The Hidden Revolution

The Hidden Revolution


We constantly read and discuss stories about how artificial intelligence better copes with many tasks that were traditionally performed by managers, programmers, marketers, copywriters. One of the most striking examples I heard in Berlin at the Native Advertising Days, a conference dedicated to native advertising in all its manifestations.


The speaker of inPowered, a platform for analytics and content management, spoke about the Mazda incident. The company considered its target audience to be young men, but an analysis of the data collected over 18 months showed that their content is more consumed by women 35+. A reference case. All such stories speak about the successful use of AI in working with Big Data. In this scenario, brands can really benefit from the accumulated data: it is better to segment the audience, customize the content and, as a result, spend the marketing budget more efficiently. And all this without increasing the wage fund! A closer look shows that, after all, this range of applications is just an improvement of the existing format. We have a product, we have a message, it is important to convey it to the audience, and then measure all the stages along the chain: how many impressions there were, how many clicks, which audience gives the best response, which creative gives the best response, adjust, repeat.


Of course, the AI takes over all the work inside the TOTE model, and this really optimizes the cost of attracting customers. But can this be called a leap forward? The truly revolutionary matures gradually, it is changing the basis of the conversion funnel, an approach to what is considered a target action.


According to Chartbeat Analytics, two out of three sessions that came from native ads last less than 15 seconds. This means that 66% of the budget is wasted. If we take into account that by 2020, according to Adyoulike forecasts, the global native advertising industry will be $ 85 billion, this is a huge amount. The solution that is becoming increasingly popular in the professional environment is to put engagement above the number of clicks, change the ratio of the weights of these parameters. The same inPowered introduced a 15-second session limit into its analytical model. Engagement instead of clicks, the economy of attention-that’s what changes the paradigm itself.



Yandex.Zen takes into account read-throughs, not just views; Google provides advertisers with statistics on viewing time in campaigns with In-Stream ads and splash ads. The Danish startup biites considers its key achievement not the number of users, but the fact that they watch 4 minutes of branded content on average, as opposed to 6 seconds in FB.


Why is this a revolution? Because this creates a completely new type of base for evaluating the quality of content, automatically eliminating the errors of CPA-oriented AI systems: clickbait headers and non-targeted traffic. And then-more. Already, the “smart” algorithm of social networks ranks content, predicting what you will like more (and what will hold your attention, the most valuable currency of the digital economy). But so far, these forecasts do not always hit the bull’s-eye.


You post on your FB page that you have received an international award with a photo from the ceremony, the AI of the social network sees that the entry has received more than 10% of the reactions from the covered audience, and recommends that you… post more photos. Focusing on the attention of users will allow the AI to learn how to track effectiveness at the level of meanings and messages. To do this, he will need to make friends not only with Big Data, but also with machine vision. We are already seeing the first steps in this direction: for example, the Amazon Rekognition service can recognize object, person, text, scenes and actions, as well as detect inappropriate content. Youscan can find brand mentions in social media, even if its name is not mentioned in the text — the logo on the published picture is enough.

& #13;
In the beautiful marketing of the future, AI will not only know what the photo format gets more views, than text, but also to understand what the plot should be in this photo, what kind of hero should broadcast a communication message and how. That’s when the creators of branded content-marketers, copywriters, PR specialists-can really be out of work, unless a new “but”arises.


Elena Nazarova, Marketing Director of the communication group “Premikom”

Go to our cases Get a free quote