Location-based check-ins — yay or nay?

Joycelyn
Government Digital Services, Singapore
6 min readMay 11, 2021

It isn’t always easy to decide what’s best for users. To mark one year of community-driven contact tracing, Team TraceTogether will be taking you behind the curtains through a series of musings.

If ensuring users’ trust in your product comes at the cost of accessibility, what would you do?

Since TraceTogether’s launch in March last year, we’ve gotten numerous requests to include location-based SafeEntry features in the TraceTogether App.

Screenshot of feedback from users: e.g. “Can TraceTogether App do auto check-out when it detects that I’ve moved away?”, “Can TraceTogether App list SafeEntry venues based on my global positioning service (GPS) location, like how Singpass App does it?”
“Can TraceTogether App do auto SafeEntry check-out when it detects that I’ve moved away?”
“Can TraceTogether App list SafeEntry venues based on my GPS location, like how Singpass App does it?”

Technically feasible to include location-based SafeEntry features in the TraceTogether App? Yes. Useful to users? Yes!

Unfortunately, this seemingly straightforward decision to include location-based features is a tricky dilemma for TraceTogether.

Before we go into details, you need to first understand how TraceTogether and SafeEntry are related.

TraceTogether helps MOH identify people who’ve been in close proximity with COVID-19 cases, using a privacy-preserving Bluetooth technology. Users can use either the TraceTogether App or physical Token. SafeEntry helps COVID-19 cases remember the places that they’ve been to by submitting a “check-in” at each place, and this helps MOH to spot clusters early.

These two systems have been working together to help MOH improve the speed and accuracy of contact tracing.

Seems quite comprehensive and well-thought out, yeah? But it wasn’t all planned from the start. TraceTogether and SafeEntry began as standalone solutions back in early 2020, when Singapore was trying whatever it could to manage the new disease that had hit our shores.

When it became clear that people needed to use both solutions to support contact tracing, the TraceTogether team started including SafeEntry features in the TraceTogether App (and subsequently in the TraceTogether Token), so that people only needed to use one item to do both Bluetooth proximity contact tracing and to submit “check-ins” at places.

“Huh, still so lay chey (Singaporean slang/dialect for ‘troublesome’). The government can use the app to track our GPS location, right? Why not use GPS data to do contact tracing?”

First, for any mobile app to track your global positioning system (GPS) location, users need to give explicit permission in the app. Second, GPS doesn’t work well indoors and in highly urbanised settings. If you’re in a building, your GPS location could look the same as someone on the floor above because of signal reflections and multi-path propagation effects. Third, using GPS data for contact tracing raises serious privacy and data security concerns.

For these reasons, TraceTogether was never keen on using GPS location data for contact tracing.

How might we meet important accessibility needs and take care of privacy concerns at the same time?

The most common location-based feature request for the TraceTogether App is to display a list of nearby SafeEntry venues for people to check in — so that persons with visual impairments (VIs) don’t need to wave their phone around to detect the SafeEntry QR code; wheelchair users don’t need to wait behind people who are crowding around the SafeEntry QR poster; and just anyone who finds it more convenient to check in without walking past a SafeEntry QR code can do so.

The alternative SafeEntry check-in method — SafeEntry Gateway — may not work for VIs and wheelchair users either, as they need to be able to see and reach the Gateway to tap it.

Imagine having to rely on others to help you at each SafeEntry point 😭 A location-based SafeEntry check-in feature seems to be the best option for these groups of users.

Screenshot of feedback from users, describing the challenges that persons with disabilities face when doing SafeEntry QR scan or tapping on SafeEntry Gateway.
The common SafeEntry check-in methods (e.g. scan QR, tap on Gateway) may be challenging for persons with disabilities.

How will the feature work?

  • Users will need to give the TraceTogether App permission to access the user’s GPS location, i.e. their rough pin on a map.
  • The App will only detect the user’s GPS location when the user taps to use the location-based feature, to retrieve nearby SafeEntry venues to display in the App. Outside of the feature, the App will not be able to collect GPS location data.
  • Also, the GPS location data is discarded once used and not stored.

So the feature should not be mixed up with apps that continuously detect and collect users’ GPS location.

Here’s the dilemma.

The TraceTogether team is all for building the feature because one of TraceTogether’s key design principles is inclusivity — that everyone can access tools that protect themselves and their loved ones from COVID-19. The feature brings clear accessibility wins, and it’ll become more important for persons with disabilities when TraceTogether-only SafeEntry is implemented soon.

The feature will be opt-in, i.e. users who don’t wish to use it don’t need grant new app permissions. This is how the location-based SafeEntry check-in feature works in the Singpass App, which seems well-accepted (even loved 😛).

However, compared to other government apps, the mere mention of GPS location can be sensitive for contact tracing apps like TraceTogether. There are people who still have the misconception that TraceTogether continuously collects data about their GPS location. 😣

With GPS data and privacy being a delicate topic, we decided to run a poll with a group made up of members from the public who are keen to discuss things related to TraceTogether, to get a rough sensing.

Screenshot of the poll that we ran a poll with a small group of members from the public, to get a rough sensing.
We ran a poll with a small group of members from the public, to get a rough sensing.

Out of 92 who voted, 74% were comfortable with an opt-in GPS location check-in feature; 26% were not.

To be honest, I was expecting clear support for the feature. So when I saw the results, I felt disheartened. I thought: “Is it difficult to empathise with the everyday challenges faced by persons with disabilities? Why do people have different priorities in this situation?”

It took me a few hours to accept that there are people who’d like a greater degree of assurance when it comes to privacy (I’m mindful that I’m measuring against my own yardstick). Even if the feature is opt-in and users need to grant explicit permission, “there will be people who don’t believe [what the product team says].”

Anything that could risk participation in TraceTogether would set us back in our contact tracing efforts. 😢

Screenshot of comments from the group, on why they were uncomfortable with an opt-in, location-based check-in feature.
Comments from the group on why they were uncomfortable with an opt-in, location-based check-in feature.

What would you do if you were in our shoes?

When it comes to national level decisions, it’s unrealistic to require 100% consensus — the diversity is too great. Some countries use rough consensus to decide.

Even though this wasn’t a nationwide poll, it brought insightful learnings. For instance, is 74% vs 26% considered rough consensus, or is 26% significant enough to disqualify the majority?

Also, the numbers game is rarely in favour of persons with disabilities. Is polling even a good way to make decisions in this case?

As designers, developers, or product people, we’ll often face such dilemmas where there’s no right or wrong, only what we as a team or a country agree on.

How should the TraceTogether team navigate from here? If you had similar experience, how did you resolve it? Let us know your comments below!

This dilemma is one of many that TraceTogether faced since we launched a year ago. We’ll be writing more posts to share other memorable experiences :p If you found this interesting, look out for more posts!

Thanks to Jason, Adrian, and Quy.

--

--

Joycelyn
Government Digital Services, Singapore

UX designer@GovTech. Design lead for TraceTogether. Behaviour science x Psychology. Always up for a game of touch rug 😊 | linkedin.com/in/joycelynchua