Engaging the right users, in the right place, and at the right moment.

The sentence above refers to what is nowadays called hyper-personalization – the ultimate form of personalization, which can be achieved by feeding massive amounts of data through machine learning algorithms. Combined with the omnichannel approach, we can see it as the likely next big thing in shaping customer experience.

However, before hyper-personalization successfully leaves the sales slides and starts doing the job actually, it looks like we need to overcome certain challenges and deconstruct at least some of the buzzwords flying around.

Data processing and privacy

Like how to gather and process said data, especially keeping in mind the upcoming changes in the way privacy is treated, connected to Google’s Privacy Sandbox and Apple’s SKAdNetwork. Or the distrust users have for companies to handle their personal information safely – according to the 2022 Thales Consumer Digital Trust Index Report , levels of consumer trust are pretty low across various industries, ranging from 12% for Media & Entertainment organizations to 42% for Banking & Finance.

Processing data directly on the device instead of in the cloud , which not long ago may have seemed unrealistic, has now become available to every company – and it would be questionable to ignore the possibilities.

The power of mobile and edge computing

Just consider this: your mobile devices already collect a plethora of information that could be used for advanced profiling. Smartphones come equipped with proximity sensors , GPS sensors , accelerometers , gyroscopes , and much more.

Now, rather than send all the collected data somewhere else for processing, mobile apps can analyze it locally, on the edge. Just like in the case of personalization, more data combined with AI analysis will allow for a greater degree of profiling ( hyper-profiling , to be exact). If the results of the analysis meet the predetermined conditions, the user will be displayed appropriate communication.

What does it mean in practice?

Let’s say there is a product campaign addressed to men who are at least 40 years old. We know the product is something they might need because we already know their habits from the profile. The best time for them to view it? Right after they come home from work because that’s when they are most likely to consider purchasing this particular item.

Marketing campaigns are normally sent to user segments based on statistical data: if the average 40-year-old man comes home from work at 17.30, all 40-year-old men will receive a notification at 17:45.

Now, because every man’s device can recognize the network assumed as “home”, the app can send the marketing message exactly at the optimal moment in the individual user’s routine , whether they get back home at 15.00 or at 19.00.

What else can we learn about said man with edge computing? Let’s take a closer look at his commute. Before he got home, his phone connected to a car via Bluetooth. The same car it had connected to on all the previous workdays, so we can deduce the car is likely his.

But can we be sure he is indeed the car’s owner and driver rather than a passenger getting a ride from a colleague? Given that he almost never performs any swipes on his phone’s screen during his commute, we’re almost certain that his hands are on the driving wheel. Good to know he may be interested in car insurance.

Getting to reach such insightful conclusions is not the only advantage of hyper-personalization. Apart from the users getting informed only about the things they might actually need or find useful, their greatest concern is also addressed – the data doesn’t leave their device, remaining safe from companies’ prying eyes.

This is only the beginning

At the same time, we need to remember that the future of digital experiences may be entirely different than the environment we currently operate in. What will apps look like in 10 or 20 years? Could it be that instead of many single-purpose programs, we will turn to one, multi-purpose, AI-driven application that will fulfill all our needs? Or maybe we’ll see wider incorporation of technologies like AR glasses or contact lenses with a UI?

All things considered, our current hardware already allows for hyper-personalization based on locally processed data . And definitely, I can think of at least several technologies that deliver impressive results in the field. Time will show whether the adoption rate will be instant or a long run, but one thing is certain: further development of hardware and software will be inseparably connected with the need to personalize all elements of products or services for each user.