In the run-up to the Future Media Lab.’s next workshop — which will take place as part of the Digital Innovators’ Summit, 25 March 2014 in Berlin — we sat down with the Director Advertising Research at IHS Technology, Daniel Knapp, to talk about innovations in social media and the impact on media advertising.

Knapp will be one of the expert speakers at the workshop, which also includes presentations from Marcel Boulogne (European Commission), Joelle Frijters (Improve Digital), Bernard Cools (Space) and Audra Martin (The Economist).

FML: Future Media Lab.
DK: Daniel Knapp

FML: How do you see the innovations in social media impacting media advertising?

DK: The most important innovation is not about advertising itself, but about its raw material - audiences. To be precise, it is social media’s ability to accumulate, retain and involve audiences in new ways despite rapid technological change and transforming consumer preferences. Social media has become ubiquitous. Not only does Facebook have 1 billion monthly active users, but the notion of ‘social’ has expanded vastly to encompass a growing range of different platforms and services. In fact, ‘sharing’ has become a shorthand for much of the activity that web users engage in, hard-coding the concept of social media into the interactional fabric of the internet.

The most important innovation that social media bring to advertising then is just this - the concept is ‘sticky’. Facebook’s acquisitions of Instagram and WhatsApp to encompass new forms of social media beyond its core platform illustrate this ability to reconfigure and reinvent the idea of social media.

This is critically important for advertising.

Firstly, in a rapidly fragmenting media world, social media provide advertisers with a rare good: scale and audience reach through a single platform.

Secondly, social media provide the fuel for advertising targeting and delivery - data of tremendous volume and variety from demographic to behavioural characteristics.

Data is growing ever-more important as the advertising industry is replacing the idea of targeting audience segments, such as the traditional TV ‘target audience’ segments of e.g. 14-49, with individual “addressability” based on personal data.

Availability of, and quality of such data has become a distinguishing characteristic for success in advertising, especially as advertisers are increasing demands for return-on-investment and press for clear evidence of effectiveness and efficiency of their investments.

To come back to Facebook, this is where I see the logic of the WhatsApp acquisition: Facebook does not want to advertise to WhatsApp users, but rather use the data which users produce on WhatsApp in order to further optimize the advertising targeting on Facebook itself.

FML: What exactly are the innovations - or should we say disruptions - the digital environment is facing in this field?

DK: The entire advertising industry is being re-architected in a way Wall Street was in the 1980s when computer-based trading desks were established that automated stock trading and placed purchasing decisions into the hands of algorithms.

Over the last three years, we have seen the same happening in advertising. Paradoxically, although more than 20% of all advertising budgets are placed on digital media, doing so has long remained very a analogue, and manual process. It often took 30 and more steps, including the use of a rather archaic technology - fax - to place a banner advertising campaign. What the industry calls ‘programmatic buying’ finally automates this process. It often coincides with ‘real-time bidding’, where the automation continues to the auctioning advertising impressions.

Whenever a user loads a website, advertisers bid on this user and the highest price wins, serving the user an ad. All this happens within milliseconds.

This has become possible through the marriage of consumer data with sophisticated algorithms and predictive models which now sit at every intersection of the advertising process, from inferring further information about users to optimizing how an advertiser should place its bids in the face of competition.

Social media is at the forefront of this, with its data fuelling these algorithmic pipes. In later 2012, Facebook launched ‘FBX’, an adverting exchange, where such bidding can take place.

FML: What do European politics have to do with these technology disruptions?

DK: Knowledge is power - this adage has not changed in the new world of algorithmic advertising. But today it is the means of arriving at this knowledge that are changing. Our entire modern society is based on the idea of causality, explaining why things are the way the are.

In the age of ‘big data’, produced also by social media, the idea of causality is replaced with correlation. Marketers today just need to know that things work, and not how. Large databases paired with immense processing powers and algorithms establish such knowledge. Those who possess the best data and the most sophisticated technology are best positioned to attract advertising budgets. This has resulted in a new landscape of media concentration.

In the US, Google and Facebook, who are at the forefront of such innovations, command in excess of 65% of all online advertising budgets. Traditional media owners, such as publishers, are losing out. The New York Times is reporting consecutive advertising revenue declines - not just in print, but also in digital advertising.

At the same time, the location of media concentration shifts from the visible, to the invisible realm of algorithms. It is here that new forms of power and control, both over competitors and consumers, emerge. And it is not just about Google and Facebook. In order to innovate, publishers are increasingly experimenting with recommendation algorithms that curate website content based on inferred consumer preferences.

These trends have profound implications not just for privacy, but also for media pluralism, and competition. But while the political and regulatory urgency to monitor and evaluate this new algorithmic landscape are evident, the tools available are blunt - how should politics, how should regulators, intervene in algorithms?