Google already knows what you are going to do (before you even think about it).


From Uber to Amazon, more and more brands are trying to anticipate our next move before we even think about it. A strategy that can pay off, but is not without risks.

It's Saturday night and you're leaving your house to go to a restaurant. As you walk out the door, your phone warns you that it might rain at 10pm and that you should take an umbrella.

You exit the building and open the Uber app. Surprise: the address of the restaurant you are going to has already been entered. And as you start to write a text message with "this is the address of the restaurant...", your phone automatically completes it with "... 36 rue de la Paix". Amazing - but also quite worrying.

You've just experienced 3 anticipatory design experiments: the ability of certain applications to suggest information or services to you without any action on your part to request it.

Based on your location, appointments in your diary or your usage habits, Google Now, Uber or the latest version of Android (which now integrates predictive SMS) try to predict your next action, even before you have thought of it.

Significant technological developments.

It is tempting to see nothing new in this trend. After all, as Sajid Sayed of SAP notes, "all user-centred design is anticipatory, since it is based on the anticipation of a user reaction".

But the history of interface design has so far been largely centred on the "user action - machine reaction" mechanism. Several technologies are now changing this pattern.

First of all, the interconnection of APIs and the omnipresence of geolocation have considerably extended the potential sources of data: the Uber application can thus know where you are leaving from, which appointments are in your diary (since January) and what your travel habits are at that time of day.

In addition, the rise of artificial intelligence and machine learning[i] increases the effectiveness of these suggestions, by constantly checking whether they are working and adjusting if the user regularly refuses the suggestions made.

What are the benefits for brands?

If the solution is well implemented, the positive effects for the brand are numerous.

The first objective is to simplify and guide the user's choice, moving to an almost subliminal level, and to reduce what has been described as "decision fatigue": we make 35,000 decisions every day, 226 of which are based on the single question of what to eat.

By suggesting an immediate solution before a problem arises, the brand has the advantage of no longer being in competition. This is what Amazon is aiming for by offering more and more automatic restocking options, from printer cartridges to food shopping to toilet paper: to rid the user of the need to think... and reap the benefits.

For a brand like Android, the aim is different: to move away from a 'convenience' brand to a service and advice brand. On a different note, this is the same challenge that many brands face when they realise that their apps are little or not used (recent statistics show that apps lose an average of 90% of their active users within 30 days of installation). By anticipating needs, the brand becomes proactive, hoping to recreate use value and a link with the user.

These risks should not be underestimated.

But effective anticipatory design is far from easy to implement. Four risks are worth mentioning:

The risk of intrusiveness: anticipatory design relies on the ability to collect data about the user in order to propose relevant choices. While many users today understand perfectly well that their phone or applications collect information about them without their knowledge, it is a risky thing to show too much knowledge about the user's habits, privacy and location. The danger is both to generate rejection by the customer and to seriously damage the brand experience.

The risk of irrelevance: this is the corollary of the previous one. If too much knowledge is intrusive, anticipatory design carries an imperative of relevance that cannot be achieved without data collection. If Uber recommended fancy destination addresses every time, you would quickly uninstall the application.

This is a particularly good example of the paradox of privacy and personalisation, revealing the schizophrenia of the modern consumer: on the one hand, we are opposed in principle to the collection of personal data about ourselves. On the other hand, we are in favour of brands that can finely personalise their offers and retain our tastes, preferences and history.

The risk of trust: Aaron Shapiro, the father of the concept of anticipatory design, summed up the philosophy this way: "the next disruptive innovation in design and technology will be the creation of products and services that eliminate unnecessary choices from our lives by making them for us."

But are we really ready to let Google, Amazon or Apple decide for us what services and products we want to use?

Last but not least, for many uses, anticipatory design requires applications that are constantly running, looking for the slightest signals of our behaviour, constantly consuming bandwidth and battery. This is the greatest risk of anticipatory design: having a laptop that gives up the ghost at 4.30pm. And who among us is prepared to take that risk?