Progressives need a UX design revolution

Could better design have doubled the number of high-quality voter contacts in 2018?

On the Saturday of GOTV weekend, my toddler and I drove 90 min to the Manteca canvassing HQ at the Josh Harder for Congress campaign in CA-10. We arrived at 11:40, raring to get out and talk to some voters. But instead we spent nearly an hour — a third of a canvassing shift — waiting in line or straining to hear a mass group training taking place over an inadequate sound system with no visual aids.

Better UX design of the apps involved in canvassing could have cut this frustrating experience down to 10 min. And it wouldn’t just have been faster — the training would have been more effective as well.

From texting apps to event creation forms, the left’s UX design is cringe-worthy. Bad design means fewer, less effective volunteers, which means losing winnable races.


The fundamental story of Tuesday’s election — and the story of the last 2 years — was progressive grassroots mobilization. That energy is being channeled through technology. Millions of volunteers used tools like canvassing apps, texting apps, phone call apps, event creation apps, and more, to organize and engage in voter contact. Many of them were contacting voters for the first time ever.

Beto for Texas’s website that tracked volunteer-driven voter contact

Don’t get me wrong. Our digital tools are better than they’ve ever been before. And I know and love many of the people who build them. But we can’t just improve marginally every cycle over the previous cycle, because users will continue to have higher and higher expectations every cycle. And so many political volunteers are brand-new each election, we have an enormous amount to gain by providing them with a great experience.

We should be aiming for every single volunteer who shows up in 2020 to find our UX so easy to navigate that they can’t wait to come back for the next shift — and then again in 2022 and 2024.

Progressives need a design revolution.


Where we can do better: Concrete examples (sorry, friends :-/)

I hesitated to call out individual tools as examples. I know how hard those who built them and managed them this cycle worked. I am incredibly grateful to you for the heart and soul you poured into these tools, and I’m aware how exhausted many of you are right now.

But I think examples are necessary to understand the depth and breadth of the problem. These are neither the worst nor the most important UX problems I encountered; they are simply representative. So, friends, I hope you can accept them in a spirit of constructive feedback from someone on your team — or at least close this tab now and reopen it in a week when you’ve had some rest. Or a month!

Example 1: Lack of onboarding process in tools like Hustle, Relay, NGP VAN, Mobilize America, PDI, etc.

When you first create an account for any successful for-profit consumer-facing app, you’ll go through a new user onboarding process. Onboarding processes increase retention, allow users to master the tool faster, and make them aware of key features that they would otherwise miss. For example, below are some screenshots of Duolingo’s onboarding process (you can read more about onboarding for apps in this helpful article).

Screenshots of Duolingo’s onboarding process: The app asks whether you’re a beginner or not, and sends you through a different flow depending on your answer.

Not a single voter contact tool I used this cycle had any form of in-app onboarding. Instead, you got dumped immediately onto a screen like the below.

Initial screen upon logging into Hustle. The user is presented with cryptic menu options like “Workflow” vs “Conversation”, with no explanation given.

Training on how to use these apps, when done at all, was done in some other venue: On the phone with an organizer; in Slack; etc. (To give credit: MoveOn had the best and most scalable training process for one of these apps that I experienced — a set of videos about how to use the tool, and quiz questions that you had to answer before proceeding. Still a far cry from a good in-app onboarding process, though.)

At the canvass in CA-20, my canvassing partner waited 20 min in a line to get assigned turf (while I tried to entertain my toddler in the parking lot…). Then we were stuck for another half hour either waiting for, or listening to, our training session. Because the app didn’t help with the onboarding, passionate volunteers had to make the best of the situation, doing training in 20-minute blocks. We waited for the previous group training to wrap up, then stand in a crowd of hundreds of people (below) straining to hear a volunteer speak over a too-quiet sound system about how to use the PDI app for canvassing, with no visual aids.

Imagine, instead, that as soon as we were ready, each volunteer got their phone out and spent 5 min going through an in-app onboarding process, then had the opportunity to approach a volunteer organizer with any specific questions we had. We would have spent an extra 30 min canvassing — and also learned the scripts and tools much more effectively.

Photo of hundreds of volunteers listening to a trainer speak at a CA-10 canvassing HQ, about how to use PDI for canvassing. The speaker has no visual aids and an inadequate sound system.

There is an entire subfield of UX called FTUE (first-time user experience). And you don’t even have to have spare engineering capacity to implement its best practices! Check out tools like Chameleon and Full Story, which integrate with and overlay your existing UX with interactive product guidance, without any engineer intervention required.

Example 2: MoveOn’s Host a Wave Event page

MoveOn is the oldest and possibly the largest among digital progressive movement organizations. The alumni of former MoveOn staff are involved in running most major digital programs on the left. But rigorous UX design thinking is not strongly embedded in the culture of MoveOn and its diaspora (my own organization, SumOfUs, included) — and sometimes it shows in a big way.

An example from this cycle: MoveOn ran a powerful and massive-scale nationwide grassroots voter contact program this year. Its #1 ask of its activists this year was to host a Resist and Win Wave Event. On mobile, the tool where you would register such an event requires you to scroll through 13 pages of dense text before you get to the form where you fill out your information. Ironically, I don’t want to copy it all in here for fear that I will lose you as a reader, but below are the first three pages…

First three screenshots of MoveOn’s “Host a Wave” mobile page (which is entirely text).

…you get the idea. No Silicon Valley tech company would ever expect users to read through a chunk of text like this, and their designers would get laughed out of the room if they suggested it. (Reminder to MoveOn staff: I’m so grateful for the work you did!)

Example 3: Failure to tailor recommended actions to users.

After I had used Mobilize America 5 or 6 times to phone bank for 3 or 4 total candidates, I returned and logged in again. I was presented with an initial list of actions I hadn’t yet taken. Even after filtering again remote phone banks, I couldn’t easily find the candidates I’d already been phone banking for.

Mobilize America’s initial screen after logging in and filtering “remote phone banks” on the day before the election. The user is presented with options unrelated to which campaigns they’ve worked on before.

In this case, the first step towards tailored recommendations is obvious and easy: Upon logging in, the first thing I should see is a list of actions of a type I’ve taken before (e.g., phone banking), that are immediately available for a candidate I’ve volunteered on behalf of before. Then you can add other layers on top: If all I’ve done is phone bank, the next tier can be other remote phone banks. Etc.

Turf-cutting for canvasses is also a recommendations problem, in a slightly different way: For the first 20 min after we arrived at the canvassing hub for the Josh Harder campaign in CA-10 above, we had to wait in line for a volunteer to manually assign us turf.

Instead, why not have the PDI canvassing app ask us 3–5 questions in-app, like:

  • Is this your first time canvassing?
  • Do you have a car with you today?
  • When do you have to finish canvassing?
  • Do you speak enough Spanish to be able to hold a conversation like this one? [EXAMPLE]

Then assign us turf automatically, using the campaign’s prioritized list created that morning, which a single organizer could be monitoring and editing in real-time.

Other examples

I could name literally dozens of other examples of obvious, major UX problems I encountered. E.g.:

  1. Not a single voter contact app — phone banking, texting, or canvassing — that I used had a way to mark voters as having already voted early/absentee, if they weren’t already registered as permanent absentee.
  2. Every single voter contact app had a portion of their script asking supporters of the candidate to make a plan to vote. This is great! Except that not a single app *explained* to the volunteer why the plan-making part of the script was important. Asking someone to make a plan is an awkward thing to do, and we can’t expect most volunteers to have attended an Analyst Institute training. I’d imagine that volunteer compliance with that part of the script could be doubled or tripled with some creative UX design.
  3. Discovery/SEO. I tried Googling things like “text bank for Democrats”, and came up short of a quick and frictionless path to volunteering almost every time. I suspect that this is because organizers assume that almost all users are returning to their tool from an email or other push notification, or through social media shares. In the context of the current political climate, we should be prepared for users who want to take a high-bar action on their own terms, not just when we think to ask them, to try to find our tools through search engines.
  4. Trying to find a texting program to join was an incredibly frustrating experience. At my first voter contact party, I had several people attend who wanted to text bank. I assumed that we would be able to find them a texting volunteer experience on the spot through Mobilize America or MoveOn. Instead, we spent an hour trying to figure out how to text-bank on short notice, failing with no explanation of why this was so hard, and eventually giving up and making phone calls instead. Even if texting on short notice actually is impossible, throw the user some kind of bone to explain why, and redirect them along a more useful path.

The list could go on and on. How many volunteer-hours did we lose this cycle because of these problems? How many budding activists gave up and won’t be as eager to do voter contact next time? How many voters didn’t turn out because the volunteer who contacted them wasn’t using the most effective tools?

My best guess? We could have doubled the combined quantity/quality of volunteer-driven voter contact this year with better UX design.


Digital departments and vendors need fundamental restructuring

The answer here isn’t: Fix the list of items Taren named above.

The answer is: Recruit, hire, and empower experienced product managers and UX designers. They should be running your product development, not staff with no formal technology background.

Most serious tech products these days are built by teams structured around the “triad” — a tech lead, a product manager, and a UX designer.

Increasingly, non-profits like MoveOn and the DNC have substantially-sized engineering teams in-house. Some of these organizations are starting to hire product managers, or at least give the title of product manager to a campaigner who has been effectively acting like one.

But I have almost never heard of a non-profit with a trained UX Designer on staff.

The job of a UX designer is simple: Design a great experience for users. But the skillset and tools required to do so are non-trivial.

I think that the non-profit campaigning world has overlooked the deep expertise available in this work because campaigners and organizers (like me!) fundamentally believe that we already know how to design great user experiences. And it’s true that there are deep similarities between UX design and organizing. But just as you would be wary about hiring only UX designer straight from Silicon Valley company onto your organizing team, you should be just as skeptical about organizers with no specialized training being in charge of your UX. Our instincts are often fundamentally good, but we only touch the tip of the iceberg of expertise available in the tech sector.

I hope that political vendors like NGP VAN, Hustle, PDI, etc. already have designers on staff. But if they do, those designers are clearly not adequately empowered or resourced. And they won’t be until the clients of these vendors demand better design.


How to be a UX design revolutionary

We mobilized unprecedented numbers of new volunteers in this midterm election. In 2020, I predict there will be even more newbies entering the arena. We MUST provide them with the same quality of user experience that they expect when they download any app from the app store, or we will lose a large percentage of them forever.

Think about it this way: Do we want the experience of being a first-time GOTV volunteer to feel more like the smooth experience of signing up for an app like Facebook for the first time — with streamlined onboarding, in-app explanation of features, and clear purpose? Or more like the experience of being a first-time voter in Georgia — facing long lines, frustrating and obscure systems, and bottlenecks that seem intentionally created to make the process as difficult as possible?

So, here are my challenges to different groups of stakeholders:

Leaders at digitally-native non-profits: Start investing now in a design revolution for 2020.

For example:

  1. Audit the UX of your 2018 tools. Make an account on usertesting.com and get some random people to walk through the tools you used this year. Ask them what they would have trouble using and how they would make the tools better. Set up some phone interviews with both power and casual users of your election tools and ask them what they loved and what they struggled with. Mix in some of the many other tried-and-true user research techniques the tech industry has developed.
  2. Hire an experienced product director and lead UX designer ASAP. It will be tempting to hire for these roles from within the movement. But, at minimum, to build a top-flight product, you need either Product Director or a Design Lead with substantial experience in the for-profit tech sector. The processes, tactics, and strategies the tech sector are too valuable and too well-refined for you to expect to be able to recreate and rival them from scratch in-house.
  3. Change the way you assess vendors. Most of the vendors I discuss in this piece are B2B2C. You pay their bills; the end user does not. They will lean into better UX design if and only if they feel like they have a mandate from you. I bet that when your staff are deciding whether to use, e.g., Hustle or Relay, they are most heavily weighing their own UX as administrators of the tool. They look at how easy is it to create new lists and assign powers to super-users, etc. You should instead be primarily evaluating on metrics like user activation (how many new users actually start texting?), user retention (how many users who text once return for another session?) etc. And you should make clear to your vendors that when it comes to allocation of their own scarce resources, in the end, the experience of end users matters to you more than the experience of your staff.

Board members, funders, tech sector workers: There’s work for you too.

  1. Non-profit board members: Insist that your staff leadership do the work above.
  2. Funders: Fund UX design trainings and positions. It’s going to be hard for grantees to invest in designer staff positions and hire great designers at their current very low level of knowledge about the specialty. And then, fund the UX designer positions when your grantees create them.
  3. Progressive tech sector workers: Support non-profits and political campaigns in their skill-building. Organize training programs in UX design and product management for non-profit campaigning staff and Democratic campaigns. And consider going to work for one yourself!

Many thanks to Nick Allardice, VP of Product & Design at Change.org; Erica Morse, Lead Product Designer at Changeorg; and Ben Rahn, co-founder of ActBlue. They helped me develop some of the ideas in this post and provided great input on the draft.