Images

Publié le

La reconnaissance faciale, une aubaine pour les entreprises, mais quid de notre vie privée ?

 

Aujourd’hui, un panneau publicitaire équipé de la technologie de reconnaissance faciale est capable d’identifier le genre et l’âge des passants, en scannant quasi instantanément leurs traits de visage. Mais la technologie va encore plus loin en permettant d’effectuer une analyse comportementale des individus qui passent devant le panneau publicitaire : détection des émotions et des réactions face au message, temps passé devant le panneau etc.

On comprend tout de suite l’opportunité réelle que représente, dans ce cas-ci, la reconnaissance faciale pour les annonceurs : obtenir des données à la fois démographiques et comportementales pour fournir un ciblage d’autant plus précis et personnalisé.

Les panneaux publicitaires équipés de la reconnaissance faciale ne sont qu’un exemple parmi d’autres. L’horizon des possibilités est bien plus large : la technologie permet non seulement aux entreprises de proposer de nouveaux services mais également de les vendre mieux, et tout cela en accumulant une masse toujours plus importante de données sur leurs utilisateurs.

Déjà présentes aux Etats-Unis et se développant en Europe, les technologies de reconnaissance faciale sont de plus en plus utilisées, aussi bien par les entreprises du web que par celles appartenant à des secteurs d’activité tels que l’agroalimentaire, dans leurs stratégies marketing et publicitaires.

 

« C’est incroyable tout ce que votre visage peut faire »

C’est un des nouveaux slogans lancés par Apple pour promouvoir la nouvelle fonctionnalité disponible sur l’Iphone X : Face ID. Le géant à la pomme permet maintenant à ses utilisateurs de déverrouiller leur Smartphone mais aussi d’utiliser Apple Pay grâce à la reconnaissance faciale.

Apple n’est évidemment pas seul à développer de nouveaux services basés sur les technologies de reconnaissance faciale, d’autres grands acteurs du numérique s’attèlent également à attribuer à notre visage de multiples fonctions. Facebook, grâce à son algorithme DeepFace, peut désormais informer ses utilisateurs qu’ils apparaissent sur des photos sur lesquelles ils ne sont pourtant pas tagués. Google dispose de FaceNet, fonctionnalité disponible sur Google Photos, qui classe les photos en fonction des personnes y figurant. Alibaba, la plateforme chinoise de e-commerce, propose à ses clients de valider leurs paiements AliPay grâce à la reconnaissance faciale.

Si toutes ces entreprises entendent garantir à leurs clients une plus grande sécurité (un visage se pirate moins facilement qu’un mot de passe) et un gain de temps non négligeable, leur objectif n’en est pas moins économique. Le recours à des technologies de reconnaissance faciale est un moyen de recueillir encore davantage de données que ne le font déjà les techniques courantes de tracking (identifiants, cookies etc). L’analyse des profils et comportements des utilisateurs devient ainsi bien plus précise et gagne en valeur aux yeux des annonceurs et des prestataires de services pour l’amélioration de leurs offres.

 

La clé d’un ciblage publicitaire ultra personnalisé

Grâce à la reconnaissance faciale, les entreprises améliorent leurs offres grâce à une meilleure connaissance client, ou optimisent l’expérience client en proposant de nouveaux services. Mais ce n’est pas le seul atout de cette technologie qui est également la clé d’un ciblage publicitaire ultra personnalisé.

La reconnaissance faciale a déjà commencé à investir plusieurs secteurs d’activité, allant de l’agroalimentaire au prêt-à-porter. Par exemple, Coca-Cola ou encore la BBC mesurent par webcam les expressions faciales d’individus volontaires face aux contenus publicitaires. Ou encore, Adidas prévoit d’installer dans ses magasins des murs digitaux capables d’identifier l’âge et le sexe des clients afin d’afficher les produits les plus adaptés, dans le but d’aider les vendeurs dans leurs démarches commerciales.

Ces initiatives témoignent d’une volonté croissante des annonceurs de maximiser leurs profits et de rentabiliser leurs investissements publicitaires. Conscients de cela, de nombreux acteurs développent justement des technologies de reconnaissance faciale dans le but de mesurer et qualifier l’audience des campagnes plus précisément, en particulier des campagnes outdoor. En effet, l’enjeu apparaît d’autant plus important que, contrairement aux campagnes publicitaires online, la performance de la publicité digital out of home ne peut être mesurée à l’aide d’indicateurs objectifs tels que le nombre de clics.

Parmi les acteurs du développement de la reconnaissance faciale au service du ciblage publicitaire outdoor, l’on peut citer la startup française Jacare Technologies. Son concept se traduit par l’intégration dans un panneau digital d’une caméra capable de détecter si la personne a vu le panneau et pendant combien de temps. L’annonceur peut alors connaître le temps d’attention et l’intérêt du consommateur pour la publicité. De plus, le logiciel est capable de qualifier l’audience en termes d’âge et de sexe, dans le but d’afficher en temps réel des publicités ciblées en fonction de ces caractéristiques. Ainsi, l’annonceur paiera sa campagne en fonction du nombre de fois où sa publicité a été affichée, sachant qu’une même publicité ne pourra être diffusée deux fois à la même personne, que la caméra reconnaîtra. Cet exemple permet donc de comprendre aisément l’intérêt que peuvent éprouver les annonceurs pour la reconnaissance faciale.

Au-delà de l’analyse de l’âge et du genre, des solutions de reconnaissance faciale tentent d’aller encore plus loin en proposant de détecter les émotions des individus : afficher une publicité selon l’humeur de la personne (si elle sourit ou non) ou encore faire payer chaque rire à ses spectateurs (initiative du Teatreuneu de Barcelone).

La reconnaissance faciale, fin du droit à l’anonymat ?

S’il est évident que la reconnaissance faciale représente de réelles opportunités pour les entreprises d’optimiser la connaissance et l’expérience client mais aussi leurs investissements publicitaires, il l’est également que le recours à cette technologie suscite d’importantes réactions chez les consommateurs, au nom du droit à l’anonymat. En effet, il est certes possible de laisser son Smartphone chez soi (pour éviter la géolocalisation et le ciblage trop personnalisé), mais son visage, c’est plus compliqué…

Les individus pouvant être identifiés partout et suivis en temps réel, la frontière entre la vie offline et celle online devient alors de plus en plus fine. Grâce à la reconnaissance faciale, les entreprises sont capables de faire le lien entre une identité numérique (un compte Facebook, une adresse IP etc) et un « individu réel » (une personne unique par son apparence). Autrement-dit, elles peuvent associer nos activités sur le web à nos activités dans le monde physique.

De ce point de vue, la définition (utopique) d’Internet comme un espace d’expression libre garantie par l’anonymat des individus apparaît clairement obsolète. En particulier, lorsque des technologies de reconnaissance faciale d’acteurs privés sont utilisées par des acteurs publics. L’on pense notamment à l’application FindFace, permettant de retrouver, à partir d’une photo, le profil de quelqu’un sur Vkontakte (le Facebook russe). NtechLab, développeur de l’application, a démarré une collaboration avec le gouvernement russe pour équiper les caméras de surveillance des grandes villes du pays de son algorithme de reconnaissance faciale.

Rappelons toutefois qu’actuellement de telles pratiques de surveillance des individus ne sont pas encore rendues possibles par le législateur européen qui impose strictement aux entreprises d’anonymiser les données utilisateurs, et interdit dans la même perspective la mise à disposition du DeepFace de Facebook et du FaceNet de Google sur son territoire.

Même si le RGPD, entrant en vigueur en mai 2018, a pour objet de renforcer la protection de la vie privée des individus, ses dispositions concernant la reconnaissance faciale se verront inévitablement bientôt dépassées par les évolutions de cette technologie. Il n’en sera alors que plus urgent de réaborder les questions de réglementation du droit à l’anonymat, du droit « à la vie offline ».

 

Hanneke Deterne

Publié le

Amazon to Instagram: Deja vu? The Reality of Retargeting and the Future of Privacy

Amazon to Instagram: Deja vu? The Reality of Retargeting and the Future of Privacy

Now: Have you ever been scrolling through your news feed on Instagram and you found an advertisement that looked eerily familiar? In the spring of 2017, I was taking part in a group project for my marketing class at the University of Richmond in Richmond, Virginia, United States. Our assignment was to come up with a product and develop an original marketing plan. We decided on an electronic door key called “Keyless”. Anyways, while conducting research for our pricing strategy, I searched the prices of similar products on Amazon. Long behold, I was scrolling through my Instagram a few hours later and voila! There was an Amazon advertisement on Instagram for the very product I previously looked up. At first, I was dumbfounded and naive, thinking that this was some crazy coincidence. For a few moments, I thought to myself “what are the chances that the products I looked at on Amazon would be advertised to me on Instagram?” However, I quickly realized that this couldn’t be a coincidence. I concluded that my Instagram and Amazon accounts must be linked in some capacity. This was an “Aha moment” that reminded me of the digital world that we live in, where so much of our personal information is shared between various networks. This scenario was the perfect example of how my information could be used by different sources to profit. The following marketing class, I mentioned this occurrence to my professor who told me that this happens even more often via the Amazon Echo. Moreover, there have been situations in which a product that someone has recently discussed out loud appears on his or her device in the form of an advertisement. This phenomenon is known as “retargeting”, which is when online advertisements are targeted to someone according to the target person’s online activity. Retargeting is not exclusive to Amazon and Instagram as it appears on many other platforms including Facebook. This implies that our devices are not only tracking our every search or click, but also that our devices are listening to what we are saying.


With advancements in online technology comes increased capabilities for many people, especially advertisers. Not only do online advertisers have a whole new way to reach their audience, but they can reach them with products that have been of interest to Instagram users. From Instagram’s perspective, there is a critical advantage and disadvantage of allowing advertisers to retarget to Instagram users. While retargeting could be highly lucrative as advertisers pay Instagram for ad space, it could also carries the risk of making Instagram users uncomfortable. Furthermore, when seeing a recently searched Amazon item in an Instagram advertisement, Instagram users could feel that their privacy has been invaded. As the platform for individuals to share pictures of their life with one another, Instagram has a brand image to maintain. Individuals, who only use Instagram to view pictures posted by their loved ones, could be turned off by precisely retargeted advertisements. These people want to feel safe when they are viewing these pictures and not as if they are being spied on. For Instagram, the question becomes, do the benefits of advertisement dollars outweigh the potential costs of upset users? In order to effectively understand the reasons for retargeting on Instagram, it is essential to look at the numbers provided by Facebook, which acquired Instagram in 2012 for $1billion. According to Facebook, when presented with a product post on the app, 60% of Instagram users see new products while 75% of users take action. Although the 60% figure was in regards to finding new products in advertisements, this could still be seen as retargeting. For example, an individual might have been browsing Nike.com for a specific Nike Lebron James basketball sneaker, but decided to not purchase the product due to its high price. In turn, Nike chooses to retarget a similar, yet more affordable basketball sneaker to the same consumer. Nike pays Instagram to display the advertisement and Nike could profit if the consumer makes the purchase. Additionally, Nike is able to see that their customer pays attention to social media ads and that the customer prefers cheaper purchases.
Instagram provides companies with so much information that could be valuable for effective advertising. If a company has an Instagram page, it could potentially monitor the accounts and topics that its own followers are interested in. For example, Nike might monitor its followers and see that many of them are also following Neymar Jr., the elite soccer player. In turn, Nike could pay Neymar Jr. to appear in a Nike advertisement that will be directly targeted to those mutual followers. Combining this data with retargeting, Neymar could appear in Nike advertisements for a product that those mutual followers displayed interest in online but never purchased.

The near future: If Instagram users feel that their privacy is already being invaded, they might feel even more uncomfortable in the near future. Apple recently introduced the iPhone X, featuring Facial Expression Recognition Software, which is marketed as a security measure. However, it would only make sense that the data provided by this software ultimately becomes a valuable item that Apple could sell to marketing researchers. Furthermore, this software could enable marketing researchers to actually observe user facial expression and therefore analyze how consumers felt when presented with a specific advertisement for a product or service. Specifically, advertisers could purchase the data containing the facial expression of Instagram users when presented with a specific advertisement. This would help companies’ marketing teams determine how consumers truly feel about their products, pricing, promotion, and even place. This would be much more effective than having biased consumers tell you how they feel. For example, if Apple has data showing that a Facebook user smiled when viewing an iPhone X advertisement, Apple knows that the person is interested in the promotion and potentially the product. In turn, Apple could retarget the iPhone X to the same individual by showing the same exact promotion on that user’s Instagram account. All the while, Instagram is paid for the advertisement. Since Facebook owns Instagram, it profits from advertising on both platforms.
The last thing an advertiser wants to do is upset consumers. Although Facial Expression Recognition Software could enable advertisers to better target consumers with material that is more relevant, it could also make consumers extremely uncomfortable. Ultimately, social media platforms like Facebook and Instagram are free because we provide value in our personal data. Therefore, if users become increasingly paranoid about their privacy being jeopardized, social media companies could face an uphill battle. However, it would be very interesting if platforms like Facebook or Instagram estimated the value of each user’s data. In turn, Facebook could introduce an option for users to pay a fee for privacy equal to the value of forgone personal data that could no longer be sold by the company.
Personally, I believe our data has been available to social media platforms and big companies for many years now and that we will only become more exposed. As someone who has nothing to hide and thinks before making a purchase, I do not feel threatened by enhanced methods of retargeting. From my perspective, we get to use these platforms for free and companies are therefore able to profit off of the information that we make available. That seems reasonable. Moreover, both sides benefit as users enjoy these free platforms, which profit off of our tendencies. Ultimately, I think there is a fragile boundary that advertisers must not cross. In other words, there is a fine line between retargeting consumers with the perfect, most relevant products and intrusion of privacy. However, some consumers will feel these two things simultaneously occurring.
Consumers could one day find themselves living in a world where companies are omniscient. On the other hand, it is possible that we are already living in such a world.

Samuel Altman

Publié le

What about the garage gig, Facebook and Google?

Clémence Fonville (exchange student fall semester)

 

We have seen throughout the course that the advertising industry is going through some fundamental changes and that the commonly used standards are now open for discussion. Mainly, the invasion of tech firms into the ad industry is causing these changes. The key players in online advertising are not the traditional WPP or Publicis anymore, it is Facebook and Google who are setting the rules of the game. Especially, the recent advertising festival in Cannes has demonstrated the power of these tech firms in the ad industry. But while the power is shifting to tech firms, they are as well under threat. Google and Facebook have been openly criticized by big consumers regarding the measurement of ad effectiveness (Bond, 2017). The problem that they are facing is that these tech firm hold all the power in their hands when it comes to where to publish advertisements and how to measure the performance of these online advertisements. This means that companies pay them to publish their advertisements and at the same time rely on them to track the performance of this publishes. Hence companies are struggling to understand how objective these performance measures are and thus how much to invest to maximize their ROI.

 

It is very ironic that Facebook and Google hold the power of analyzing their own performances of linking a company to a publisher. It is as if a middleman is telling you after you have already paid him for guiding you to a clothing store, that the dress you are trying at that store is the perfect dress for you and that you will not find anything better. This middleman will then advocate that he is the best middleman to contact if you want to find the perfect dress, and no other person is able to judge this performance.  In my opinion, and I am joining that of many companies, the current practices for online advertising are not objective and there is an urgent need for change.

 

This urgent need for change is becoming more and more strongly resented in the industry. Not only because of the questionable objectivity of ad performance tracking but as well of the accusations that these tech firms are at the source of spreading fake news and hate speech. Google has very recently attempted to calm down this resentment by announcing to sharpen their YouTube policy for ad monetization. This is mainly following the bad publicity it recently got with YouTubers who have published questionable content but still were able to monetize the content through ads published alongside their videos (Litman-Navarro, 2018).

 

This action from Google follows another coming from Facebook who has announced January 11th that Facebook’s newsfeed algorithm will be changed. The change will prioritize posts that spark conversations and meaningful interactions between people and will prioritize posts from friends and family over public content (Cohen, 2018).

 

Unluckily for Facebook, this announcement has not generated an overall positive reaction, mainly brands and publishers are considered to become losers. The change in general means that brands will need to redesign their social media strategies to remain part of the conversations. And publishers will have less chances to sell their publishing space. But it does not entirely mean that these two groups will be the losers. The main downside that this change is bringing along is that it will make it much harder to connect with your fans or stand out if you are a small business. Facebook will prioritize brands that you as a user have interacted with in the past and with which your friends are interacting as well (Castillo, 2018). Besides, brands will still be able to be involved in the conversations or advertised along the conversation by paying a certain amount of money. But the prices of these advertisements will go up since the advertising space will diminish. This means that brands who have already much awareness and who have the possibility to pay the higher prices for advertising will continue to stand out. But smaller business will suffer. Before through social media they had a chance of reaching audience and consequently customers with small budgets, but now this seems to be a thing of the past.

 

As well, it seems that Google’s Youtube policy change will have the same consequence. Youtube’s Creator Blog has mentioned the following statistics regarding the policy change: “99 percent of those affected were making less than $100 per year in the last year, with 90 percent earning less than $2.50 in the last month.” This means that very unfortunately with the Google policy change also mainly the smaller YouTube channels (read: small businesses) will suffer, just like Facebook’s new newsfeed algorithm. In the the case of YouTube these small channels are paying the price of a blunder made by a well-known YouTuber who with the policy changes will still be able to monetize its YouTube channel in the future.

 

To summarize, through these changes Facebook and Google have partially reacted upon the critics they have faced publicly. They have shown to be willing to improve the online content shown on their platforms along with how I call it premiumizing the display of advertisements to their users. These changes do not lessen their power within the advertising industry and the problem of advertising performances remain. Facebook and Google will still need to reassure the public of their performance practices. This will mean providing some degree of transparency but unless forced by law to do so they will not be willing to give in their power.  And what can be concluded of today’s policy changes is, the main losers are the least powerful smaller businesses.

 

References

Bond D 2017 https://www.ft.com/content/9a9ac60a-575a-11e7-9fed-c19e2700005f

Castillo M 2018 https://www.cnbc.com/2018/01/16/facebook-news-feed-changes-not-doom-for-publishers-adam-mosseri.html

Cohen D 2018 http://www.adweek.com/digital/the-ad-communitys-reaction-to-facebooks-news-feed-algorithm-change/

Litman-Navarro K 2018, https://www.inverse.com/article/40270-youtube-partner-program-monetization-changes-logan-paul

Quitter la version mobile