Separate news from Google, Facebook and Salesforce.com this week highlight different approaches to advertising on different media. While the focus for television advertising will remain around brand-building among relatively broad audience, the Web will increasingly be about targeting new and existing customers using known first-party data whenever possible.
Google announced that its Google Fiber service would start a trial selling TV ads in its recently established Kansas City system. It would have the capacity to target based on geography (a capability that existing MVPDs already possess), type of program being shown (ditto) or viewing history (technically possible but not widely deployed). The company noted that advertisers will only pay for ads that have been shown (standard industry practice) and can limit the number of times an ad is shown to a given TV (this part would seem to be somewhat unique — the idea of frequency capping in TV ad sales).
The underlying dream of buying discrete volumes of precisely targeted TV ads persists for many. Pulling out the clichés, it’s about making TV as targeted as the web, ensuring that dog food ads only go to dog owners and allaying concerns for the advertiser who knows half of her advertising is wasted but doesn’t know which half. While Google will undoubtedly develop TV ad products that are more advanced than the industry’s state-of-the-art, as a practical matter, even if Google Fiber was a much bigger service, the most sophisticated brands would have difficulty with assigning budgets to Google for related initiatives. Beyond the limited footprints of individual MVPDs and the inconsistencies in ad product offerings in many instances, concepts associated with advanced TV are often held back because of technical inconsistencies across MVPDs and business model issues between programmers and MVPDs which make the whole process unduly laborious and unnecessarily expensive.
These very practical limitations are only part of the reason why most advanced TV advertising services have a hard time achieving meaningful scale. In general, TV is primarily used by (and is uniquely powerful for) large marketers amplifying general awareness of brand attributes. Towards these ends, most marketers know why they prefer to focus their media spending on (and negotiate against) specific demographic targets, but few will believe that these are the only audiences that matter, given the sales of products that can occur outside of a target audience and the difficult-to-identify sources of influence from out-of-target-demo audiences that might contribute to business outcomes.
For example, a luxury car maker knows that few people making under $100,000 per year will be able to afford to buy that car, but the marketer knows that the aspirations of the masses can make a product more valuable to those who can afford it. Targeting has a role in improving TV advertising and making existing TV advertising incrementally better (for example, with improved secondary audience targeting metrics across a campaign or by better balancing reach and frequency characteristics), but do not think that precision to the level of the individual will necessarily have long-term legs.
In large part, this is because an advertiser seeking to reach discrete audiences in an interactive and/or highly targeted manner already has a medium of choice to use: it’s called the Internet, and it’s not running out of inventory any time soon. Our view has been that spending on digital advertising by large brands has primarily focused on notions of “engagement,” or interactions between a brand and their desired audience segments. It capitalizes on an architecture which allows for one-to-one targeting across essentially all inventory in what is a relatively automated fashion, and allows a brand to capitalize on the lean-forward nature of the medium. These factors lead to superior conditions for targeted media trading, made all the more valuable when desirable data is appended to media inventory.
As an example, consider this week’s news that Facebook would expand its video-focused SSP LiveRail to include in-app advertising on mobile devices. Facebook-based age and gender-based data will more readily be included across digital advertising inventory well beyond Facebook’s own walls among publishers who use this product. High quality data such as that which Facebook can append to media trading is highly useful in helping marketers target audiences for engagement in digital environments, and will help Facebook continue to expand the share of digital media advertising sales it captures or influences. Targeting by audience characteristics remains important to digital advertising and will likely remain a significant basis for targeting for many years to come.
However, an alternative (and possibly more valuable) targeting paradigm involves identifying segments of known customers, look-alikes of those customers and individuals who fall outside of both of those groups. This is where another product also in the news this week from Facebook called Custom Audiences comes in.
Custom Audiences was highlighted when Salesforce.com announced the general availability of an updated version of a product it offers called Active Audiences. In its new form, Active Audiences can integrate a marketer’s customer data from one of Salesforce.com’s Sales or Service Cloud products into Salesforce.com’s Marketing Cloud tools, which in turn buys ad inventory from Facebook via Custom Audiences (or from Twitter via a similar product called Tailored Audiences). This allows marketers to buy ad inventory including audiences of individuals on a highly targeted basis using known customer data rather than audiences of individuals based upon proxies of what customers are expected to look like.
Custom Audiences has been popular for marketers using Facebook, contributing to the company’s capacity to gain share-of-wallet from those who use it. Salesforce.com isn’t the only company powering the buys of Custom Audiences (they likely account for a low single digit percentage), as there are many alternative approaches to buying inventory in this manner. Alternatives are also emerging from agencies such as Merkle, which will enable advertisers to buy known customers through other publishers, applying CRM-based consumer data towards a much broader range of digital inventory.
The trend towards targeting digital advertising to engage with known customers rather than a more generalized (if still targeted) audience is an important one, it will be more valuable to advertisers than alternative approaches. While there will be new approaches to selling TV advertising to new and existing advertisers, our guess is that the relative strengths around television will remain around brand building for relatively broad audiences. Targeting on the web will increasingly be about targeting new and existing customers using known first-party data whenever possible.