Just In Time Nudging

Just in time nudging can be powerful and dangerous if used for the wrong reason.

Have you ever wondered how amazon always seems to know what you want to order? The way those recommendations just pop up after you purchase an item and it is a must have product? The same thing happens when you think of a person and suddenly they appear on your Facebook list of people you may know. Type only two characters on google search and it will complete it with the full query you were thinking about.

This is the state of technology today. It is powerful enough to kind of predict our next move.

This is not necessarily a bad thing. It is exactly how I imagined computers would work when I was a child. When computers succeed at guessing what we want, we notice it. When they don't, we simply ignore them. How many time have you ignored the list of people you may know on facebook? Or all the other suggestions from Google?

But that innocent looking list of recommendation is a powerful tool in disguise.

Internet Suggestions


The idea in the movie "Inception" is to influence a person's decision by implanting an idea in their mind at the subconscious level, in their sleep. When that person wakes up, she is more likely to make the same decision as the one suggested in the dream.

In the real world, influencing decisions is not so dramatic. In fact, it is so subtle that it might have even happened to you multiple times today.

When you head to a grocery store to do your shopping for the week, your intention is to buy groceries, or essentials. Check items off your list one by one and the next thing you know your cart is full. Now that you are done you roll to the cash register and you are presented with the most flashy products in all the store. Candies, chocolate, tabloids. All the sort of things that will attract your attention. Chances are you will grab something that was not in your list. You have been influenced to buy something you had no intention of buying.

A similar thing happens online. Here the influence is disguised as an innocent looking recommendation list. Amazon shows that "People who bought this laptop also bought a laptop bag with it". Of course it makes sense. Are you going to be the person who owns a laptop but not the bag? Add to cart!

Reading these two examples make it seem like this is always a bad thing. Yes, it is a bad thing to exploit people in this manner. But the same tool can be used for the greater good.

Stackoverflow is one of the great examples. Anyone can ask a programming question on stackoverflow and anyone can answer it. While this opens the door to a flood of bad behavior in other online communities, stackoverflow is where you will find it the least.

People who get an error after a single attempt at programming automatically try to ask questions without doing any research. To prevent this, before they submit their question the first thing they will see is a list of similar questions people have asked in the past, with answers. Chances are they can find their answer without ever posting their own question.

If they post a not so well written question anyway and another user decides to tell them how stupid they are, this user is also presented with a message right before he submits his answer. It will be something in the line of "Criticize the question, not the person". This Just In Time reminder prevents bad behavior before it ever occurs.

Another example, where the tool is used for good, is the private social network Nextdoor, an app that allows neighbors to communicate. With the app popularity soaring, users started using it for something other than what it was intended for. They used it for racial profiling. People of different ethnicity were constantly being reported as suspicious for simply being present in a neighborhood.

The solution to this problem was not a large protest in the streets of suburbia. Instead it was a little nudge disguised as a Just In Time reminder. When users are about to report a suspicious activity, they are reminded not to report someone simply because of the color of their skin. Instead they are asked to describe the person (what they wear, color of their clothes, height, etc) and the suspicious activity. Because of this small change, racial profiling has dropped by 75%.

The voice in your pocket

The emergence of technology outside the confinement of a store or a desktop computer, but into our pockets makes us subject to these influences more frequently. You don't need to be on amazon.com to see the list of recommendation. You don't even need to be shopping. The app can ring at any moment it thinks you are more susceptible to buy.

If Google notices that it does not have pictures of the interior of a particular restaurant, all they have to do is send a suggestion to people who are currently there. A notification pops on your phone and shows you a frowny face because there are no pictures of the restaurant. Now you want to turn that frown into a smile. So you turn on your camera, snap a few pictures, and upload them to their servers.

We know that people are more likely to do something that their friends also do. So facebook, can simply tell you that your friends bought something just to influence you to buy the same thing.

For example, if a user were to like Starbucks' page on Facebook and the coffee shop chain paid to advertise a new variety of coffee on the social network then that 'like' could reappear on friends' news feeds next to a Related Post promoting the new product.

I'm not surprised when Facebook or Google say they do not listen to our conversations. It might be true. But that is only because they already know everything else we do and can nudge us at any moment. Hearing our conversations is not as important to them.

Who is making those nudges.

It’s easy to say we should be angry at Google, Facebook, or Amazon. But we are also the ones praising them every time there is news about Machine Learning or an AI triumph. All these tools are powered by our data, collected through every interaction we make on the web. This data is then used to feed machine learning algorithms that will classify us in buckets.

This classification is sometimes used by programs like Alpha Go, or WaveNet voice generator, which is great. But the bulk of the data is used by advertisers to precisely sell us stuff. But what do we expect from for-profit companies to do with our data other than increase profit?

Remember, this is not as obvious as it used to be. Back in the days (very few years ago), you would see an advertisement following you around from website to website until you buy the product. Now, the advertiser can figure out the exact moment you are most susceptible to make a decision and show it to you right then with the most compelling words you are more likely to respond to. They know this because they studied your behavior for an extended amount of time. The machine learner has been fed all this data and can model your weak points that you are not even aware of.

Unlike a grocery store that will put random candies at the check out line to attract your attention, machine learning and AI can nudge you to go to the store and tell you the exact brand of candy you will get, based on the time of the day and your relationship status.


The price of connecting all of humanity is having little devices in our pocket. Although they can connect us, they also have the power to influence our decisions or suggest something new all together.

They can use their powers for the greater good, but also for their own personal agenda. We can't simply leave them at home because we are addicted to our phones.

The fight for our attention is real and the least we can do is recognize it so we can figure out what to do about it.

The real solution is to... hold on. My watch is buzzing, it is time for me to exercise.


There are no comments added yet.

Let's hear your thoughts

For my eyes only