Iowa Caucuses Should Not Have Used An App

Four days later, the Iowa Caucuses have finally finished. Yet, we're just now seeing the results. What the hell happened?

As an Iowan (and a Software Engineer), I was disappointed to hear it was a failure by technology. I was surprised that our tech community would have let this happen. So, I did a little digging.

If you don't read anything else in this article, know that:

• The caucus app was not made in Iowa, or by Iowans.
• The mismanaged, under-budgeted app had no chance of success.

To understand the full story, we have to first go back to 2016.

2016 Election#

After the close race in 2016, new reporting standards were introduced to provide more clarity on the caucus numbers. Precinct leaders were now required to submit three numbers – including the number of backers on the first and second alignment. Previously, they only had to submit one.

There are over 1,600 precincts in the state of Iowa.

This change meant it would take longer to gather the data on the night of the caucus. So naturally, party leaders looked for a solution. Is there a tech solution that will simplify the reporting process and improve our confidence in the results?

The App#

Enter Shadow, a Denver based company contracted to build a reporting app for precinct leaders. This app would simplify the reporting process to allow caucus results to be submitted the same night, even with the heightened reporting standards.

Why not hire developers in Iowa? (Hint: Money).

In my opinion, this was the first failure by Democratic party leadership. There are plenty of development shops in Iowa which could have executed flawlessly. But, hindsight is 20/20. So what happened?

What Went Wrong#

The night before the caucuses, the Iowa Democratic Party chair seemed pretty confident.

In reality, some of the precinct chairs weren't able to test the voting app until a few weeks before the caucuses. As you can imagine, that didn't turn out well.

To report results in the app, you had to:

• Sign in with email and password.
• Open Google Authenticator for two-factor authentication.
• Enter a one-time code from the app.
• Return to the voting app.
• Enter his precinct PIN.
• Report the results.

But they didn't make it that far.

When trying to enter the one-time code back into the voting app, it failed. This frustration prompted many precinct chairs across the state to abandon the app and call the telephone hotline to report their results.

In the wake of national media collectively losing their minds, Iowa's democratic leadership failed to communicate effectively.

But the app did fail to work as intended, causing the phone hotline to be overloaded and reporting delays to ensue. The app can't take the full blame. Ideally, the phone hotline as a backup should have prevented the delay.

What Should Have Been Done?#

The app shouldn't have been built.

Did the steps to submit results in the app seem over-complicated? They did to me. There are many ways you could verify vote counts digitally.

Hell, a Google Sheet would probably suffice.

Firebase? React Native? Talk about over-engineering. Shadow's budget was $60K from Iowa and $58K from Nevada.

For what they were trying to accomplish (even if it was the wrong choice), that budget wasn't going to cut it. Employing even a few Software Engineers is going to blow that budget almost immediately – let alone think about best practices.

I have empathy for the engineers at Shadow. Some were new and inexperienced. They lacked the leadership and direction to understand they were making poor technical decisions.

So Now What?#

• We need to rethink how we're funding tech initiatives in politics.
• Iowa Democratic leadership needs to understand the core issues and prevent this from happening again.
• Encourage the use of open-source software in politics.
• Fund tech efforts for politics in-between cycles.
• Improve training and preparation for volunteers.

We're not ready for voting apps. Then again, we may never be.