How a search engine fired Nancy Pelosi
By now every observer of American politics is familiar with the phrase “Fire Nancy Pelosi,” especially the soon-to-be-former speaker herself. But even the wonkiest of political creatures will be surprised to learn how a software algorithm led to that meme entering the political landscape. What is bound to shock data-driven technology people is that many Republican insiders completely disagreed with the data, and thought that a campaign targeting Speaker Pelosi wouldn’t ignite public ire. This is the inside story of how Speaker Pelosi found herself the target of the brand of politics she has practiced for years, and how this may well be the year that consumer technologies chose a winning political message; it is also a warning tale to people in the public opinion and mobilization businesses.
In the weeks before Speaker Nancy Pelosi succeeded in ramming through President Obama’s effective takeover of the American health care system, the leadership of the Republican National Committee was engaged in a ground battle of sorts to stop the legislation. Mimicking a political campaign, all of the divisions of the RNC worked together to help people in swing-districts get informed about Obamacare and get in touch with their representatives. We were far from alone in that effort. Separate from our campaign, Tea Party activists took to town halls, groups like American for Prosperity organized rallies; Americans, many of them first-time political activists, called the Capitol to voice their concerns about the government making their medical decisions for them. Despite all of this, the fact remained that, by virtue of the Democrats’ overwhelming majorities, the Obamacare bill would pass. (However, this was largely a charade since Pelosi announced she would use an obscure procedure to “deem” it passed otherwise.)
If Pelosi slammed the bill through, Americans had made it clear they would be outraged. They would want a place to focus that frustration, and an immediate way to fight back against a Congress that was at the pinnacle of their already impressive arrogance; not only were they refusing to listen to a growing majority of its citizens, they seemed to mock the very idea that Congress should ever listen to the people — “this is my town hall, I make the rules” and “we have to pass the bill so you can find out what’s in it” — should serve as the American versions of “let them eat cake” (falsely attributed to Marie Antoinette). Part of our job at the Committee is to work with people at moments like these, to help them get the outcome they want through electing people who want the same things. In this case, Americans desperately wanted the chance to repeal a bill that even the most committed Tea Party activists believed would be jammed down their throats. The question for us was, how do we best capture the moment with a message that is compelling, big enough to stick and, most importantly, honest?
The prevailing wisdom among the political long-beards was that we should launch a movement based upon the message “repeal it.” While that was (and is) unquestionably our legislative goal, the view of those of us new to politics was different: since we knew the world would be full of people calling for repeal, it was clear that would join a cluttered field of messages and therefore advertising terms would be expensive. We also favored a tactic that would most ideally meld our message — the Democrats are not listening — with an immediate action our supporters could take, one that could last well past the health care debate and into November. The way we expressed this message had to be undeniably the single most potent and defensible words we could use.
While many of the longtime political operatives pushed a series of good ideas like “Repeal It Now” “Repeal, Replace & Remove,” and “Make Them Listen,” a few of us stuck with our thesis that the real target of our campaign should be the leader who was most visibly pushing that piece of legislation, and was already reviled by a majority of the country: Speaker Nancy Pelosi.
The prevailing counterarguments were politically sound. Here are a few of them: Nancy Pelosi may well be disliked, but she has very low name recognition; people in the middle of the country don’t care about Nancy Pelosi; and we cannot nationalize a race around a Congresswoman from San Francisco.
The two camps were far apart on tactical implementation, but not on the absolute importance of doing this right. All of us insisted that we test which phrase best communicated our plan to supporters and independents. But, with zero time to launch a poll (and no budget with which to do that), we had limited options for learning what would work best. So we turned to the one device we could tap immediately: search engines.
To be perfectly blunt: we asked the Google “machine” for the best way for us to drive our message home and raise money to fight to repeal Obamacare. Our team gained agreement from even the technical neophytes in the building that they would agree to accept the findings of Google, Bing and Yahoo! search advertising data. We plugged all of $50.00 into search marketing campaigns, and agreed to go boldly with the meme that won the race for click-throughs (meaning it resulted in the highest percentage of people clicking on it). In less than one day, we would have our answer … but, not our consensus. The longest-tenured political operatives — brilliant, accomplished people whom I deeply admire — had an array of reasons for disagreeing with the data and thinking our tactic was stupid. To their enduring credit, they stuck to their promises to go with the search-winner. The data won out, and what impressive data it was.
For those who remain unfamiliar with what search marketing is, and how it works, a quick primer (for those already familiar, skip ahead one paragraph): When you search for something on Google, Bing or Yahoo!, the search engine scans the Internet to find the sites they think are most relevant to your search term. Those returns are called “organic search results.” Getting to the top of those results is worth a lot of money, as you can imagine. Groups can also be “keyword advertising”; this allows you to runs ads on Google or Bing adjacent to these organic returns. So, if your site is not on the top of the organic search returns, your advertisement can be. We used keyword advertising like a poll, and the data we got turned out to be spot on.
We ran the phrases we wanted to test next to searches for “Obamacare,” “Repeal Obamacare” and a myriad of related terms. Among the phrases we tested — there were roughly ten of them — there was a clear winner. Any advertisement that contained the phrase Fire Nancy Pelosi drew an extraordinary percentage of clicks: well into the 60-70% range (a normal keyword campaign gets a 1-2% click-through-rate, and 5-7% is considered outstanding). Given that data, and the fact that no one else online was using that phrase, the decision seemed clear to those of us schooled in consumer technologies and the data they provide. It was less clear to people used to dealing with polls that cost tens, sometime hundreds of thousands of dollars to field.
The message delivered from Google, Bing and Yahoo! was loud and clear: anything other than “Fire Nancy Pelosi” was at best ordinary and at worst a loser. The long-beards went with the New Media hippies and, together, we launched the Fire Nancy Pelosi money bomb, which became by far the most successful online fundraiser in the history of the RNC. The potency of the message can be displayed in the efficacy of a three-day campaign:
● $1.6 million raised
● 53% of contributions came from first-time RNC donors
● Six contributions came in at the federal limit for donations to the Committee
● $17,000 spent on advertising for the full campaign, equaling close to a 100:1 return on investment
But, of more strategic importance, we owned Nancy Pelosi’s name for a full week on Google (our site, and stories about our site, were the top organic return for her name). We also caused people to search for the phrase “Fire Nancy Pelosi,” which became the third-most-searched-for phrase on all of Google on the second day of the campaign; read that sentence again: a phrase chosen by Google and other search engines, propelled by social media (as discussed below), and predictably leaped upon by marketers of political-chatter on cable TV, became the third-most-searched-for phrase on all of Google. Yet, more important to us from a viewpoint of the staying power of that message was the way that it came alive in the social sphere. Yes, technology not only helped us name our movement, it introduced us to the people who would help us grow it.
Days before the event launched, we shared the Google data and our site design with people who have large social media followings. The moment we launched — which was at the exact moment the bill passed — Twitter became a sea of #firepelosi hash tags (a method people on Twitter use to help categorize their tweets). By the first morning of a campaign that began at roughly 11pm, social media had proven to be the early driver of the campaign; while we had spent only $400 on advertising, we had earned over $400,000 above that 100:1 ROI. In the days that followed, we watched as Facebook sharing proved the most potent means of delivering traffic — and new donors — both to our site and our movement. These new people were extraordinarily active. In fact, the people who came in the door through our Fire Nancy Pelosi campaign became the most active and reliable activists we gained in the 2010 cycle.
In the months that followed, the people whom we met through that event helped us perform at record-breaking rates in email marketing. When we asked for their help, they were there. While the RNC’s old email list had open rates of about 3%, the GOP Action Team email program (peopled connected to the Fire Pelosi & anti-Obamacare campaigns), had an average open-rate of 45% and an average action rate of 29%. Days before the November election this team of people were the ones using our Volunteer Action Center to find opportunities to assist campaigns and utilizing our call-from-home application (40 Phone Calls To Fire Pelosi) to make voter contacts.
In the weeks before the election, the meme that search marketing had helped name became the focus of renewed chatter among political reporters and cable talkers; The Drudge Report featured several above-the-fold photos of people holding the Fire Pelosi signs that RNC staffers had given out at rallies. When Chairman Steele decided to undertake a bus tour, the name of the tour was a fairly obvious choice. As the days ticked by, Democratic congressional incumbents began to actively run away from Pelosi (something we had predicted would occur if we successfully marketed that meme). A handful of Democratic incumbents even ran television ads critical of the speaker. It was no doubt a grassroots phenomenon, a Tea-Party victory, a win for Congressional “Young Guns” — Republicans, most of whom are completely new to office, who bring new blood and bold ideas to Congress. There is also little doubt, however, that the idea of firing Nancy Pelosi was a catalytic notion that helped join all of these forces — in Beltway and blessedly out of Beltway — together on a goal that resonates immediately and emotionally, but also ran the long term.
Digital mobilization strategists will remember this as a watershed event. It is the first time that a search engine ad campaign inspired a national movement, and in all likelihood was the first time such a movement was named for only $50. It will not be the last. Pollsters and non-digital mobilizers would be wise to bookmark the moment as well.
But this is not the only way consumer technologies made a massive leap toward acceptance in national politics. These technologies, especially Twitter, are far ahead of purpose-built technologies, like the Associated Press’s Election Watch, for immediacy and accuracy. On the night of November 2nd and long into the morning of November 3rd, I worked in the “War Room” with our amazing political team as we watched returns roll in from around the country and made the all-important last-minute efforts to get out the vote. Throughout an eight-hour period, Twitter proved to be far ahead of the AP’s election monitoring tool, and the five television networks we were following. In many cases, it was literally hours ahead of the more established tools. It was also better at self-correcting.
Late that evening I received word from someone monitoring cable TV that a race had been called in favor of the Republican. I tweeted it out, and it was immediately re-tweeted by a well-known and widely followed Republican operative. Within minutes I was being corrected, and I quickly tweeted that I was in error. I then apologized to the candidate and operative, who re-tweeted that clarification to their followers. Consumer technologies are not only faster with the news, it appears their users are quicker to admit fault.
This means something significant for people in the business of political campaigning, and it’s well worth reflecting on what you’ve just read as it relates to en vogue conversations about technology and politics. These two disciplines are a natural match: technology processes an amazing amount of data very quickly, and political campaigns mobilize huge numbers of people in a small amount of time. The conversation about technology and politics often turns to turnout-the-vote measures — a natural focus, of course. But, for better or worse, American politics has always been about money, message and marketing, which lead to turnout and make get-out-the-vote technologies necessary. But, the technology of persuasion is just around the corner, and it is going to be a game-changer.
This year, I believe we have witnessed the tiny birth cries of a powerful new mix of technology and politics, one that signals a sea change in the relationship between political entities and public opinion. I am the first to admit the Fire Nancy Pelosi case is an isolated one, pursued non-scientifically — but that is the point. The science in public polling, opinion determination, and social mobilization around a political movement should recognize that point. In the same way eBay has made it easy to make money online and YouTube has given birth to the citizen broadcast star; just like easy blogging platforms interrupted the newspaper business, and micro-blogs the breaking news business; consumer technologies like Google, Bing and Yahoo! are simply first shots at changing polling and mobilization. More advanced technologies already exist — technologies that operate well above the pixie-dust of most “semantic analysis” products. The ability to listen to and understand the populace without a poll — without invading people’s privacy — already exists. Polling and mobilization firms have the same choice that others had before them: fight or embrace the new. These technologies will take hold. You cannot stop them.
One of my favorite quotes, which I saw on a sign in Hollywood years ago, said this: “Embrace that which you cannot avoid.” With technology becoming the cornerstone of opinion monitoring, and technology-driven mobilization becoming the core, not the adjunct, of the business of helping people move toward an action, this is what we must do.
Embrace what you cannot avoid.
Todd Herman is a dad, husband, and digital media pioneer. He has been a start-up founder and CEO, angel investor and Microsoft executive; he was most recently the chief digital strategist at The Republican National Committee.