Research Now SSI: Scale Matters, Accuracy Matters, Confidence Matters
Shopper Intelligence

Sunny Without a Cloud (Computing) In Sight

Users love personalization. However, brands are struggling to build consumer trust in the midst of data privacy concerns. The solution would be the creation of smart devices that that can learn, process, and deliver without the need for a computing cloud.

Editor’s Note: This post is part of our Big Ideas series, a column highlighting the innovative thinking and thought leadership at IIeX events around the world. Ofri Ben Porat will be speaking at IIeX North America (June 11-13 in Atlanta). If you like this article, you’ll LOVE IIeX North America. Click here to learn more.

Hyper-Personalization. Users no longer tolerate anything less. Impersonal content is directly related to disintegrating loyalty between consumers and brands. It’s either personalized or irrelevant, immediate or non-existent; and once the users experience hyper-personalization, they can never go back. It’s not by chance that the greatest innovators are built on the back of personalized experiences. Giants such as Facebook, Amazon, Netflix, and Spotify all have a common denominator when it comes to personalization. They are the best at collecting, analyzing and utilizing first-party data.

With the introduction of the cloud not too long ago, companies were able to process quite a lot of information about their users, much faster than previously available. These, then new, computing powers enabled the rise of third-party data. Brands were able to respond to what users proactively did. However, in reality, personalization has become a euphemism for manipulation, we are more trackable, targetable and impressionable than ever before. That type of personalization tech focuses on coercing us into outcomes that mean everything to the business and little to the consumer.

But third-party data has met its destroyer: what started out with random incidents of leaked nude photos, or email addresses of escort service subscribers, has escalated to the Facebook and Cambridge Analytica debacle and brought forth new European data regulation and augmented awareness of the public with regards to their personal data.

This doesn’t mean that brands are going to stop using data to drive engagement, nor does the public really want that. (We love it when our tech knows us, like when our in-home voice device calls us by our name and plays our favorite playlist.) What needs to happen is the reinvention of customer interaction. Brands have to thrive to anticipate customer needs, bring first mover marketing through proprietary data, engage only when it is valuable to the user, eliminate privacy concerns and build lasting loyalty.

The fact that our private data travels across the web, all the way to the cloud and back, going through endless amounts of optimized algorithms whose sole purpose is to cluster us and predict what we will engage with next, is the main problem. As consumers, we build trust with brands that we want to bring into our lives. In fact, we literally bring some of these brands straight into our bedrooms, daily, and on our smart devices. But we would like to keep that relationship private, as in not intercepted by data companies we don’t know, who “sleep around” with other brands, that we did not agree to.

One great solution that has been in development for the past couple of years (and is still in its infancy) is the ability to analyze and understand data directly on the device itself! This means that the data never leaves our device, and therefore, it can never be intercepted. For that to happen, the deep learning algorithms used in the cloud by data aggregators need to be able to operate on our devices. But running any process so heavy on a smart device could have quite serious implications on battery life and overheating. Deep learning, unfortunately, as its name suggests, is quite deep or in other words, massive. The cloud allows heavy processes to run as it happens away from the device. Google mentioned this a year ago, “Unfortunately, training is still too computationally intensive to be performed on smartphones. If models can run on a device, at the edge, they can avoid the cloud and internet all-together.”

On the other hand, cracking this process has the huge potential for hyper-personalization without threatening consumer choice and privacy. The race for this new capability of eliminating the need for the cloud is already in place. “As on-device processing becomes more powerful, and AI grows more prevalent, our future will increasingly be defined by the convergence of these two game-changing trends” (MIT Technology Review Insights).

So, don’t put your umbrellas away just yet, as the cloud is still here for now. But soon, an engine that trains, learns, and delivers entirely on a smart device—so data is never exposed, and privacy is guaranteed—is bound to be one of the most interesting developments of this time.

Please share...

Join the conversation