A good user experience goes beyond digital channels and includes in-person and print-based interactions. A piece of paper is more than tree pulp. It can hold the key to something as important as an election. Small usability issues can be magnified by the complexity inherent in interactions that involve multiple pieces of paper which cross-reference each other.
I recently encountered an example of several such usability problems in vote-by-mail materials from the state of Massachusetts. First, to recognize what was good: the state diligently sent my vote-by-mail packet in plenty of time, with all required materials. It was also easy to recognize the mailer as containing the official ballot materials, as opposed to being junk mail.
However, the vote-by-mail process is not seamless, and the printed instructions will likely be ignored or misinterpreted by some potential voters. The outcomes could be very serious: anxiety, wasted time, incorrectly submitted votes that end up ignored, and potential COVID-19 spread due to people feeling the need to vote in person. Each vote should count and should not be thrown away due to poor usability or service design.
Small Usability Issues but Drastic Effects
I opened my vote-by-mail envelope and found inside 2 more envelopes, a long ballot form, and a full page of instructions.
In the first moments, I noticed several problematic aspects:
- There were 2 envelopes, and one was supposed to be inserted in the other — an unusual occurrence.
- The signature goes on the manila envelope, but some people would expect to sign the ballot.
- The signature line on the envelope is not very visible — it is small and sandwiched between the section for the election office and the voters’ address label. Also, there are several other signature lines for people who may have assisted the voter.
- The vote-by-mail process is too complicated if it requires users to read instructions in order to be completed correctly.
- On the instructions page, there is a lot of text and no images. Especially in these cases, people often ignore or just skim the instructions.
- Some instructions refer to the color of the envelope rather than using a clear label (or number) on each envelope that could be easily matched with the different steps.
- The due date appears in a box at the bottom of the page, but that could be easily disregarded because it looks like an ignorable footer.
- The due date also appears at the top of the ballot in a bright yellow box, which could also be ignored due to banner blindness.
Some of these very issues have already caused nearly 18,000 votes to be trashed in Massachusetts in October, according to an article by Channel 10 Boston.
Residents attempting to vote are not all methodical, savvy, calm people. Massachusetts voters include those of us facing anxiety and stress, in a rush, seniors, with low literacy, low vision, cognitive disabilities, or speaking English as a second language. Any of these could make the vote-by-mail process difficult to follow.
Details of the Massachusetts Vote-By-Mail Packet Design
A summary of the Massachusetts vote-by-mail process is detailed in the figure below.
These steps do not seem excessive or highly complex, and to be clear, the vote-by-mail packet has some very good design qualities due to the contributions of UX people, such as:
- The ballot comes with comprehensive instructions.
- Instructions are formatted in chunks with subheadings; subsections are bulleted and keywords are bolded.
- Black text has high contrast with the white page.
- The text is written in a simple, straightforward way. The passage is categorized as “fairly easy to read” by both the Flesch Reading Ease and Gunning Fog and assigned the 6th-grade reading level by Automated Readability Index, Flesch-Kincaid, Linsear Write, and SMOG Index. Two other reading scores are a bit higher, with Gunning Fog at 8th grade, and Coleman-Liau Index at 9th grade.
- The fact that you can track your ballot is advertised on the instructions page.
However, some of the issues with the instructions are:
- The instructions are verbose, and there are no images to help explain the process. Some of the potentially confusing or unexpected parts of the process are called out in bold, which is great, but these are somewhat lost in all the steps.
- It is unclear which instructions are required versus nice to have. For example, if I use a red pen will my vote be counted? What if I neglect to sign the outside of the envelope, or worse, I use only the manila envelope or only the white one rather entombing my ballot in the sarcophagus of two envelopes?
- Does by November 6 really mean before November 6? Would delivering the ballot on November 6 be okay, or is that grace period for only mailed ballots? (Amazingly, in my first draft of this article I believed you could also hand-deliver a ballot on November 6. Further investigation revealed that the November 6 grace period is for only mailed-in ballots, and hand-delivered ballots do not need to be received by November 6, rather, they need to be received by 8:00PM on November 3.)
Small Service Design Issues but Drastic Effects
My 84-year-old neighbor confided in me that she dropped her ballot off at our Town Hall, but plans to also go to the polls to ensure that her vote is already there. Allowing mail-in-voters to also go to the polls is a helpful safeguard and probably a good call for my neighbor. But what does this mean to the voting service design?
At the polls, workers will need to double-check each voter’s status and whether a ballot has already been cast. This step not only introduces an opportunity for errors, but also adds time to the close physical interaction between a poll worker and the voter; plus, it means longer wait time for all in-person voters and, during a pandemic, higher risk of getting sick. This is especially problematic if many of the volunteers and folks in line are older.
WBUR reports that, for the 2020 presidential election, “Any ballots returned without the signed secrecy envelope are known as ‘naked ballots’ and they will not be counted. Secretary of State Bill Galvin says voters who submitted naked ballots will be given a second chance to vote ahead of the election, but with Election Day coming up fast we recommend getting it right the first time.”
Making people aware of the ballot-tracking feature on the web could ameliorate these issues.
We should also consider all these parts of the service design:
- the effort it took to create the packets: the personalized ballot (to vote in your district), 2 envelopes, a label on one envelope, and instructions
- how people or machines counting the ballots will deal with ballots in unsigned manila envelopes or missing white or manila envelopes
- how to provide support for people who used the ballot-tracking form on the website and learned their ballot was rejected
Voting Usability Is an Old Issue.
This is far from the first time UX folks have vented about voting design. On November 8, 2000, the first time I voted in New York City, I recall talking with Jakob Nielsen about how some of us worried that we voted wrong using archaic voting machines and he included my story in his newsletter. (Jakob later wrote more extensively about the infamous 2007 Florida ballot.) Twenty years later, my question is still this: Where are all the UX designers and researchers, service designers, and content writers and editors when voting process and materials are designed? Not there or constrained or beaten by bureaucracy, laws, and deadlines?
UX people know how to research user journeys and service design. I once spent weeks with a financial institution testing, iterating, and retesting the envelopes and print materials related to credit-card bills and interest-rate offers to make that information comprehensible. The results were a success. It’s amazing that a credit-card company would hire us for this work, yet our governments don’t invest the same for our vote-by-mail design.
If UX had been involved, the Massachusetts vote-by-mail process would undoubtedly be better. In just moments I identified issues, and any good UX person would look at this packet and do the same.
Summary
To be clear, the issues written in this article have nothing to do with voter fraud and have everything to do with poor design and organizational and political issues that cause it. Such design issues might exist in any state or country. But Massachusetts has at least 6 universities with established UX programs, we are home to some of the best UX-research and design firms, and there is an active UX community (the UXPA Boston chapter has more than 3,000 followers on Twitter; the UX East Boston Meetup group has 1,365 members). All UX folks should take more initiative to fix the voting process (#voteUX on Twitter).
Sadly and appallingly, state and local governments are not known for their UX prowess, and redesigning anything related to a ballot form is constrained by laws and could require input from politicians and others who have zero design experience. Still, governments should pay close attention to service design and user research when it comes to voting, and get some serious, professional help or listen more to the help they have.
Share this article: