Programmable: Die menschliche Expertenrolle neu denken

With more than 50 percent of display advertising in Germany, real-time Programmatic Media Buying has taken hold with a prospect that all advertisers can only dream of: individualizing communication with each potential customer on an industrial scale and at no extra cost. Unfortunately, even though most of the technological barriers have already fallen, obstacles remain that prevent the ultimate realization of this goal.

The first obstacle is related to the lifecycle of the technology. In order to catch on and allow advertisers of any maturity level to benefit from programmatic efficiency, programmatic players had to simplify the functionalities offered. This, of course, brought with it the risk of hiding the wealth of technological capabilities. As a consequence, the performance of the more mature advertisers is limited.

The second obstacle is the proclaimed reign of "machinization." A higher and omniscient consciousness that predicts every individual behavior with minimal risk of error and locked in an opaque box. Advertisers are hostage to this dictatorship, which prevents them from personalizing their targeting, better understanding future digital customers, and constructing their own programmatic experience curve. They are, in a sense, slaves to a system in which the machine, having helped humans with their repetitive tasks, becomes an uncrossable horizon.

Inevitable repositioning of the market

Allerdings ist der Markt an einem Scheideweg angelangt: Werbetreibende haben es mehr und mehr darauf abgesehen die Effizienz ihrer Mediainvestitionen zu verbessern. Es wurden technologische Entscheidungen getroffen (DMP, Analytics…) und Profile eingestellt, die auf neue Bedürfnisse eingehen und neue Möglichkeiten erforschen. Sie sind heute reif genug, um die Kraft des RTB (Real Time Bidding) einzusetzen: Die Möglichkeit, die Kontrolle über Maschinen zu übernehmen, um ihre Marketingstrategie ohne Filter auf ihre Mediastrategie anzuwenden. Dieses scheinbar einfache Ziel kann nur mit einer ausreichend offenen und flexiblen technologischen Lösung erreicht werden, eine Lösung, die es schliesslich ermöglicht, das bereits angeeignete fachliche, wissenschaftliche und technologische Wissen aufzuwerten.

Das alles verspricht Programmable Advertising, welches mittlerweile in der US-Presse als Weiterentwicklung des Programmatic dargestellt wird. Eigentlich könnte man meinen, dass Programmable nichts Neues mit sich bringt und dass es sich nur um eine, von wenigen Akteuren unterstützte Angelegenheit handelt, aber der Paradigmenwechsel geht um einiges weiter. Es geht darum, die menschliche Expertenrolle in seiner alltäglichen Zusammenarbeit mit immer mächtiger werdenden Maschinen neu zu denken.

This new paradigm alone makes it possible to create optimal solutions because they are truly customized, as opposed to generic, algorithmic black boxes. This new type of interaction between machine and human (and no longer between human and machine) makes it possible to focus more quickly and efficiently on the same target, to have more freedom in the elaboration of targeting and buying strategies, and to obtain information on the behavior of advertising-exposed audiences.

Programmable on behalf of the performance

In der Programmable Herangehensweise, wo das Personalisierte überwiegt, interveniert der Programmatic Experte zu verschiedenen Etappen des Implementationsprozesses. Zusammen mit dem Werbetreibenden und der Agentur wird zuerst über die Herausforderungen des Werbetreibenden reflektiert: Seine Marketingzielgruppe und die am meisten unterscheidenden Attribute unter allen gesammelten und darstellbaren Daten, also familiäre Situation, Interessen, Entfernung zur Verkaufsstelle, Zeit der Ausstrahlung, und so weiter. Ein guter Marketingsinn und eine ausreichende Kenntnis der Zielgruppe sind massgebend in der Identifizierung dieser Attribute, welche oft starke Prädiktoren von Interesse, Besuch oder Kauf sind.

The result is the development of a customized algorithm based on these attributes, which allows the advertiser to personalize his target audience, his purchase price and his message. In other words, these decisive criteria form an original processing architecture of the auctions and media broadcasts of each advertiser.

Once the campaign is online, the machine takes control under the supervision of the expert and gradually enriches the algorithm. Meanwhile, the performance data is collected by detecting additional attributes and refining the original model.

This evolutionary architecture is and remains open: it allows the expert and the advertiser to analyze the campaign performance of each selected attribute, to verify the hypotheses established beforehand and to guide the decisions of the machine. Furthermore, this architecture allows to interpret the new attributes detected by the machine, which can enrich the knowledge about the advertiser's digital audience.

The Golden Age of Man and the Machine: "Hominis Ex Machina

Programmable advertising does not deny the added value of the machine, on the contrary: only machines are able to detect less intuitive micro-signals that can predict the future behavior of the Internet user. However, only a head of marketing is capable of knowing the most distinctive attributes of his target audience, some of which even the most competent machines would detect only after a long and costly learning phase.

It is the complementarity of the human experience and the calculating power of the machine that enables the extraction of Programmatic Media Buying value: this is the principle of the Programmable.

On the one hand, this maximizes the efficiency of the campaign on the part of the advertisers and, on the other, minimizes the intrusive nature of advertising on the part of Internet users. In addition, this positive cycle enables the publisher to provide high-quality content and better sell its advertising space.