Header Ads

Predictive Behavior in Product Development [Mind the Product]


Imagine a world where predictive applications take control of our day-to-day lives.

I currently use a suite of apps that take care of most of my social media accounts separately, and I have an additional application that connects everything together.

I watched a video from MIT research fellow and cyborg anthropologist Amber Case on calm technology, and it caused me to pause and reflect on the products I build and the value they add to my life. Technology should not add stress to our day with constant notifications and unnecessary interactions. Amber continues her view on calm technology furthermore expressing concern with applications that are automatically set to notify the user with updates and popups. I’ve found there are apps – particularly those that I pay to subscribe to – which continually notify me with updates and popups and make it difficult for me to turn off these alerts or edit my preferences. I understand that these apps need to make money, but surely there needs to be a balance. A simple fix would be to ask the user for their preferences on installation and allow them to be disabled with ease.

I imagine in the near future, products using past behavior to anticipate everything. The example I give below is an extreme use case for this type of feature implementation.


A Potential Use Case

All my applications would communicate with an artificial intelligence (AI) bot to create a new event for a party. This AI bot would then choose a time when all potential attendees have an opening in their schedule. Then it would find a location that meets all criteria, invite the same people who attended the last event, especially relevant have the ability to add new friends to the group. It would then provide food according to everyone’s likes and allergies, and choose music that everyone enjoys.

Through using historical behavior to think for us with the applications we use today, everyone would already know everything. Our individual jobs, feelings about our bosses, family affairs, and vacation pictures for the past year are already known. Everyone in our lives already knows our religion and political preferences. The past few events I attended were mostly with people that agree with my views and actively follow my every move on social media.

Personal Example

I was at a party recently where I began to see the future I just described. We were all around the kitchen island, as a result talking and catching up, and I noticed a common theme. There were no new stories to talk about; everyone had already seen the updates on social media. Naturally, the conversation automatically shifted to stories with the theme of “Remember when.” These stories are interesting since they normally go back to a time everyone shared together without posting on social media.

Technology did not help that conversation since pictures or video of that moment did not exist. I selfishly place a higher value on stories of this kind, because those stories belong to all the times my friends recreated that moment with storytelling.


Perspective

Using predictive behavior to enhance product applications should not take away the human experience, but rather increase the productivity for the user. It should decrease the number of times a user repeats simple tasks. Predictive behavior should be used to make accurate suggestions. Remember that humans change their minds often and products should allow them to do so easily.

I personally use predictive behavior logic in the applications I build, but I find that I need constant reminders to make sure it adds value to the user’s experience. The first way I use this technology is with proactive user interfacing. Based on historical usage of the application, the different screens change and adjust to put the most used features right in front. The second way I use predictive and historical behavior is in analytics. We use certain metrics to understand how users interact with the application. Historical behavior, business feedback, and user interviews can be collected to make future product decisions.

Real World Application

I try to use predictive behavior logic in subtle ways to improve a user’s experience.  The logic should be able to focus on a single user or group of users. I try to perform a/b testing within a group of users, giving a different version of that app to half and measuring performance against the same metrics mentioned above. I have recently been experimenting with adding variance to predictive behavior logic. The difference here is in the calculation. We build in product feature variables that have room to grow and expand. In order to do this, the product needs data. All features need metadata and themes that could be plugged into predictive behavior reference matrices.


Case-

  • Instead of: If User A performs X, then next time the product should do Y.
  • Allow the following: If User A performs X, allow the product to offer Y or Z.
  • An example is Netflix. However, I personally cannot verify if their predictive algorithm allows for variance.

I have yet to see a product that offers a feature as my extreme event creation described above, but if you consider what is published online every day, it could be possible in the near future.


References

Amber Case- Calm Technology

The post Predictive Behavior in Product Development appeared first on Mind the Product.


Source: Mind the Product http://ift.tt/2rEFVnU
Powered by Blogger.
// check for file exists steve added below // check for file exists steve added above