Brand Advertising and Google: Relevancy vs. Targeting
The recent SES Conference in San Jose included a panel called "Meet the Search Ad Networks". The panelists seemed to have broad ideas about the identity of their ad networks, which is likely why Emily White of Google recommended the panel simply be called "Meet the Ad Networks", citing the increasing mediums (i.e. radio, print, video and games) and advertisement types distributed by the companies represented on the panel.
To the dismay of the panel's moderator Danny Sullivan, who understandably wanted to keep the conversation search focused, in turned took the opportunity to expand on some panel members' comments (Emily included) to briefly address the ability of these ad networks to distribute brand advertisements across publishers/content. Although it is clearly the goal of Microsoft, Yahoo! and Google to become one-stop-shops for advertisers and it is absolutely the goal of brand advertisers to find new channels to garner some of the consumer attention that is increasingly directed online and to the infamous Long Tail/User Generated Content, it doesn't mean they are the answers to each others problems... just yet. Brand advertising does not fit into the current ad distribution models - it requires something different. Let's look at why:
Relevancy vs. Targeting
First and foremost is the issue of targeting versus relevancy. When an audience member asked if Google would continue to provide highly targeted advertising opportunities over the new mediums served, the response was "targeting has always been, and will continue to be, a basic tenet for Google". We all know that targeting is important, but the core of what Google provides is relevancy - namely relevancy of search results and relevancy of advertisements. There are infinite reasons why Google has been successful, but the most basic is that they have always focused first on relevancy. This has allowed them to capitalize on the concept that ads are actually part of the content, not simply a negative externality. Ads served with search results on Google are in some cases more valuable to the searcher than organic results. Ads served by Google on a publisher's web page add value to the publisher's content through relevancy. What exactly does relevancy mean? It means that value is added to all members of the advertising ecosystem. The user receives value, the publisher receives value and ultimately, the advertiser receives value - though none of them have complete control over the process. The question remains; how can you apply this to brand advertisements?
The proposed distribution methods for brand advertisements across vast publisher networks revolve mostly around site targeting and demographic targeting. This is because contextual relevancy/distribution does not apply to brand advertisements in the same way that it applies to transactional advertisements (something we will get into later). Targeting has a number of problems including that site targeting isn't scalable - brand advertisers have a difficult enough time negotiating buys across the current television landscape (hundreds of channels) let alone across the seemingly infinite spectrum within the long tail of content online. They are now experimenting with eBay to create a more efficient market for buying ads across those hundreds of television channels. Online, we must also consider the motivations of the content creators and how often the content changes (something else we will get into later). For these reasons (and countless others), let's assume that site targeting isn't the end game.
Demographic targeting, which aggregates information about who the content reaches and allows advertisers to target based on desired demographic reach for their brand campaign, isn't the end game either. Why not? Television does it this way, right? Wrong. Ads on television are 30 second spots between content. Ads online are part of the content. Even more so than on television, online page content ‘bleeds' onto the brand ad, and in turn the content of the brand ad ‘bleeds' onto the page. Allowing the advertiser to target gives too much and not enough control at the same time. It says ‘forget if the publisher wants this brand ad, forget if it is relevant to the content, advertisers are paying for it'. It also says ‘forget if this is going to add to or detract from the user experience, forget if it is going to interrupt the page/video/radio consuming experience'. In short it is not based on relevancy. This is not how Google has done business in the past and it would be a mistake for them to do so going forward. On the ‘not enough control' side, demographic targeting does not say to the brand advertisers that the images, sounds, art or even text put together by the site to convey an image will enhance/amplify the brands image; only that the audience of the page is the audience they would like to enhance/amplify their brand to.
Creating one market for distribution of transactional and brand advertisements
The idea is that there is one pool of inventory where ads (whether brand or transactional or some other form) can be displayed, so there needs to be one market for bidding on this inventory. The problem is that while the search companies can use their massive amounts of data to estimate an acceptable CPM bid based on what a CPC ad might generate in the same location, the value of the two cannot be measured in the same way and in many cases would represent very different uses of publisher inventory. What is needed is a separate market for brand advertisers that determines the market value for brand advertisement distribution within image enhancing content reaching particular demographics. The two markets could then be combined to determine the most efficient combination of distribution/placement of brand and transactional advertisements across publisher networks. There is a lot more to this idea with regards to optimal combinations of CPC and CPM as well as determining measurability of effectiveness, but we can go into that at another time.
Intentions of Long Tail Content Creators
The final piece to this we have already spoken about at length, so we will try to just restate it a little more clearly. The intentions of those whom create long tail content, especially social content, are not the same as the intentions of a studio creating content. Studios consider possible audiences and advertisers when creating content, because they know they have to/want to monetize the content. This is very different from the generation of long tail of content, which almost never considers advertisers when deciding what content to publish, at least in a traditional media sense. We will write more on this in another entry.
So what is the missing link between the brand advertiser's need to increase reach online, television's reach/ad value being increasingly compromised, and search ad networks' desire to attract these brand advertisers' online budgets? The link is a whole new system for determining relevancy of brand advertisements; a system which will create a market for brand advertisements and facilitate the rethinking the advertising ecosystem, taking into account brand advertisement goals as well as the intentions of the ecosystems participants and their effects on a market for brand advertisements. And what does that look like?...
Chas Edwards has a great post he pointed us to over on Chasnote (http://chasnote.com/?p=189) on brand advertising and scalability. He also pointed out that the trouble is between the bots and human recognition of context. We certainly agree, but we do think it is necessary and eventualy possible to have a scalable, technology based system; it just needs to be something different then what the options are today.
Excerpt of our response on Chas’ page:
The problem is you need the brand marketers to bid to set a price for display because each site displaying brand ads is truly unique in its value to the brand. With transactional, display on a site is worth exactly the marginal value from acquiring transactions regardless of the context (in most cases). For brand advertising display within various contexts has very different value, even if the viewing demographic is the same. And we certainly agree contextual relevancy isn't the answer. But there are other types of relevancy...