Helping Users Prepare Their Data for After Death

Time to read: 7 min
Death is a part of life. Knowing this, we write wills and set up our estates to make things easier for our loved ones when we die. Yet, very few people turn their attention to their digital data. Think about a lifetime of photos lost to iCloud or important documents locked beyond two-factor authentication. How can we help users deal with a lifetime of data?

To prevent painful and costly mistakes, companies need to evaluate and support how users can pass on or delete their data in the inevitable event of their deaths. Through this research project, I outlined the best ways to approach this issue.

Stats

The emotional side in a way, my personality traits, who I am as a person, those aren’t things I type up on the internet. Those aren’t things that can be represented by photos and search history. That would be missing.

The Challenge

Have you ever wondered what happens to all of your photos, videos, messages, and digital files after you die? Unlike our physical possessions which are easier for others to deal with, our data can outlive us. Also, our digital data lives many different places — distributed across our online accounts, hard drives, or cloud servers. Lastly, the overwhelming amount of data a user has at the end of a lifetime poses a challenge. How might we help users make plans for their post-death data?

In this academic research project, I conducted generative research to understand user needs and test a variety of design concepts. Would users prefer erase their data, bequeath digital memories to their loved ones, or something else entirely? Talking about death isn’t always easy, but through this process, I learned so much from our participants and their views on death, love, and the value that our personal data can hold.

🔍 Explore media coverage on this research project: CTV, UBC CS, Built In, UBC News, Vancouver is Awesome, North Shore News, New West Record, Tri-City News, Squamish Chief, Prince George Citizen, Tech Xplore

✨ My research was financially supported by the National Sciences and Engineering Research Council of Canada.

The Research Process

The design sprint

My team and I planned a condensed version of the Google Design Sprint to generate a range of diverse ideas to explore. The design sprint participants included a mix of senior and junior researchers with expertise in HCI (human-computer interaction) and design. Through Crazy 4’s and voting, we arrived at 28 initial design concepts. We used Miro and Zoom to do this online.

From this, I created a final presentation of 12 design concepts for the interviews. Each concept had a sketch of the idea and a short description. The design concepts ranged from the mundane to the weird — from building on existing tools like Google Cloud to integrating virtual reality replicas based on a lifetime of personal data.

🔍 Find the design sprint slide deck here (lightning talks, how might we’s, crazy 4s exercise).

The user interviews

Over the course of two months, as the lead researcher, I was responsible for recruiting, screening, and interviewing 20 participants. We looked for participants with diverse ages and parental statuses. This was because we wanted to understand perspectives from people who had prepared for death in other ways (e.g. writing a will) and wanted to pass on their possessions. As we expected, older participants (65+ years old) thought differently about their data and reacted more negatively to the technology-based solutions. For example, they were less interested in gamification  of the process, and preferred tangible representations of their data to virtual.

“That would be fascinating actually… if I’m getting something out of it, I can see more of the value of passing data onto my kids”

💭 Reflection

Due to the COVID-19 pandemic, I quickly pivoted from in-person to online interviews from March 2020 to the summer months. Finding older adults who could use Zoom was a challenge! Our sample of older adults leaned towards professional class and more tech-savvy which likely biased our final results.

Highlights

Chicken nachos — digging deep into a research area that is universal and able to make a difference to people’s lives through better UX was genuinely so thrilling for me. Talking to actual users, with diverse backgrounds and experiences, is still my favorite part of this job. I got to ask deeply personal questions with remote participants from Toronto to Ireland. It was a humbling experience to learn about people’s experiences with grief and loss. My favorite quote is about chicken nachos:

What I want might not be what they want. For me, it could be a publication or something, but they could belike, yeah, she was really good at making chicken nachos. 🌮 🌮 🌮

Key user insights

The process of preparing for death should feel personal and intimate.

Participants wanted to create an intimate process by carefully involving loved ones. Most participants wanted control and agency over their self-represenation after death.

Data cannot fully represent an individual.

Many participants thought that preparing data for a digital legacy was valuable. However, some ways of trying to remember the dead were not acceptable, like AI replicas.

Familiarity with existing tools can be naturally extended to death preparation

Existing services like Google Drive, Instagram, chatbots, and smart photo systems provided familiar and easy gateways for novel purposes.

User feedback

Here’s what some of the users said what they wanted for their data after they die.

I have so many papers and I’m tryingto get rid of some because I don’t want this task to be daunting for my kids. I cleared my mother’s house and it’s a struggle. The time is limited, you’re feeling very sad and you don’t have the judgement

If I were to see an AI replica of any of my loved ones that are gone right now, it’d just make me sad and a bit angry because it would be nothing like them. You can’t electronically recreate a human or human essence. I can’t see it making anyone feel good to see that.

The Research Impact

On the academic side, we published our findings in an academic article at the most prestigious HCI venue: the Conference on Human Factors in Computing. Our article ended up receiving a best paper honorable mention award (top 5% of submissions). As my first first-author paper, I was beyond excited! I presented our findings in a video and answered Q&A live to an audience of HCI researchers.

On a broader scale, we were lucky that a number of media outlets were interested in our findings. There were over 10 media outlets that covered our research findings. I even went on a radio talk show to briefly talk about this research project. I was also contacted by a start up to provide consulting on their product in the space of post-death data management.

Three design directions for technology supporting users in preparing data for death:

What I would've done differently

As my first time leading a research project, I definitely stumbled over quite a few parts of this project. I was lucky to receive amazing mentorship and guidance from a senior PhD candidate in my team. Mentoring others is very important to me!

Recruiting more older participants — although it was a challenge to even recruit the 4 out of 20 participants who were at least 65 years old, hearing from more older adults could have strengthened our findings and provided more diverse perspectives. We definitely found differences between younger and older adults, but we didn’t have enough resources to recruit older adults who were able to quickly use online tools at the start of the COVID-19 pandemic.

Integrating spiritual and religious beliefs — perhaps due to our more Western academic perspectives, we were less inclined to include spiritual and religious beliefs into our design concepts. It would be fascinating to understand how religious differences could affect user reactions to the design concepts.