Two steps forward, one step back: digital programs in the 2018 midterms

By Alex Kellner

It’s been two weeks since Election Day, so it’s probably time to talk about 2020, right?

Before we get too far ahead of ourselves and focus fully on the upcoming Presidential election, we want to take a moment to look back on what we learned from the past two years.

In the House, Democrats truly did get a sizable Blue Wave, winning more seats than we have since the post-Watergate election. While the tough Senate map led to predictably bittersweet results, Democrats dramatically outperformed their 2016 results across the country thanks to strong performance among college-educated, young, black and Latinx voters, flipping a number of Governor’s mansions and down-ballot elections along with the House. It’s clear that the playing field is changing and nowhere is that more true than on digital.

In 2018, Bully Pulpit Interactive was proud to partner with a number of organizations to run digital programs for several Democratic campaigns, running over $100M in spend, including:

• The largest ever non-Presidential digital persuasion advertising program
• The largest ever House-focused mobilization program
• The largest young voter-focused program ever

Our biggest takeaway: political groups finally took a lesson from the corporate world and invested seriously in digital programs for brand building and persuasion. But campaigns still silo their digital programs and haven’t fully adopted a true audience-first approach — we still have a long ways to go. Here’s what we learned in 2018:

You Can’t Win Without Investing in Digital

For the first time since digital programs became a fundamental part of campaigns, leading Democratic organizations and campaigns didn’t fall for the 15% fallacy, instead making digital a core part of their strategy and planning their online programs and budget with an audience-first approach. In many cases, Democrats invested serious dollars in digital programs and hired more in-house digital staff than ever before. Leading campaigns and independent expenditures identified a path to victory, determined how many people needed to be persuaded and mobilized, and budgeted digital programs accordingly.

Digital-First Video Is Now the Expectation

The average American voter spends over 5:53 per day online — split across their computer, mobile device and connected TV. Each of those is a different experience and it’s critical that we tailor our creative based on where it’s consumed. In 2018, Democrats embraced this challenge.

Candidates like Alexandria Ocasio-Cortez realized that they didn’t need a six-figure budget to produce effective video for the Internet, all they needed was an authentic story to tell. And Super PACs identified ways to supplement candidates without stepping on their toes — using compelling visuals and an informational approach.

What isn’t needed? 1 million individual pieces of creative. One of the biggest misnomers from the 2016 cycle was the press-driven narrative that Trump’s campaign tested tens of thousands of pieces of creative. If you’ve spent as much time as I have looking at the Facebook & Google transparency reporting, you will notice a lot of repetitive creative with very slight tweaks in accompanying text or targeting methods, but far fewer variations of video and static visuals that accompany them.

But Television Still Gets the First and Last Dollar

Even with larger digital teams and budgets than ever before, television still gets an outsized portion of the budget. According to Kantar Media, there was $3.2 billion in television advertising, compared to only $900 million online this cycle. In some markets, there were well over 2,000 points in the final weeks, as many Democratic decision makers opted to match Republicans point-for-point instead of allocating incremental budget where it can be most effective. Do we really think that seeing a television ad for the 27th time will persuade someone in a way that the previous 26 times did not?

Transparency is a Double-Edged Sword

No campaign manager has ever been fired for matching TV point for TV point, which has led to millions of political dollars wasted. This red bar/blue bar fallacy is potentially creeping into the digital realm thanks to the biggest change from previous cycles — the introduction of transparency dashboards.

In 2018, digital spending was much more out in the open, thanks to data published by Google, Facebook & Twitter. Let’s be clear, this is absolutely a good thing for democracy. Voters have a right to know who is paying for political advertising and how much is being spent. However, this new development has a strategic downside for smart digital strategy — the publication of these numbers game-ified digital spending, which results in politics spending more money than is strategic on platforms that published spend than those that didn’t. I worry that as this trend continues, Democrats will attempt to match Facebook dollar for Facebook dollar, leading to the equivalent of the 27th television ad problem. Especially when you consider that even with this additional information on political spend, much is missing — namely, the purpose of the ad. It’s easy to confuse list building dollars with persuasion or mobilization expenditures within the current transparency tools.

Smart Buying is Harder Than Ever

Every impression is not created equal. No two voters consume online media in the same way — some may watch a lot of video on social channels and others may watch none at all. It’s true that it’s easier than ever to launch an ad online. But to make sure that someone actually sees the ad, Democrats have had to establish more advanced relationships with ad platform algorithms. For example, when we targeted our persuasion audience on social channels it disproportionately delivered impressions to younger voters even though they aren’t necessarily the best targets. By using expertise buying on the platform we were able to optimize delivery to reach the voters with whom our advertising has the highest ROI, instead of simply who is easiest to reach. The rules of the game are constantly changing, so we need to strike the balance between accuracy and delivery to break through.

In 2016, Democrats learned that over-relying on models and polling can have serious disadvantages. If anything, in 2018 we overcorrected — opting to take more of a broadcast approach (talk to a broad audience) to digital targeting than the direct mail approach (talk to a very narrow audience). For example, at BPI our 1:1 targeted political spend was down 21% compared to last cycle. The answer lies somewhere in the middle. We absolutely shouldn’t take offline data and the models that build on it as gospel and spend all of our money on 1:1 targeting; however, we also need to ensure that we don’t forego it in its entirety because programmatic buying is the best tool we have to drive a persuasive frequency online.

A Single Approach to Testing Doesn’t Work

Perhaps the shiniest new tool in 2018 was the explosion of online panels to test message effectiveness. Unfortunately, there is no silver bullet to creative optimization. We tested the results of an online panel against real-world measurement and there was a substantial difference between what voters said in a closed panel environment when they knew they were part of a study thea what they said after being exposed to the same content in the real world.

Like with the use of data, the answer isn’t either/or — it’s a combination: Democrats should use online panels for the qualitative feedback and to help narrow many pieces of creative to a few pieces of creative, while leveraging real-world data to map content to specific voter groups.

Cooperation Makes It Happen

Together with Priorities USA and Civis Analytics, BPI created a coordination dashboard for Democratic Super PACs that facilitated smarter spending and ensured that the ROI of Democratic digital spend went as far as possible. From divvying up different audiences or messages for each group to own, to ensuring multiple groups weren’t bidding on the same search terms, the dashboard was a key first step that Democrats need to build upon in future years.

What’s Next?

In a cycle that will undoubtedly have dozens of candidates vying for the nomination for President, a challenging slate of Senate races, and the mandate to defend the 35+ seats we gained in the House, Democrats must build upon the digital success of 2018. We should focus on:

• Budgeting based on the best method to reach each voter — instead of treating each method (doors, TV, radio, mail, digital) in its own separate silo.
• Testing messages early in the cycle, so they can be iterated upon and leveraged when it matters most.
• Expanding upon coordination efforts to ensure that the ROI of digital spend is maximized.

The last thing we can afford as a party is to become complacent with progress made this cycle, or we’ll end up repeating 2016.