Privacy & Sharing: Attitudes and Behaviors

A study on what current Tile users think of potential in-app location sharing features.

Project Overview

Goal: To learn about and understand the attitudes and behaviors of current Tile users regarding privacy, including the sharing of personal information. What motivations and challenges do people face in the arena of privacy? How do they feel about privacy?

Background: The Tile app allows people to see the location of their important belongings and share that location with friends and family. Tile wanted to find out if their current users would be interested in sharing their personal location with each other as well. I was brought in to develop a research structure, interview users, extrapolate themes, write and conduct large-scale surveys, and present my findings regarding privacy and personal location sharing.

Team: I acted as a research-team-of-one for this project. My work was overseen by Tile’s Head of Design, Winnie Wong. Four product teams were invested in the research as well, so their Product Managers were my primary stakeholders.

Challenges: The nature of “privacy” as a potentially difficult to discuss topic. The absence of any prior research system or team at Tile.

Methods: A two-stage process wherein I first uncovered the habits, beliefs, and values of a small group of users, and then surveyed a statistically significant portion of our user base to determine how Tile users act and feel overall. Stage one utilized narrative interviews, empathy mapping, and 2x2 relationship diagrams to uncover qualitative insights. Stage two utilized a large online survey and data analysis of said survey to uncover quantitative insights.

Results: While personal location sharing is not generally regarded in a positive light, due to its convenient applications, 70% of current Tile users already share their personal location in some manner. Most often (80%) these people share reciprocally, i.e. they share with someone who also shares with them. Sharing one’s location with a romantic partner or a friend is the most common use, with half of all location sharing Tile users doing so. Usually (70%) these people share their location in order to coordinate meeting up or in case of an emergency.

Reflections: Being a team-of-one researcher is challenging. So is implementing a research system from scratch. While I have proven that the skills to do so, I would not necessarily pursue a solo opportunity like this again. The research itself was fascinating, the individual interviews enjoyable, and the results interesting, but the most gratifying portions of the work always came from partnering with people from the Design or Data Analytics teams.

Keeping one’s coworkers motivated with cheery doodles is an important element of collaboration.

Project Story

The first and most important step of any research project is to exactingly pin down the objective of one’s research. Because I was investigating a potential future direction for Tile, every product team at Tile would have a stake in this project — creating potential for conflict between varying goals. By practicing research methodologies even during my planning meetings with Product Managers and other stakeholders, I was able to discover & ask thoughtful questions, pull together consistent themes, and identify unifying goals between their five individual perspectives. It became clear that there were several directions new social and community-based features could head in, but if any feature were to succeed we needed an accurate understanding of our users’ feelings and habits around privacy and personal location sharing.

The Tile app allows you to find things that have “wandered off”!

Once I had established my research goal, “To learn about and understand the attitudes and behaviors of current Tile users regarding privacy, including the sharing of personal information.” I began to construct a Discussion Guide which I would use to maintain focus and spark questions in upcoming interviews with individual Tile users. This Guide was informed by my conversations with Winnie, head of Design, and our four Product Managers, and its creation allowed for additional feedback from these stakeholders — further clarifying our intended direction. As I created drafts of the Discussion Guide I shared them with the four PMs and together we went through several rounds of edits in order to ensure the discussion topics were relevant to their goals. Because each team had slightly different goals, some of my discussion topics were able to overlap in intent, while others diverged. I also explained to my stakeholders that not every interview would cover every topic, and that we also might uncover new information in the course of the interviews, but that these elements were a natural part of the interview process.

My original recruitment email was too open-ended.

Next, I requested a random selection of 500 current user emails from our Data Team. I wrote and sent an email to these 500 people asking them if they would be interested in speaking with me over a video call to share their opinions about Tile. My outreach received numerous responses and the challenge became how to schedule the individual interviews so that I could complete the sessions one after the other yet also have enough time to unpack critical Initial Impressions after each call. Leveraging the online community of user experience researchers known as Mixed Methods, I discovered that is a useful tool for team-of-one Researchers and began using it to better organize my interview scheduling.

I next conducted seven interviews via video conference calls. These were structured yet still casual conversations in which I gently guided the interviewee towards certain topics by asking open-ended questions. Our conversations covered a range of topics, from “How did you start using Tile?” to “How would you thank a stranger for finding a lost item?” and everything in between.

Something I would do differently next time is to be pickier about who I select to talk to. Selecting interview subjects was a fairly disorderly process — I believe it was because this was my first time doing so without the guidance of an instructor. My future screener surveys will be much more rigorous so I can be highly selective in who I reach out to. Overall it went extremely well, however. My cluttered schedule was always tidied up when inevitably some interviewees canceled our appointment, and the interviews themselves still yielded useful results.

Conducting an unpacking session with design team member Madison.

In order to engage co-workers in the research process some of the interviews included a partner, usually a designer or an engineer, who wanted to learn more about qualitative research and get to know an actual Tile user. This person was asked to, at first, simply take notes but also encouraged to bring up their own questions toward the end of the interview. Pair interviews are fascinating and useful — you get to watch how someone else thinks about yet another someone else’s thinking. Bringing in the unexpected element of an interview partner can help you shake yourself out of your ingrained interview habits and personal frame of reference and into seeing a new and different perspective.

After each individual interview, I would update and revise my Discussion Guide. This was necessary because I was constantly discovering better ways to discuss the topic of privacy — how a similar question could be most effectively posed with different individuals, how to ask tricky or sensitive topic questions, and how to guide the conversation while also letting it remain natural and comfortable and keeping it within our time limit.

Discussion guide, V3

Simultaneously with my user interviews, both in order to grow engagement and interest in the study amongst Tile staff and to satisfy my own curiosity, I would draw surveys on a whiteboard near the design and engineering desks. An office-wide email asked people to respond with their own opinions about privacy by marking down their names or a check mark in the category they most identified with. Anyone who responded got to select from a wide variety of Pocky as a reward. While we all learned some new things about each other though the whiteboard surveys, an unexpected perk of this informal research was the number of conversations I had with coworkers about their own privacy experiences and beliefs. These conversations helped spark new insights and ideas that were then incorporated back into the next phase of the project.

One of several coworker-outreach boards.

Once the group of seven interviews were completed, Winnie and I worked together to review my findings pull out several themes. We learned that some people are very open to sharing, others hate sharing but use it anyways, and yet others value their privacy over everything. Winnie found it surprising that people would share their locations in order to avoid the annoyance of frequent “Where are you?”-type questions, while I was surprised how heavily phone battery drain rate figured into some user’s decisions to use or not use Bluetooth and location sharing features.

Once I’d presented our mid-project status to VJ, it was time to begin the second half of the project.

Part Two of the study was to pull from the themes, beliefs, and habits identified in the individual interviews and create a large-scale survey of a significant population of current Tile users. At the time, Tile had about 1.4 million active users, and because I wanted to have a 95% confidence level with a 2% margin of error, we needed 2397 responses. Given that these sorts of surveys receive about a 5% response rate, I would need to invite 47,940 users to participate.

I asked the Data Analytics team for a list of 50,000 emails to contact, and began working on an extensive questionnaire, which we would host via SurveyMonkey. This survey asked our users if they shared their personal location, what software they used to share their personal location, who they shared their location with, why they shared it, and what thoughts they had regarding both location sharing in general and Tile’s potential offering of personal location sharing. In order to create an easy to answer survey, I taught myself SurveyMonkey’s “advanced logic” systems and set up the questions so that, for example, if someone responded that they did not share their location at all, they would not be asked who they shared their location with.

The Lean Survey Canvas was a useful tool that helped me organize and plan.

The survey was sent out over the course of a week, partially due to my own curiosity regarding which days of the week we’d receive the most responses, but also in an effort to contact people with different lifestyles checking their emails at different times. As always, compensation was offered for the user’s trouble, this time as a chance at a $100 gift card to use on Amazon. I had been asked to create a database of users for future studies and interviews, so I additionally mentioned that if one opted-in at the end of the current survey they might be contacted for future interviews which would be compensated at $50/hour.

Friends and significant others are our primary use case for location sharing.

The survey was quite successful, gathering 3954 responses instead of the hoped-for 2400. SurveyMonkey automatically provided some basic analytics and response breakdowns, but for deeper answers I again turned to our Data Science team who were incredibly helpful in parsing out my complex system of questions into solid numerical answers.

Ron’s presentation broke down the research I’d done into numbers our PMs could understand.

Finally, both I and my data team partner Ron presented our findings to the PMs and the Head of Design. It was interesting to discover how much more attention the results received from the Product Managers than had the process of uncovering them. It seems that my background in design has led me to place additional value on understanding how answers are obtained, and my presentations would have gone more smoothly had I already known this is not necessarily standard! Additionally, I came to understand that being embedded in a team from the very start is incredibly helpful in speeding up the process of deciding which questions to ask. Chasing down the true intent of a request to add this or that to a survey or interview can be challenging, both to team dynamics and in the amount of additional time and effort it takes. However, seeking to uncover the wants and needs of one’s colleagues is in many ways the same process as uncovering the reasons behind a user’s statements or actions, and so can be a bit fun and exciting, too.

Thanks for being fabulous to work with, Tile Team!

You’ve reached the end, thanks for reading! Want to head back to my portfolio?

I don’t really use my Medium account anymore :)

I don’t really use my Medium account anymore :)