incremental revenue will follow.

Targeting in today’s ad network environment usually means either by behavior, demographic, or context. And all can work to raise performance of the average direct response campaign — or create additional recognition and recall for brand campaigns. Just like transparency means many different things to many different networks, not all targeting is created equal. For instance, behavioral to one network may be just re-targeting to the next network. And then there’s contextual. Most don’t know that in many ways behavioral is based on contextual. In this scenario, behavioral is a contextual solution with a time element.

So what’s the best targeting solution out there to ask for? This usually involves determining what targeting will provide the most performance enhancement for your campaign. And that, of course, hinges on what your key performance indicators are going to be: Is it clicks, acquisitions, brand awareness or a combination or something else entirely?

In many ways contextual targeting may have a leg up on the rest. Behavioral offerings almost always have a contextual component driving their segmentation, so it is one of the most mature technologies out there. Semantic relevance engines have been around since the early days of Knowledge Management in the early 90s — even before the first AT&T banner ad was sold by Doug Weaver to And contextual side-steps any privacy issues, as it derives its relevance from the page content as the ad is being served and does not need to ask private questions or save little bits of private data behind-the-scenes. Perhaps most importantly, contextual targeting has frequently shown to offer both more lift in clicks and more brand recall than other targeting solutions.

From a practical standpoint, when you are out there shopping for an ad network and everyone is pitching transparency, great publisher sites and more performance, it would be well worth your while to stop and ask about targeting: Don’t be afraid to ask about technology, either. Most likely you will find little behind that curtain besides some basic channels, a little re-targeting, and a status report. Ask for a proof report and see where that leads you. Can they offer proof of why they targeted a certain page with a specific ad? If not then there is probably little technology back there. I’d keep looking, because as targeting continues to ascend in importance, so too will the need for ad networks to add real technology and targeting to the people and publisher relationships already in-place. At the very least, it will be interesting to watch how ad networks adapt, adjust, and try to differentiate themselves in an environment that is perhaps more focused than ever on targeting.


  1. We are seeing that targeting, as the basis for optimization, has become the single most critical component to advertisers and agencies. This is especially true now that the economy is tightening marketing budgets and flattening ad spend. Advertisers and their agencies need better insight and more lift in the tough economy and targeting is the driver.

  2. “Targeting is rapidly overtaking inventory quality among ad networks”

    That feels right — except its not just among Ad networks — because the networks compete in the very same ecosystem as the publishers direct sales efforts.

    In other words, for the first time in the 500 year history of Publishing, Content, especially quality content, is simply NOT king. And when quality content isn’t king, it is no longer economical to spend more money to create it. Which means the end of quality, or at least expensive, publishing.

    Isn’t there a moral issue here aside from the Privacy issue? The proliferation of networks with their inherent targeting advantages are causing massive externalities to the content environment. Local journalists are getting laid off by the truckload, as ad networks route around them and capture the bulk of the value created.

    Perhaps I am a Luddite on this — the shift caused by technology may be inevitable…like, say, nuclear power, its pretty much impossible to put the technology genie back in the bottle, and you just have to live with the consequences. But the next time one crows about the wonderful ability of ad networks to provide targeting that supercedes the need for quality inventory, consider what you are really crowing about. Its at least a little sad.

  3. Great article.

    What we do (using our own solution) is the following:

    We help marketing teams/agencies to

    – collect a huge number of social properties where their target communities hang out (ex: teen parenting blogs, frugal people interested in finances, IT security geek, personal injury lawyers, cosmetic professional, green …) – we use our own technology to build help client build these custom lists.
    – listen, craft their messages based on what’s going on in relevant conversations
    – advertise with laser sharp relevance in these properties CROSS DIFFERENT ad networks.

    Our experience is that (beyond Google), if you want to have BOTH fine grain precision, volume and affordable CPM, you’ve to tap several ad networks. Not to mention that the way some ad network build their “vertical/behavioral” lists is not very accurate.

    Our clients build campaigns that cross Google Adwords placement targeting, BlogAd, Federated media aso, test with low budget and move with premium networks when it makes sense.

    Doing it this way, advertising also becomes the last part of the social marketing process that starts with listening, consumer insight gathering, buzz, message testing and therefore has more social relevance !

  4. You definitely *should* ask about targeting, as the technology approach does vary.

    In contextual, most networks have to arrive at and sell a limited set of categories. The more targeting you apply, the smaller the audience, so having “bigger buckets” to sell provides scale and efficiency to the sales process.

    The easiest thing for networks to do is segment their publisher base on a site basis, regardless of the content on their pages — is in the “social media” category. The problem is that more and more sites are no longer vertically oriented, they have content about everything! So if they’re selling you contextual then you want to make sure they’re segmenting their inventory on a page-by-page basis. And how are they taking each individual page and arriving at a category? Remember, the broader the category the more useless the targeting.

    For most of these networks too, behavioral simply means retargeting — where the category (likely very broad and useless as well) is made up of people who historically went to a page that fit into that category, rather than an affinity-based measurement.

    Affinity targeting allows you to reach an audience based on the level of their interest. If you’re an advertiser, you know what audience you’re trying to reach. You know what keywords are working for you on Google too. If you’re targeting exercise enthusiasts and your best-performing AdWords are “running marathon” and “cycling”, then why wouldn’t your behavioral strategy be to target an audience with a heavy affinity to those activities?

  5. When you say “Can they offer proof of why they targeted a certain page with a specific ad?” – what kind of proof would you look for? Tracking and storing ad selection results for each page view seems excessive.
    Also – targeting based on behaviors may be weighted more toward targeting the person than the page, so perhaps folks would also ask for proof of why they targeted a /person/ on a certain page.

  6. Mike, by proof I am referring to a “transparency report”. It is entirely possible for your ad networks to disclose exactly which page–down to the URL–where your ad ran, why it ran there, and what content and concepts or behaviors justify its placement. We do this in something we call a proof report which shows the advertiser or agency any and every URL, if they choose to examine them, and the justification for each placement. When you can understand performance at the page level then you can optimize to the page and drive significantly more lift in clicks or acquisitions or whatever KPI is important. Blind networks limit optimization making it a black box that you just have to trust in and that has caused the backlash. The new breed of networks can optimize and disclose to the page level and that is where you can find you can do more with less when marketing budgets are tightening in every industry.


Please enter your comment!
Please enter your name here