USER RESEARCH • STRATEGY

How can we use technology to make smarter decisions?

How can we use technology to make smarter decisions?

TechFIIN

Introduction

Introduction

An exploratory research study around how we can use technology to make smarter decisions. Focused on people's attitudes, behaviors, and feelings towards technology that makes choices on our behalf. Resulting in a model to explain user’s different mindests when using technology to make decisions.

An exploratory research study around how we can use technology to make smarter decisions. Focused on people's attitudes, behaviors, and feelings towards technology that makes choices on our behalf. Resulting in a model to explain user’s different mindests when using technology to make decisions.

My role

My role

I planned, conducted and analyzed a one month long qualitative research project with the help and feedback from mentors.

The challenge

The challenge

Ten years ago, I walked to school listening to one of the 1000 songs I had on my iPod. Today, my Spotify subscription expands my choice to a library of 40 million songs. 

Technology has made my world more convenient and enjoyable. But it has at the same time filled it with nearly endless amounts of noise and choice.

spotgrid
Advancements in technology has given us more choice
LightSpotifyVsIpod
Advancements in technology has given us more choice

Today we make around 35 000 decisions per day (source Psychologytoday ). There is a belief, especially in the western world that a maximized amount of choice leads to maximized freedom and happiness. 

However, that is not entirely true we as humans have a limited amount of energy to make decisions each day. Every decision we make drains our energy and lowers our ability to make smart decisions. A phenomenon known as decision fatigue.

In an Israeli study, researchers examined a parole board's decision of granting or denying prisoners parole. The study showed the real consequences decision fatigue can have for people.

 

QouteLineVer
”There was a pattern to the parole board’s decisions, but it wasn’t related to the men’s ethnic backgrounds, crimes or sentences. It was all about timing.
Prisoners who appeared early in the morning received parole about 70 percent of the time, while those who appeared late in the day were paroled less than 10 percent of the time.”
”There was a pattern to the parole board’s decisions, but it wasn’t related to the men’s ethnic backgrounds, crimes or sentences. It was all about timing.
Prisoners who appeared early in the morning received parole about 70 percent of the time, while those who appeared late in the day were paroled less than 10 percent of the time.”

In the morning the parole board is well rested and can make smart and thoughtful decisions. But as the energy decreases during the day so does their ability to make effective decisions. To a parole board suffering from decision fatigue denying a prisoner parole seems like a better option as it preserves the status quo.

Internet has given us an abundance of information and choice. The increasing amount of information makes it essential to remove noise and focus on not overwhelming the user. 

We are moving towards an era of mass personalization where users will expect products and services to be tailored and adapted to them. At the same time, data and artificial intelligence enables systems to understand users and their needs on a deep level.

The approach

The approach

To focus my research I broke down my general question of “How can we use technology to make smarter decisions?” to a few key objectives I wanted to answer with my research.

Research objectives: 

• Understand users' general attitude towards assisted
decision-making technology. 

• Understand users' general attitude towards assisted decision making technology. 

• Understand how humans make decisions.

• Understand where the limits are for this kind of technology.

I decided to do interviews as my primary research method as the objectives revolved around understanding user's feelings and attitudes. I planned to recruit people from two groups; people who use technology on a daily basis and people who use technology to make decisions. 

But, when I recruited participants I realized that the categorization of the different samples was inadequate. It is hard to find people who use technology on a daily basis but do not take any form of decisions with it. Therefore, I decided to create a spectrum based on the two extremes of my interview subjects.

MaryaPaul
The spectrum of people I interviewed for the study
SpectrumMobile
The spectrum of people I interviewed for the study

On the lower end of the spectrum is Mary, she is an elderly person. Mary uses her computer at home to stay in contact with friends and family over Facebook. Mary is excited about technology and she is actively trying to see how it can help to improve her life. She is currently attending a course that helps her learn how to use her new smartphone. 

On the other side of the spectrum is Paul. Paul uses technology and digital tools to organize his private and professional life. An example of his organizational mindset is his process of finding new music through discover weekly on Spotify. 

He forces himself to listen to his Discover Weekly every week before Monday when a new list comes out. He does this to not miss any potential good songs. He adds the songs he likes to one of the three playlists depending on what kind of song it is. 

By putting the rest of the interviewees on this spectrum I got an understanding of the different profiles I had interviewed.

Paul's and Mary's name and image has been changed.

The interviews

Before conducting the interviews I designed a discussion guide that would act as the foundation during the research. I divided up the interview into 5 parts and each interview lasted for roughly 40 minutes. 

DiscussionFormat
Discussion guide I created for the interviews. Click to enlarge

In the first part of the interview, I explained the practicalities and the purpose of the interview. I cleared up any potential questions and concerns the interviewee had and asked for permission to record the interview.

In the second part, I asked the person to present themselves and talk about their life. I got to know them better by understanding who they were, their routines and their relationship to technology.

In the third part, I wanted to understand how they find and consume content online. I asked what technology and platforms they use to find new content. Did the recommendations and suggestions from the system feel tailored to them? My goal was to understand how they value the recommendation and suggestions from a system.

In the fourth part I wanted to understand how people make decisions and the part technology plays into the decision process. To understand a person's decision process I used two different scenarios. I asked half of the people I interviewed about the latest product they bought online. Then I asked the other half about their latest trip and how they found and booked accommodation. What was the process to make the decision? What sources and opinions did they consider? How did they use technology to help them?

The final part was to understand people's fears associated with this type of technology. Where is the limit? Are people afraid of technology taking over their whole life? I summarized the main talking points and asked if there was something else they wanted to add.

When all the interviews were done I transcribed them and put them on the wall using post-its. I began to cluster the information, first by generic groups like pains and gains. But, as I continued to work with the information more specific clusters and patterns started to emerge.

Photo1x
Clustering the information to see patterns and insights

The discovery

I realized that people have a lot of contradicting feelings around technology that make decisions on user's behalf. The insights I gathered could be divided into three different levels.

Mindset - general mindset around life and technology.
Feelings - feelings and thoughts around decision making.
Recommendations - ranking and evaluating recommendations and suggestions from different sources.

By synthesizing the insight I understood why users had these contradicting feelings. The findings helped me build a model for how personalization needs to differ depending on the current mindset of the user.

 

MindsetTaggFin

People like familiarity because it makes them feel safe. But they also want variety to grow and advance. 

There were a lot of mixed feelings about technology that helps people make decisions.  People are excited about technology being able to help them in their daily life. But, at the same time, there is a strong fear that it could take over their whole life.

QouteMind1

 

MindsetTaggFin

People want to understand what problems a product or service tries to solve.

For a suggestion service to not to feel like it controls the user, it has to communicate the problem it is trying to solve.

Mind2FinQoute

 

FeelingFIntag

People need an overview of the relevant alternatives available to them to be comfortable that a decision is right for them.

Users want the ability to choose between all relevant alternatives. It makes them feel safe in that a choice is right for them. However, too many alternatives make it easy for users to doubt their decision as a better alternative might be out there.

Qoutefeelingfin

 

FeelingFIntag

People rate to improve the community or tool. However, rating experiences interacting with humans is hard as it feels like you are judging the person.

People have more motivation to rate a choice in the system if they know that it will improve their experience or the community. Rating an experience where they have interacted with a human is though as it feels like they rate a person rather than the experience itself.

Feelings23z

 

RecommendationTagFIIN

Presenting recommendations with motivation and a clear context gives people trust and a better ability to evaluate the suggestion.

People evaluate a decision based on how good it matched the criteria they set up before making a suggestion. Presenting suggestions in a context gives people a better ability to understand and evaluate its relevance.

Rec1Qoute

 

RecommendationTagFIIN

Knowing the source of a recommendation helps people to evaluate if it is trustworthy or not.

People want recommendations from people they trust. It allows them to evaluate if the other person's taste is similar to theirs.

Knowing the source of a recommendation helps people to evaluate if it is trustworthy or not.

People want recommendations from people they trust. It allows them to evaluate if the other person's taste is similar to theirs.

RecQoute2

 

RecommendationTagFIIN

People feel like suggestion technology can inspire them with their suggestions.

People have different behaviors using Netflix compared to SVT Play (Swedish streaming service). The suggestions on SVT Play are unlike Netflix not tailored to a specific user. 

When people use SVT Play they usually enter the service with a goal in mind of what they want to consume. When people use Netflix they trust that they can find something interesting among its suggestions. 

People feel like suggestion technology can inspire them with their suggestions.

People have different behaviors using Netflix compared to SVT Play (Swedish streaming service). The suggestions on SVT Play are unlike Netflix not tailored to a specific user. 

When people use SVT Play they usually enter the service with a goal in mind of what they want to consume. When people use Netflix they trust that they can find something interesting among the suggestions. 

Rec3QOuote

Summary

Mindset
Feelings
Recommendations

Summary

Mindset Copy
Feelings Copy
Recommendations Copy

Synthesis

During the research, there were two reappearing subjects I wanted to explore deeper. The importance of selection criteria and people's mixed emotions around decision technology.

People have a very clear goal or motivation for why they want or buy a specific product. For example, " I need Bank ID on my phone” or "this camera has a better lens”. 

I also noticed that people enjoyed filtering products by criteria during the decision making process. This made me understand that a decision is not only about selecting the best product. Part of the decision process is to decide what criteria to prioritize in the decision. 

As previously mentioned I saw a lot of contradicting feelings around decision making technology. Users are excited about technology's ability to streamline their day and simplify tasks and routines. However, there is a fear that technology will remove variety and make it harder to experience new things.


The matrix
I wanted to better understand the contradicting thoughts the user had of making decisions using technology. Therefore, I created a matrix to better illustrate the different mindsets a user can be in when using this kind of technology.

The x-axis shows the width of the decision criteria by which the user is evaluating different options. For example, using Netflix with the goal of “I want to see a movie with Adam Sandler” is a much more specific criteria than “I want to be entertained”. 

Meanwhile, the y-axis illustrates how familiar the user is with the content and how open the user is to consume something new.

During the research, there were two reappearing subjects I wanted to explore deeper. The importance of selection criteria and people's mixed emotions around decision technology.

People have a very clear goal or motivation for why they want or buy a specific product. For example, " I need Bank ID on my phone” or "this camera has a better lens”. 

I also noticed that people enjoyed filtering products by criteria during the decision making process. This made me understand that a decision is not only about selecting the best product. Part of the decision process is to decide what criteria to prioritize in the decision. 

As previously mentioned I saw a lot of contradicting feelings around decision making technology. Users are excited about technology's ability to streamline their day and simplify tasks and routines. However, there is a fear that technology will remove variety and make it harder to experience new things.

The matrix
I wanted to better understand the contradicting thoughts the user had of making decisions using technology. Therefore, I created a matrix to better illustrate the different mindsets a user can be in when using this kind of technology.

The x-axis shows the width of the decision criteria by which the user is evaluating different options. For example, using Netflix with the goal of “I want to see a movie with Adam Sandler” is a much more specific criteria than “I want to be entertained”. 

Meanwhile, the y-axis illustrates how familiar the user is with the content and how open the user is to consume something new.

MindsetMatrix
Matrix of mindsets, x-axis defines the width of the selection criteria and y-axis the familiarity of the content
Matrix of mindsets, x-axis defines the width of the selection criteria and y-axis the familiarity of the content. Click to enlarge

There are four different mindsets a user can be in; Explorer, Specific, Vague, and Freedom. The user is not fixed to a specific mindset but moves between the different mindsets depending on the mood.

How could this matrix be used by companies to offer better personalization services to customers? To create a tangible example I used the matrix to map out Spotify’s different personalized playlists. 

SpotifyMatrixFIn
Matrix applied to Spotify’s personalized playlists 
Matrix applied to Spotify’s personalized playlists. Click to enlarge

Mindsets

Explorer - ”I want to explore…”
The user wants to consume something new and has a clear idea of the style the content should be in. The user has a clear reference point in mind and all suggestions should relate to that reference point. The user accepts wrong suggestions as long as there is a clear connection to the original reference point.

In Spotify, a user in the explorer mindset consumes the daily mixes. There are different mixes depending on the type of music the user is looking for. The thumbnail of each daily mix shows the artists the mix is based on creating a connection to the user's reference point.

Specific - ”I want this…” 
Specific is the most fixed mindset out of the four. The user is familiar with the content and determined to find exactly that. This is not the place for the system to present recommendations or suggestions. Instead, the system should focus on helping the user find the content as fast as possible.

In Spotify, this is equal to searching for a specific artist, album or song. It could also be a custom playlist where the user is familiar with the content. 


Vague - ”I want something like…”
The vague mindset has a lot of similar traits as the specific mindset. The user is familiar with the content and knows what to look for.  However, the reference point is wider and less defined. The user is open for suggestions from the system but it has to be content that the user is familiar with.

In Spotify, playlists such as Your time Capsule and Your summer Rewind caters to the vague mindset. The content reminds the user of a feeling or time but it does not specify the type of music it contains.

Freedom - ”I want to feel…”
The freedom mindset is the widest of the four. The user is looking for new content and inspiration. The selection criteria are wide and the acceptance of error is high. The user is very open for suggestions and recommendations as long as it is in line with their taste. 

In Spotify, Discover weekly caters to the freedom mindset. 

Conclusion

In the research I understood that people have contradicting thoughts around decision making technology. Synthesizing the information I realized it is because their lives and choices are complex. 

Users' preferences and needs change depending on the time and context they are in. Products that offer suggestions and recommendations will need to adapt their offering based on the mindset of the user. Otherwise, companies run the risk of making the user feel trapped or controlled.

Designed by Gustav Dybeck

Nuura