How we designed a tool for you to get your friends to vote
Using public data to spark meaningful outreach and sway elections
Co-authored with Christopher Scott
If you feel like you don’t have a lot of power in US elections, you aren’t alone. We’ve conducted over one hundred user interviews in the past year that suggest a lot of people are feeling that way.
It turns out, though, that we all have a superpower to influence elections right at our fingertips: encouraging our own friends to vote.
In fact, our rigorous research shows that sending a simple text message to a friend is between 2x and 20x more effective at getting them to vote than sending a message to a stranger.
That’s where VoteWithMe comes in. VoteWithMe is a free app that helps everyday people have more influence in US elections by increasing voter turnout among the people they know.
VoteWithMe shows you information about the contacts in your phone like:
- Elections they can vote in and how tight those races are predicted to be
- Which party they registered with (e.g. Democrat)
- Whether they’ve missed important elections in the past (meaning they could really use your reminder!)
The information VoteWithMe shows you about your friends is all publicly available, but is rarely accessed or used by the public (although political campaigns have been using it to target their outreach efforts for decades). VoteWithMe evens the odds by putting all this information in the palm of your hand.
The path to building VoteWithMe was riddled with major design challenges, many of which led us to decisions that lots of folks have asked about. And so in this post, we would like to share more about why we designed the new VoteWithMe in the way we did, offering a look into the questions we asked and the findings that informed our design decisions.
1. How might we present election and voter data in a way that is immediately meaningful and helpful to you?
Solution: Supplement users’ knowledge of their contacts with voter data, rather than replace it.
If you were using VoteWithMe back when we first started, you’ll remember that the previous version of the app would simply present you with a curated subset of your contacts for you to remind to vote. The app sorted your friends for you based on things like their voter participation history. We quickly learned that there were three fundamental problems with this approach.
First, users frankly didn’t trust the algorithm and didn’t understand why the app was suggesting seemingly random assortments of people for them to remind to vote. Many of the users we interviewed wanted to see the data used to rank their contacts in order to understand how they ended up sorted that way (spoiler alert: we listened!). Asking people to reach out to their own personal networks about anything ‘political’ is a big ask — so big that in early versions of VoteWithMe, users often asked if they could opt to text strangers instead. (We didn’t pursue this path because research has found that texting strangers is much less effective.) Given the extremely high social cost of reaching out to one’s friends, the uncertainty that users felt about our algorithm and how it ranked their contacts was often enough to stop them from using the app altogether. We knew that going forward, it would be critical to reduce users’ uncertainty around our data and contact ranking.
Second, displaying only a subset of your contacts was a crapshoot. Our algorithm couldn’t possibly take into account your own relationships with your contacts. Surely it’s more comfortable (and likely more influential) for you to remind your best friend to vote rather than your dentist.
But we couldn’t detect the depth of your relationships with your contacts — what are the odds that if we pick 10 people at random from your phone’s contacts, they would be people you are comfortable texting right now about voting? Exactly.
We also couldn’t take into account any of your own relevant knowledge about your friends. For example, you might know that your friend has recently taken an interest in social and political change — they post a lot on Facebook about the Kavanaugh hearings, started attending March for our Lives protests, etc. But if this change of heart wasn’t visible in our historical data, we did not rank them as a high priority for you. You know better, and it would be a mistake to prevent you from engaging with them.
In the new version of VoteWithMe, we solved all three of these problems by getting rid of the ranking system and showing you the data for your contacts directly. Instead of being opaque and prescriptive, we became transparent and empowering.
Now you can come up with your own outreach strategy by filtering your contacts in a way that is meaningful to you (e.g., show me which of my friends are ‘Democrats’ that ‘can vote in a tight election’) and reach out to the people you’re comfortable talking to.
We’ve heard repeatedly that it’s more powerful for you to directly see the voter information about your contacts and decide for yourself whether it’s important to remind them to vote.
For example, seeing that a friend is eligible to vote in a race that is predicted to be a “toss-up,” but didn’t vote in the last midterm election, is far more compelling than just having the app say “Trust us! It’s really important to reach out to them!” Through this new approach, we are showing people public information they didn’t have easy access to before, which some may find unsettling. We felt this too, and offer some additional thoughts in our last post about VoteWithMe. But since making this change to share more data with users, we’ve found that people actually understand and trust the app much more when they can see the data for themselves and make their own decisions about who to contact.
2. How might we help you initiate productive conversations about voting with your own friends?
Solution: Provide users with brief, relationship-based sample messages that include actionable voter resources.
In our research, once users decided they wanted to reach out to a friend, they often struggled with how to approach the conversation. They were unsure of what to say, concerned that they didn’t know enough about their friend’s election, and/or fearful of the conversation snowballing into something more controversial.
We found that providing short, actionable sample messages made users more likely to reach out and helped to get the conversation rolling. During user testing, we noticed that people often cited the nature of their relationship with someone as a reason they would or would not reach out to them. In response, we introduced sample messages you could send to your contacts based on your relationship with them (e.g., your best friend vs. the guy in the next cubicle over vs. your high school crush).
These sample messages reduced the burden of figuring out what to say, and the relationship-based message labels cued to users that they could start a conversation about voting with someone, even if they weren’t close friends.
We made the content of the sample text messages about voting in general, and not the specific election, to address the understandable fear users have about entering political debates with their contacts via text message. Of course, you can add or change whatever you want in the messages (we encourage the use of emojis)!
In addition, we generated personalized voter resource links for you to include in your messages to your friends. These links include voting information specific to your friend, like their election deadlines and polling location. We’ve found that the inclusion of these links positioned users to feel helpful and mitigated their concerns that they weren’t knowledgeable enough about their friends’ elections to prompt them to vote.
3. How might we assure you that your data is secure and that you are in control of communication with your contacts?
Solution: Only ask for data that is absolutely necessary, explain why it’s necessary, and make sure the user feels in control of communication with their friends.
In order for VoteWithMe to give you election information for your friends, you must grant the app permission to access your contact list. This was a major point of hesitation for users across the board, especially following the Cambridge Analytica scandal. Specifically, users were often concerned that if they synced their contacts, we would sell their contact data or spam their contacts directly.
We will never do either of these things. Our Executive Director, Mikey Dickerson, is a recognized champion of data security and user privacy and protecting your privacy is a core value of our team. We respect our users for being cautious about their privacy. Rather than try to skirt around this issue, we made numerous design and product decisions to assure our users that they are in control of communication with their contacts.
First, any time we do ask you for a piece of information, we do our best to explain why we need it and what you will get in return. Rather than throwing an iOS permissions pop-up at you asking to access your contacts, we have an interactive page that shows you the types of information that you’ll be able to see about your contacts if you opt in.
We’ve tested these permissions pages extensively with users in order to find the language and structure that they understood best. With the redesign, users now better understand what we’re asking for, and are empowered to make an informed decision about whether or not to sync their contacts (they almost always agree to sync their contacts now!).
Second, we needed to make sure you would feel comfortable and in control as you navigate throughout the app, unafraid that each thing you tapped would trigger automatic outreach to your contacts. We’ve intentionally used visual cues and messaging throughout to assure you that this won’t happen.
For example, in the redesigned version, the button at the bottom of a contact’s voter record (see image to the left) is intentionally worded as “Pick a starter message”, not just “Remind Tina to vote” to make it clear that there will be another step before any communication with Tina and that tapping the button won’t trigger an automatic text message to Tina.
In user tests, in addition to paying attention to signs of hesitation, we made a point to explicitly ask at every stage, “Once you tap this, what do you think will happen next?” Then we iterated on the design accordingly.
VoteWithMe has come a long way since we tested it earlier this year.
Every step of the way, we’ve relied on user feedback to make decisions about how to move forward. From listening to our users, we completely changed how the app works — from opaquely ranking your contacts to transparently sharing information about them — and made hundreds of tweaks to the user experience design.
As a result, our users have found the app to be more compelling and engaging. They return to use the app multiple times over multiple days, spending twenty, thirty, or even forty minutes reviewing and filtering the information about their friends. They also reach out to more of their contacts than they used to and those contacts who receive a message sent from VoteWithMe are more likely to download the app than they used to be.
And we’re not done yet. We’re still actively working to improve the app and make it more accessible and powerful for you. Taking a user-centric approach to VoteWithMe has been a big win for us, our users, and the progressive movement.
Now, let’s go sway some elections.
In the current climate, causes and campaigns too often lack the time, expertise, and flexibility to work beyond immediate deadlines. The New Data Project (NDP) is a new 501(c)(4) organization built to address this gap by testing new approaches, looking beyond the current cycle, and serving as an advanced technology research lab for progressives.