AJ O'Leary

about • portfolio • essays • reviews • contact

Murphy’s Caucus: A Product Manager’s Take on Iowa’s Nightmare

Our democratic experiment is not the place to stress test your mobile app

This article was written on February 5, 2020, and references the American political climate of the time period.

The results are (sort of) in from the Democratic Party’s presidential nominating caucus in Iowa, and the winner is… a “total technological breakdown”.

What should’ve been a roaring start to a primary campaign to decide who will earn the chance to face off against Donald Trump in November instead turned into a mess reminiscent of Florida’s 2000 Presidential election, as actual results were delayed well into the next day due to claims of “quality control” issues from Iowa Democrats. Accusations flew and conspiracy theories ran wild as the major campaigns began issuing their own statements of frustration and vote total claims.

Eventually, a culprit emerged: a new mobile app rolled out to support the caucuses didn’t work as intended, glitching out and causing mass confusion amongst the people tasked with using it.

As someone who manages a mobile app for a living, I’m all too familiar with the challenges that come with launching a new product. Based on what I’ve learned, I have some takeaways to offer after analyzing the situation like I would if this were a product I was personally responsible for.

Manage Your Moving Parts

There’s a saying I like to bring with me into development and strategy meetings: “Plan for the worst, and work your way up from there.”

This doesn’t mean “be a pessimist and act like the sky is falling” — rather, it means you should be a realist. Technology isn’t foolproof. You and I are not perfect. You can’t just roll out something as major as a new mobile app and assume that anyone will inherently get it, because you cannot assume anything will just fall into place with product management and deployment.

Worth mentioning, too, is that the concept of a caucus is so absurd and difficult as it is — in other words, it’s got way too many moving parts — that calls to scrap Iowa’s caucus system existed even before Monday night’s debacle. Political concerns aside, if your event is already so convoluted in nature that it requires a term paper-length explanation, perhaps a new mobile app that requires its own handbook isn’t the best thing to add to the mix. If there’s too many variables at play, something will go wrong.

The First Cut is the Cheapest

Let’s say you’ve got a decent budget and you’re in the market for a new car. Your options are:

• Extremely cheap with a high risk of failure

• More expensive, but with a far better safety record

Which would you choose? Unless you’re a masochist, I’d assume you’d pick the second car.

I would like to think the same line of reasoning would apply to software development, and yet, for the Democratic Party, it apparently didn’t — not if stories about their spending somewhere around $60,000 to launch their caucus app in Iowa and another $50,000 in Nevada are to be believed. I’m not a procurement expert, but I know enough about development costs and schedules to tell you that you shouldn’t trust software to reliably support an actual election of all things for the cost of a Tesla. Engineering and tech support and quality analysis and everything else doesn’t come cheap, as is the nature of human labor.

If you plan to develop a worthwhile app, test it properly, and educate its userbase to a point that potential error is diminished as much as humanly possible, you should expect to pay a premium, especially when the event you’re supporting is the first and most visible step towards nominating a candidate for one of the most powerful offices on Earth. Even worse, cheaping out on a stage like this further erodes the already-fragile public trust in the electoral process. Never compromise your integrity just to save a dollar.

You Can’t Just Wing It With Support

A recurring theme in the thick of Monday night’s madness was what industry types call a “cascading failure”: everything at every point in the process broke down in some manner or another. In two moments that would be funny if they didn’t concern an actual election, an Iowa precinct chairman trying to manage the situation live on CNN was hung up on by the state party around the same time that one of the caucus chairmen found themselves tasked with physically driving the results over to the party headquarters in the state capital in the dead of a cold Iowan winter night.

Hearing these stories, I couldn’t help but think how differently things might have played out if Iowa Democrats had put more physical, knowledgeable subject-matter experts on standby than they obviously did.

As humans, we have a tendency to either overestimate our own individual resources and capacities or underestimate the complexity of situations we’ve been asked to take responsibility for. (Sometimes, we even do both!) I don’t see it as a logical leap to think “It’s just a caucus. We’ve done these before!” or something similar was a thought expressed at one point.

Again, however, the Iowa caucus is the worst possible time to take these things for granted. Trusting one person, or a very small handful, of people with well-informed support and assistance for an app serving even 10,000 or so people in one confined physical location would be daunting enough, and likely result in the tech equivalent of someone getting pelted with soccer balls. For an event servicing hundreds of thousands of people in a geographic distribution as spread out and, in many cases, isolated as Iowa, your likely outcome is the cascading failure that ended up occurring. Without proper support in place, questions will go unanswered and make failure inevitable.

There’s an entire technical and social discussion to be had over all the things that went wrong in Iowa, but if I were the person tasked with answering to this series of unfortunate events, I’d bring the above takeaways to the postmortem meeting — and go from there. Put simply, no matter your own political leanings or preferred candidate, if you’re looking to regain your competitive digital advantage as a political party and restore faith in “the system”, a poorly-planned development cycle and launch is never the way to do it.

If you liked this article, let me know.

Click here to view my other essays.