Failure to launch: What happened with the app in Iowa
Hint: user experience matters
“Did you make sure it’s plugged in?”
“Have you tried restarting?”
It sounds condescending, but how many times have you found yourself asking someone these questions before you dug into why the technology wasn’t working? Sometimes we assume that user error or common mishaps are the cause of the failure or unexpected results.
Perhaps the obvious question for the failed app designed for the Iowa Caucus was, “Did you even test it?”If we back up a few steps before testing would occur, we’d find that the project plan was non-existent. We know that they had about two months to develop the app. The Iowa Caucus isn’t a surprise or last-minute event. Quite the contrary, it’s a tradition and something that Iowans look forward to hosting. It marks the official launch of the political season. For political nerds, it’s like the Black Friday of the holiday shopping season. It signals the first of many trips leading up to the big day.
This year, the caucus resembled a frantic and chaotic Black Friday experience. People waited in line, spent hours, and in the end, weren’t able to get what they wanted. Two days later, they’re still without the goods. When they called customer service, they were on hold for over an hour. That’s what happened in Iowa.
This was a day that the Democratic party knew was coming, one where they hoped to digitalize and ramp up their efforts. Typically, viewers know who wins the Iowa results the night of. All of the news outlets are monitoring and you can see voters raise their hand and tally up who got how many votes and then move onto the next round until the winners are determined (sometimes that comes in the form of a coin toss) but hey, that’s a caucus for you.2020 was different.
By 1 am eastern time, there were still no results and likely wouldn’t be until the next day because of outputting technical difficulties. Party officials said they wouldn’t report the results until they had all precincts in and an actual account of what happened. Come Wednesday, people were told by 5 pm there would be results, but Wednesday came and went and so did Thursday.
The app was supposed to help put faith back into the caucus process. Caucuses are different from primaries but they also carry a unique set of rules and tradition is something that is celebrated more than anything. Each cycle, though, candidates hold grievances on the fairness or accuracy of the results that help give momentum. These complaints put more pressure on Iowa to have a spectacular and seamless night, and they changed the way votes would be counted. Instead of each precinct announcing the winner only, they would have three different sets showing each round and vote. It was complicated before tech even got involved, but a solid development strategy could have solved the problem.
In an age where tech can help diagnosis and give treatment plans in healthcare, provide precision medicine in surgery, and predict consumer buying patterns, one would think that it could help people vote and tally accurately.
So where did the app for the Iowa Caucus go wrong? In two months, the company rushed planning, development, and testing into one. Did they gather specs and build out iterative sprints? Did they even think about who would be using the product or what the purpose was?
The business objective was much more than to just allow voting. The primary function of the app was to populate data to meet the many rules of caucuses: understanding both the popular vote and the raw data for delegate distribution. Counting hands is easy. But understanding the raw data and output is what was most needed in Iowa. For the company to say that there were no bugs but that the outputs weren’t calculating properly is a massive flaw.
At it’s best, tech helps solve problems. At it’s worst, it makes problems seem bigger and causes more of a mess than manual labor. The caucuses are the latter scenario. Complaints about reporting not being fair in the past created a strong need and gap for tech fill. Each precinct even has its own rules for calculating results. In one precinct I saw captains from other campaigns who were leaders of counting a competitor’s votes — this is based on the honor system. Other precincts have attendees sit or stand in a line and when they are pointed to, they put their hand down.
Once these objectives could have been outlined, emphasis should have been put on the importance of tradition in Iowa as going first in the nation. Understanding that complaints have sparked against caucuses in previous years and knowing the tension with Iowans wanting to keep their tradition and impact in the state, the app could have enhanced the brand and improved the process. Instead of upholding tradition, the app all but destroyed it.
Should the app have been developed in Iowa? It’s now been dubbed as the Iowa app fail, but in reality, it was the Democratic machine’s failure. The company that built the app is located in DC. Former Clinton staff and Buttigieg contacts work for the tech company. Competent tech startups exist in Silicon Valley and Austin that could have handled the task. Iowa entrepreneurs could have stepped up to the challenge. They would have understood the unique value and challenges. They could have provided brand identity and trust to the caucus and their home state.
How did this year-old, DC startup app miss the mark? We’ve established that their vision and goal for the app were off-mark. But so were the intended users. As noted earlier, each precinct has different rules and team leads. Looking at some of these leaders, they varied in knowledge and probably interest in using the app. When the app wasn’t as user-friendly as one would have hoped, customer service did not pick up the slack, or the phone.
My favorite article about the debacle was told by an actual participant, a Democratic county chair in Iowa. The purpose of his role is to provide backup to the precinct chairs in the county — and boy, did they need a backup that night. He drove to one location for one voter. When they say that every vote counts, they mean it. Here was one participant waiting to cast one vote and they couldn’t do it. When he called in the result, the state party said the precinct had already been counted and reported that they were having trouble tabulating single-person precincts so they went back to pen and paper.
It took time to even get someone on the phone because caucus chairs were experiencing confusing results and difficulty with the app so now volunteers were in charge of customer service for an app they didn’t create or understand.
Let’s unpack this rollout, or as the Democratic county chair put, dropoff on the doorstep. Hit-and-run might be more applicable.
Getting the app was a challenge.
Typically you would go to an app store and download your app. Maybe there would be some instructions on how to use it. Some quick tips to help the user navigate the UI and UX journey in the platform itself.
That might have taken too much time for this rushed debacle, so instead, you had a five-prong approach just to get the app:
1. Fill out a survey
2. Receive an email with a link
3. Click on the link and download a different app (not the actual one you would use)
4. Enter a code from your email
5. Finally, get the real app
Security in tech is very important. People can use their knowledge for good or for bad. Hackers can be critical in the tech industry, and even more in politics. Especially when the past and current cycles were experiencing conspiracy theories on hacking and interference with the election process. Making sure you know your end-user and provide a journey that they can follow is also important.
I remember looking at some of these precinct captains and thinking they look older. Not Ok Boomer old, but like they probably have Boomers for children old. Not to say that this generation cannot use tech, but it’s likely that they aren’t trailblazers in the industry. This is a new concept and we’ve already discussed how important tradition is in Iowa, would these users be likely to adapt and learn this new tech? Likely not, but it doesn’t matter since they didn’t even provide training on the app function and how to use it.
The actual users of the app tried it the night before the big caucus day so they could be prepared (Note: the day before launch typically isn’t going to help with bug fixes). Only two of the 22 captains were able to successfully download it. That was just for one county. Iowa has 99 counties. They knew the night before this would be a failure but they advertised high expectations the day of the event and said there were contingency plans.
When the UX failed, CX wasn’t there to help either. The system figuratively and literally had shut down. Volunteers and captains gave up and went home because this wasn’t what they signed up for. Developers can’t be expected to man the phones and explain every step in the process or counter every complaint. But someone in the company needs to understand and create messaging and training so that customer service can be supported and knowledgeable. If the app doesn’t have quick tips or a user-friendly help section then a real person or chatbot will need to step up.
A week later and the state party can report that Pete Buttigieg won the most delegates with Bernie Sanders close behind. With New Hampshire’s primary up tomorrow and the Nevada Caucus next, voters and the country are looking to see what Nevada will use. The state party of Nevada already announced its decision to dismiss the same app that Iowa used, but recent reports show that they are moving forward with an iPad tool, which is not an app. It will come pre-loaded on the iPads that are given out, but nothing has been confirmed. As Nevada considers this tool or app or whatever process they choose, let’s remind them of important steps no matter what type of tech you use:
1. Know the purpose or problem
2. Know your users
3. Build iterations that give value
4. Test your product
5. Have customer service
6. Make sure customer service understands the tech
This close to Nevada, I’m doubtful that they will have time to do QA testing in-depth and build iterations. Hopefully, whatever tool or app they use, though, will have come after sufficient research and hands-on application.
The app for Iowa was not an MVP (minimum viable product). The real MVP (most valuable participant) were the people like the county chair who stayed up and sorted through all of the data manually and reported it — twice — because the state party lost the results the first time.
The big joke for Iowa was that it got stuck on 62% in calculating the results. People were saying 62% of movies that ended sadly were great at the 62% mark. 62% of a developed app isn’t great either, especially if you haven’t been building iteratively.