Big Things Have Small Beginnings 

1. March 2024

As we approach a new election year, Ben Pearton our Head of Activation & Analytics explores possibly the first use of paid ads in a political campaign

I have always been fascinated by the power of the message when it comes to political campaigns.  

 Depending on the region, or the position up for election, words and themes are carefully crafted in focus groups and demographic studies to create the catch phrase that can encapsulate the entire process.  

As online marketing has grown at an astonishing rate over the last decade one feature that both political and digital messaging have in common is that it should be short, sharp and memorable if it hopes to have any kind of resonance with voters. 

In 2008, Barack Obama’s collection of upbeat slogans (‘Yes We Can’, ‘Hope’, ‘Change We Can Believe In’) are still some of the first things I think of when casting my mind back to being at school, and seeing it on the TV or hearing it on the radio.  

5 November 2008: Obama becomes first African American to hold the office of US president - editorial | US elections 2008 | The Guardian

This campaign was also arguably the first time that digital strategy played a decisive part in a presidential campaign. Mary Crosse at estimates that the Democrats may have saved up to one hundred million dollars in TV buying by utilising social platforms. She also points out that Barack Obama “was the first president to rely on micro-donations of hundreds of thousands of supporters rather than big ticket donations of select high net worth individuals”, something else that likely would not have been possible without embracing the new landscape being formed online.  

To compliment these new strategies, Obama’s 2008 campaign website was also built in a simple and effective way, with clear call to action buttons and messaging (the evolution of which can be seen here.) Search boxes where the public could enter their zip code to find out where to vote are also present, allowing the capture of user data for remarketing purposes such as email campaigns.  

But, despite the 2008 presidential election putting this strategy on the map for many future campaigns (Trump and Brexit come to mind), it was not the birth of the idea. This of course led me to the question: 

When did paid media get its first break in the political sphere? 

In November of 1998, the race for governor of New York was reaching its climax. Peter Vallone, a Democratic New York City councilman, was attempting his first run at the position, and showed interest in diversifying his tactics to help his ratings in the polls.  

In October of that year, Crain’s New York Business reported that $100,000 was being invested into a paid online marketing strategy by Vallone’s campaign, making him “the first political candidate in the country to launch a significant on-line ad buy”. While both Vallone and his republican opponent, George Pataki, had hired specialists for website development, only Vallone was willing to invest such a large sum of money on what was, at this time, an unproven tactic. 

It is worth pointing out that many candidates before this point had put out contracts to have websites built for their political campaigns.  

Despite this, the late 90’s represents a time where the jury was out on how effective online strategy could actually be, and so it was often neglected. This was characterised in an example given by Rebecca Rainey discussing the quality of two campaign websites in the race for governor of California between Gray Davis (click the link to take a look at the site in all its 90’s aesthetic glory) and Dan Lungren (whose domain is now, rather amusingly, a Japanese molecule removal website).  

Writing for the New York Times in 1998, Rainey quotes Michael Cornfield and ‘Mindshare Internet Campaigns’ as they express their dismay at how limited the sites for each contender are, especially when the winner would govern Silicon Valley, a beacon of cutting-edge digital technology. They also make the point that the winner would have an important voice in the build up to the 2000 national ticket party ticket and presidential race, suggesting that they had limited long-term thinking when it came to how to use their websites moving forward (something that Barack Obama perfected with the evolution of his online presence). 

A Closer Look 

In researching this article, and Peter Vallone’s unconventional use of campaign funds for banner ads, I began to uncover a little more detail surrounding how the $100,000 would be spent.

I managed to contact Kevin McCabe, and get his view on the campaigns efforts. Despite the lack of enthusiasm for allocating budget towards digital strategy in the wider political environment, McCabe and others working for Vallone clearly saw its potential, and how its importance was only going to increase.  

McCabe recognised that there was no social media to “support the effort and broaden the base”, but the “electoral tech evolution” had begun already, something which he thought would surely turn into a revolution in the near future.  

“Our online presence was clearly effective although limited by lack of funds” explained McCabe in back-and-forth emails over several weeks.  

“Only 40% of the electorate was using the internet at the time”. This claim is supported by VNS Exit Polls from 1998 showing that only 40% of the New York electorate ‘used the internet regularly’. A combined CBS & New York Times poll from the same period suggested that 60% of the electorate had never used a computer to connect to the internet.  

The Strategy Behind the Strategy: E-Voter 98 

These numbers may be discouraging when looking for a strategy that would make a difference in the polls, but it starts to make more sense once you research the hypothesis behind this digital media buy.  

A case study released in January 1999 called ‘E-Voter 98’ that I discovered after trawling through countless links and articles, features a section written by Alan Gould that effectively serves as a PCA (post campaign analysis) of the digital strategy.  

Here, Gould explains that; “using online ads, we would reach a limited number of highly motivated voters and engage them in a discussion of key issues that the media was ignoring”. Here we can deduce that the team in 1998 were focused on audience refinement, and not campaign reach.  

This is also supported by the estimated percentage of this audience that were registered voters (“70% of The New York Times on the web users”). Gould also writes that other segments they deemed to be important had been identified, but not acted upon due to limited funds: “The original online media plan included web sites whose viewers were more female and younger, but again, with our reduced budget, funding one impactful buy on The New York Times on the Web was going to be difficult enough”.  

Geographic information, in this case zip codes, that a user would have to provide to the paper to sign up to their site was useful first party data the NYT could provide as an indicator of whether a user would be registered to vote in the elections for the city.  

Two strategic plays were then made with this data.  

First, the campaign would segment the zip codes to target ‘suburban voters around New York City’, an area which held strong support for Pataki, Vallone’s opponent.  

The second was to heavily target Manhattan, as many of the suburban voters targeted in the first strategy would commute into the city to work. For Manhattan alone, Gould reported that they were able to produce a unique reach number upwards of 150,000 people, and 600,000 impressions overall, resulting in a 4.0 frequency per user (whether this was deliberate, I cannot tell). 

When reading a PCA that is 20 years old, one thing that sticks out is how primitive display banner ads were as a channel. Gould explains that the NYT offered three banner types; small, medium and large (no advanced native format strategies here unfortunately).  

The clearly defined audience that was being targeted also allowed for a clear messaging strategy that would “take advantage of the higher than average educational and income levels of The New York Times on the Web users”.  

Bringing these aspects together, it shows that the team had a clear idea of how they wanted to use this space.  

The last question that remains, then is this: 

Did it help? 

The short answer to this question is: no.  

As mentioned previously, the lack of funds to fully realise what the NYT data could be leveraged for could be seen as a warning sign that the strategy was not going to be as effective as possible.  

Alan Gould gives more context to the plan, suggesting it wasn’t a complete failure. As Vallone was outspent “10 to 1 on television”, and in a race that the media had “decided was over in September”, the display activity was still able to drive 600,000 sessions to the campaign website. 

As the campaign had correctly predicted, digital was a seed that would soon grow into a vastly complex network of ideas and strategies and a billion-dollar industry. Much has been written over the last couple of years about the use of online advertising in political races due to Cambridge Analytica, Facebook and others.  

Sometimes to understand where we are, we have to understand how we got here in the first place. 

On this page