In my last post, I tried to establish the notion that brand advertisers were under-represented in the digital ad world since it is so technology and math driven and therefore mysterious and difficult to understand. In short, digital ad creation is not the same as print and TV ads, and for those who create them, the usual tactile feedback of creative testing, focus groups and market research is missing. The creativity is missing – digital ads have become something for the pocket protector crowd. And the media planners and strategists are outsiders.
But what if media planners and strategists could have all the algorithms and math buried behind a user interface that restores and reveals creativity to its rightful place? That is the design approach we took here at ADmantX. That is not to say, however, that there is not some deep technology at work. There is. In fact, the best technology is often hidden, hard at work doing magical things you never see. That is the essence of good technology design.
At ADmantX, we built semantic processing power as the core technology innovation for matching content to ads. So behind the user interface, a deep analysis of every content page is happening 24X7. Literally ADmantX is reading the page just like a human reader does. It looks for people, places, things and how they are connected, colored by sentiment, negatives or other modifiers. At a higher level, it looks across content to arrive at what multiple sentences, paragraphs or the whole document is trying to convey. In essence what the author was shooting for when they wrote it.
That alone is a firm basis to achieve a superior match with ads. But it is not enough. We cracked the code on emotions, motivations and behaviors that are contained in the content as well. This is a distinguishing method on which to match ads. Even if the words are not present does the content prompt an emotion in the reader? Does it prompt the reader to do something, think something? These are the questions ADmantX answers.
But once you know all that is contained in the content, how do you engineer an algorithm to match an ad to it? That is where the ADmantX user interface shines. This is how it works:
Media planners enter the core concepts they wish their creative to convey. ADmantX understands them as feelings, topics, entities, as well as negatives. Planners can then test this algorithm against a set of content to see how well it will match. This picture is below.
Immediate feedback in terms of coverage – overall, by concept and by URL, is presented to the planner. Editing and repeating the test gives the planner a way to compare and contrast algorithms. Note that the second URL contains a negative term match on “sadness.” Brand safety concerns show up as part of the test and can be corrected as shown in the picture below.
A simple operator “!” followed by the concept to be eliminated, ensures the algorithm will not match on this negative association going forward. Note, this provides more flexibility than having a standard list of negative terms. There may be times that brands will want to be associated with “sadness” – as in the case of advertising depression medication, for example.
Building an algorithm to match ads to content with preconceived notions is only one of many ways the ADmantX user interface supports creativity. For example, allowing the news of the day, updates to website pages, blogs and RSS feeds to influence the algorithm may provide a better match. This is done by looking through a tag cloud with the suggestions link.
Iterating from broad to more specific tag clouds term by term allows creativity and the most recent content to occupy the same space.
Accessing social media is another way to tap into the “voice of the consumer” for a superior ad match. Here we show a simple paste of a tweet or blog post. ADmantX reads and interprets the social media snippet into an algorithm for the planner.
Of course, an entire page of content may be a good representation of what the planner wants to accomplish. This can be done by pasting the URL of the representative page into ADmantX and allow it to interpret the whole page. This is shown below.
Finally, crowd sourcing or social collaboration may prove a fast start to the creative process – both from within an organization or from publicly posted algorithms. Simply review by type, rank and test results – then borrow and make it your own, followed by edits and corrections. The start to this process is shown below.
Of course, any of the methods to support and test creativity can be combined with the other.
In the end, media planners, buyers and creative strategists need tools to more effectively design and test digital ads, where testing means more than testing the ad itself. It must mean the ability to know quickly and thoroughly what content the ad will be matched with. This avoids match mistakes, ensures brand safety and more. In today’s environment, these are painfully absent. We created ADmantX to begin to address these shortcomings.