Jeff Johnson was an apparent underdog. Even those who believed he’d win or wanted it, stretched their imagination to think the big money and establishment infrastructure of Tim Pawlenty could be overcome. Not many people knew he would win.
Make no mistake, Jeff’s message was a winning one. His political instincts are impeccable, his message genuine, and his integrity unmatched in the world of politics.. He will be our next Governor because of that. But at NATIV3, we knew he’d win the primary because that’s what social data was telling us.
Let me explain.
NATIV3 was hired by the Jeff Johnson for Governor campaign on June 13th, shortly after his endorsement at the Minnesota State Convention. At this point in time, the campaign challenges were obvious: name ID in the prospect of Tim Pawlenty, fundraising in the prospect of Tim Pawlenty, and ensuring we could get OUR people out to the poll August 14th in the prospect of Tim Pawlenty. NATIV3 partnered with graphics guru Derek Wehrwein to ensure – from a social media perspective, a paid search perspective, a re-marketing and display ad perspective, and from an overall digital strategy perspective, that we lapped Tim Pawlenty in creativity, execution and hustle. We’re a small team at NATIV3 – 7 of us to be exact. Understanding the fundraising gap, every dollar we spent, every initiative we undertook, and every message we deployed would need to be maximized, analyzed, tested and pinpointed to a controllable audience. We won’t get too much into the inside baseball of what that took, because we now have a general election to win.
But I wanted to cover the high points of what worked digitally. It worked so well that my message through our basecamp project management software read like this when we sent our final report to the campaign: “When Jeff or Justin (campaign manager extraordinaire) or anyone on the team write their autobiography, I want to be the first one on record: We’re going to win the primary tomorrow.” The polling numbers were scoff-worthy in the face of hard social data.
A LOT went into the victory of Jeff Johnson – especially on the digital front. But what it boiled down to was the following 3 main points:
- immense target testing
- diverse and contrasting creative assets
- giving the voter something to do.
First, Immense Target Testing
One of the best data assets that the Johnson campaign brought to the table was the list developed from Jeff running for Governor before. As an agency, we run list-based targeting for businesses of all shapes and sizes – and you usually get a hodgepodge of old friends, names with spelling errors, and half correct information. I’ve noticed a similar trend with typical political lists and Secretary of State information. NATIV3 segmented the lists by issue specific trends, modeled the data to match likely primary voters, and ultimately developed 50 separate campaigns to beat the living pulp out of that list with diverse ads, compelling calls to action, and actionable ways for them to engage with the Johnson campaign (more on that later). Keep in mind with list targeting, you can upload a controllable universe to Facebook advertising, Google Search, YouTube Preroll, and Display Advertising, and can even utilize certain Twitter targeting features.
Lists are the lifeblood of a campaign. I’m more convinced of that than ever before.
Later, as the vote got closer we further broke down the list and actually created lookalike audiences to broaden out the scope of our targeting. As the campaign team and I discussed, we thought that there were names and numbers and emails and personality types that could not be accounted for in any lists: namely, Trump voters. We modeled demographics and behaviors off of the large list we already had, so with nearly every advertising campaign created we dealt with 3 audiences: Jeff’s loyal supporters (an immovable base and an incredible campaign asset), our modeled primary audience (which we segmented from Jeff’s list), and a broader audience of people who matched similar traits as our listed audience. This equated to a universe of 400,000 people. Our goal was to use our two precious months to drive home name ID, build contrasting messaging on all platforms to Tim Pawlenty, and get a commitment from the prospective voter in some form or fashion. The data operation coordinated between the Johnson Campaign and NATIV3 was one of the complex executed in Minnesota politics (hourly monitoring, rapid spend increases and decreases, and rapid A/B testing). As campaigns began and ended, NATIV3 manipulated budget towards the ads that were most compelling, altered text and calls to actions, and created micro-targeted assets that maximized both engagement on their respective digital media platform AS WELL AS ensured that the voter would want to see the landing page beyond the ad.
As we closely monitored data and competition – socially, organically, video content, and paid search, it was incredibly apparent that the Pawlenty campaign did not prioritize digital strategy as fiercely as they needed to. In fact, the ONLY campaign that was comparative to impression numbers in this race (from what we could tell) was Tim Walz – which primarily came from a video campaign that ran the last month of the race. Our estimates put that at a 6-figure investment and the targeting was VERY broad.
At the end of the day, we honed in on a finalized voter universe of 329,000 voters, served them nearly 10 million impressions (30 impressions per voter over the course of 2 months), garnered 50,000 conversions (voter commitments), and spent very little money comparatively to do so. The Johnson campaign ran a lean, tight ship. We loved being part of such an execution-oriented team. But in order for those numbers to work, the creative must also be spot on. The creative needed to be on point, which leads us to the next H2:
Diverse and Contrasting Creative Assets
Before explaining what went well, I want to give another tip of the hat to Derek Wehrwein. Derek ran most of the graphics production, which was an incredible strength of the Johnson for Governor primary campaign. Understanding we needed to not only foster better name ID statewide for Jeff, but also create a contrasting message for the future from Tim Pawlenty, we catered almost every post, every ad, and every theme around that premise: build Jeff Johnson’s name up and contrast him to Tim Pawlenty. As we dug deeper into some data around the Pawlenty name ID (something viewed as a great strength by most pundits statewide), we quickly realized that not all of that ID was positive. We needed to put the proverbial boot on the proverbial throat repeatedly, every day, to that controlled audience. Part of our challenge was not only helping boost Jeff’s name ID online, but also reminding Minnesotans of who Tim Pawlenty was.
NATIV3 and the Johnson campaign both handled day to day posting on social media platforms. Obviously, it’s critical to show campaign activity. Bruce Miller, the Johnson Campaign driver (and SO many other things) captured imagery that NATIV3 could use for daily posting. We made sure maintenance activity and posts were regular, effective, edited, AND always tied back to Jeff’s messaging. It’s one thing to talk about being at a business round table. It’s another thing to talk about being at a business round table AND talk about how Minnesota’s tax system is broken for small business owners. The engagement numbers between those two types of posts would blow your mind.
But beyond maintenance and static posts that I liken to 3-4 yard off-tackle hand-offs (to use a football analogy), we also threw Kirk Cousins-like bombs. A lot of them. Remember that targeting we talked about in the last section? We dropped some of the most compelling and creative imagery, video content, boosted posts, and audio assets in front of that list. Facebook was our main driver. But we also utilized Paid twitter (VERY little), Youtube Preroll and Adwords to drive home the engagement and impression count. The two days before election, Jeff occupied most early vote searches on Google, Yahoo, and Bing.
We ran through several issue specific campaigns, utilized earned media and boosted it to the universe we had, and then a massive GOTV and engagement push the final 3 weeks of the campaign. You’ll notice, even in polling and predictions, Jeff was climbing statistically. The polling, though, still did not reflect how effectively the base was turned out.
Grabbing impressions within a controlled universe and using those impressions and creating compelling content are all critical for building name ID (what what our world calls “top of funnel awareness”). But you can’t complete a deal – if you’re a politician or an e-commerce website or an HVAC Contractor – unless you give your user something to do. Impressions without action don’t mean a hill of beans.
Giving Voters Something to Do
At every turn, with every ad, we needed the voter to do something and commit to something once they clicked through. This is simple in nature, but it modeled Donald Trump’s digital strategy. We knew that farming ads targeted toward those who cared about ag issues in our modeled universe would be attractive to click on. What happened next would be far more critical than a mere click. Our creative revolved around “Here’s what Jeff believes”, “Subtly, here’s how Tim Pawlenty hurt you”, and “Now we need you to do something about it”. With each landing page, we either asked for a voter commitment or a small contribution RIGHT AFTER and RIGHT NEXT TO a presentation of compelling emotional issues or political concern. We did not let an impression go to waste. We needed to ensure that there were actionable items we could give to the potential voter. At every turn, getting a voter commitment was the critical end goal of each campaign we ran: be it adwords text ads, YouTube preproll, re-marketing campaigns, or Facebook ads/boosted posts. When it was all said and done, with new online voter commitments and DEFINITE supporters within our database, we knew our floor was 70,000 voters. For computing purposes, assuming 70,000 voters brought .3 voters along with them the polls, we were 99% confident we had victory in our grasp August 13th.
Giving voters something small to buy into allowed us to create new email audiences, new remarketing audiences to ask for another donation/ask, and ultimately a very targeted GOTV effort to those who already supported our campaign.
Making the ask, even if it’s online, proved to be incredibly effective.
Jeff Johnson is an amazing candidate. The grassroots participation in his campaign, and how he inspired so many people restored my faith in Republican politics. Volunteerism, especially Republican volunteerism is a force to be reckoned with. Experts will debate in the coming weeks what his upset of Tim Pawlenty means. As a firm, we are so looking forward to the coming months of this campaign and pivoting to a general election strategy. Jeff will continue to inspire voters with his message, drive grassroots activists to action, and hustle for every vote touring the state. He proved that in this primary process. NATIV3 will need to be sharper than ever in matching his veracity for the general election.
But at the end of the day, the data proved that this was not nearly the upset we might want to make it out to be. Data-driven targeting and inbound marketing principles played out. You can adapt or you can lose. Jeff’s vision for innovation in this campaign is one of the major components of what won him this election and what will help him win in November. As a small digital firm, we’ve been humbled to be part of the process. The Johnson campaign took a chance on us. We loved being part of this process.
Digital is the present in political marketing, as it is for big brands and small businesses alike. There is no “future” in digital and inbound marketing principles.
And you need to start if you want to win. It’s not just throwing hundreds of thousands of dollars towards a campaign. It’s building a measurable relationship and asking for “buy in” with every impression you pay for. Jeff Johnson had a winning message. We ensured that voters heard it online, cared about it, and derived actionable next steps from it.
As for our next step: all eyes are on November. Our universe is considerably bigger, but the methodology doesn’t change.