Is there a larger study available? Did you find online ads were more important towards raising funds from activists or as a persuasion tool for Republicans/moderates? Were banners worth the cost, and on what kind of sites did they work best (news, entertainment, sports)?

Veeeery interesting, as they say. Joshua Shimkin has posted a response over at EchoDitto -- since I share his questions, I'll quote them here:
"...persuasion and registration successes need be included into the mix. Michael, do you think the marketing campaign was a success? What types of ads worked best? Did you see any difference in effectiveness given the length of the ad, style (i.e. cartoon, video of candidate, just text with no animation, etc), or subject matter?"

Thanks for the questions, Oliver, Nicco, and Joshua. It's encouraging to see increased interest in this topic. I hope these answers are helpful.

First, a clarification, and then I'll address the specific questions above.

Joshua Shimkin noted in his response on EchoDitto that "These results show that the marketing effort didn't break even if you measure success only by money." It's my fault for not being clearer.

The study only reported dollars raised from specific fundraising banners (i.e. Give $50 Now). There were, however, a few different campaigns: Fundraising, Post-Debate, Last Minute Persuasion, Voter Registration, etc. The fundraisiing campaign posted very high returns, which, in turn, allowed the DNC to launch non-fundraising campaigns.

But here is the interesting news: even the non-fundraising campaigns drove positive returns. The DNC's Post-Debate efforts, for example, drove people to a site explaining why Kerry won the debate and why Bush lost. At the bottom of the page, well below the fold, there was a link to contribute. As you can guess, people clicked to contribute and gave enough money to completely offset the costs of running the banner campaign, and then some. Unfortunately, we weren't tracking donations very closely for the persuasion campaigns, but it's something that I wish we had been monitoring. Thanks for bringing up that point.

Did you find online ads were more important towards raising funds from activists or as a persuasion tool for Republicans/moderates?
Hands down, the ads were considerably more important for persuasion. Each of the studies we conducted found that our ads greatly impacted those who were exposed to them. It was much more important to persuade the 100% of people exposed to the ads rather than try to get money from the less than 1% who would actually click and make a contribution. Fundraising was always a goal, but persuasion was much more important, especially in the final month of the campaign.

I hardly ever suggest using banner ads for fundraising. If making money and signing up supporters is your goal, stick to the Blogs and contextual ads on Google. To educate and persuade the electorate, use banners.

Were banners worth the cost?
Banners are a bargain. For less than a penny per impressoin and less than 1% of their entire advertising budget, Kerry and the DNC managed to purchase over 800 million ad impressions on leading websites. If you didn’t see one of their ads, you were probably living under a rock. The Web is by far the least expensive advertising medium and can be extremely effective in reaching highly targeted, motivated audiences. In addition, creative costs are tiny. So, yes, they were certainly worth the cost.

On what sites did they work best (news, entertainment, sports)?
Ads performed best in terms of fundraising, persuasion, and click-through on national news sites like CNN.com and Yahoo! News. Local newspaper websites also performed well. However, a lot had to do with the creative message and the context in which the ads were placed. On sites like iVillage.com and InStyle.com, for example, ads addressing issues of importance to women performed better than standard fundraising or persuasion ads.

We also did a campaign for www.twofacesofbush.com, placing a slew of funky ads on sites like MaximOnline.com and Space.com. These were some of the best performing ads of the entire election. It really has a lot to do with the creative and the placement.

Timing is also key. Ads featuring John Edwards started running within 24 hours after he was selected as Kerry's running mate. These ads gave a tremendous boost to the banner-driven fundraising efforts.

Surprisingly, ads on sports and weather websites also performed well. Here’s my theory. If you have enough time in the day to check the weather and see who won last night’s game, you also have enough time to click on a banner ad. In addition, banner ads are great tools to reach information-seekers. Sports and weather sites certainly fall into the information-seeking category.

Was the marketing campaign a success?
I think I’m a bit biased, but since there are facts to back up my assertions, I feel safe in saying that the campaign was a success. Kerry’s online efforts were limited in scope and scale. However, we proved that their attempts to raise money, influence voters, and register Americans living overseas far exceeded expectations in terms of both direct response and persuasion metrics.

Which types of ads worked best?
In terms of fundraising, the most simple and straight-forward ads performed the best. Cartoons were also helpful for fundraising at the onset of the campaign (novelty was the driving factor, we suspect), but they lost out when tested head-to-head with basic appeals, such as “If You Want Kerry To Win In November, We Need Your Help Today” placed next to a nice photo of Kerry.

Ads that tried to convey a large, complex message failed miserably unless they were genuinely funny or cute. Rich media (i.e. flash, streaming video, roll-over tabs, cartoon animations) worked best for persuasion. Video and cartoons stood out in helping all of our clients generate buzz and free media coverage for their online marketing efforts.

Some other things to keep in mind:

1. Creative Limitations: One of the most important features of an advertisement on any medium is the overall visual quality of the ad. The challenge all designers face with the Internet is how to portray a message within a small banner ad that looks great and also fits within website file size limitations (usually less than 20k).

2. Legibility: Ads should be easy to read. The font size of the text (and disclaimer, if applicable) should be large enough to read without straining and should stand out on the page.

3. Animation Pace: If the ad is animated, it should animate at a speed that allows you to read and digest the entire message.

4. Message Salience: The average Internet user is exposed to 50-75 different banner ads each time they go online. While political ads tend to stand out from others based on the relative importance of their content (i.e. we don’t all need a new mortgage, but we are all impacted by the leaders we elect), you should look to build ads that stand out and attract attention.

5. Message consistency: Banner ads do not exist in a vacuum. Rather, they are placed above, around, and in the middle of pages filled with content. Perhaps a banner catches your eye when you first surf to a new webpage, but perhaps you only catch the ad after the animation has come to a stop (usually after 15 seconds). You should be able to read and understand the message and purpose of the ad, even if you catch it mid-way though or after it is done animating.

6. Clear call to action: Banner ads generally attract click-through rates in the .25% to .5% range. As such, ads are almost always designed to appeal to the 99% who will not click on them rather than the small percentage that will. However, the Internet is an interactive media and the call to action should be a clear indication of what the advertiser wants the voter to do (i.e. vote, learn more, contribute, etc.)

Thanks again and keep the questions coming...

It would be interesting to compare online versus offline conversion metrics. An analysis could be helpful in proving the potential of online marketing to the political sector. Can you place the online spending in context of the campaign's overall media spend? Did it outperform other media efforts in persuasion and fund-raising? Should more have been spent online? And lastly, how did the online marketing integrate with the offline if at all?

Hart...the answer to each of your questions can be summed up by an answer to the last question you pose. I'll deal with the final question first, then address some of the others below:

1. How did the online marketing integrate with the offline if at all?
Online marketing activities were conducted in a vacuum. Over the course of the entire election, I can't recall a time when online ads mirrored offline messaging. You never saw a TV commercial asking for money or trying to register voters. Persuasion ads that ran on television were never re-done for online usage. Generally speaking, demographic targets sought on TV were not targeted online.

In many ways, this is a good thing. It shows that the campaign understood that the Internet is a different medium, capable of doing things that traditional media cannot.

However, this is also part of the problem most campaigns face. General consultants do not understand online communication, so they do not fully integrate the Web into greater campaign goals.

One exception was geographical targeting. Online ads were targeted to the same states that were running television ads. But aside from that, there was no coordination between the traditional and interactive campaigns.

To that end, your other questions are relevant, important, and critical to our understanding of how the Internet fits into a greater marketing campaign, but we do not have access to any data relating to the offline efforts.

However, all hope is not lost. We have some very compelling cross-media studies regarding consumer packaged goods (CPG) advertisers. These studies, incidentally, were conducted by Dynamic Logic, the same company who did all of the Kerry and DNC online ad effectiveness studies.

I know what you’re thinking right now. “How can you compare selling soap or detergent to political communication?” There are some similar metrics (i.e. product awareness = name recognition), but CPG is particularly helpful when trying to determine what percentage of a media budget should be spent online.

Don’t take my word for it, though:

- Levi's Type1 Jeans X-Media Study

- Quaker Oats Breakfast Squares

Some other findings from Dynamic Logic:

- Unilever Branding Campaign: The exercise revealed that spending 15% of Unilever's media budget online provided a 24% lift in branding metrics, compared with a 19% lift from plan with the same total budget, which only spent 2% online.

- Colgate Branding Campaign: Whereas advertising on television alone increased purchase intent by 3.4 percent, Colgate was able to increase purchase intent by 4.3 percent by keeping the overall budget the same but allotting 11 percent to the Internet.

- Non-Profit Advertising: The closest we have to political marketing research is another great study from Dynamic Logic on non-profit advertising. It's not a cross-media study, but still interesting. It found that:

"Exposure to online advertising leads to significant impact on all key metrics including: increasing awareness of the organization or candidate (up +5 points on average), building message association (+8 points on average), raising favorability (+2 points average) and increasing likelihood to donate, volunteer or vote (+2 points average).

2. Should more have been spent online?
To recap (or preface my conclusion), the vast majority of banner ads purchased by the DNC and Kerry were focused on fundraising. They drove positive returns and also proved that fundraising ads positively impacted voter perception. We also conducted a handful of other campaigns, from voter registration to post-debate education, that all exceeded expectations.

It goes without saying that 2004 was a huge step in the right direction. More was spent in 2004 than in 2000 and 2002 combined. More than I had predicted. As an industry/medium, we're on the right path and it's hard to look back and criticize what was, at the time, a great step forward.

But, to answer your questions directly, more should have been spent online. Absolutely. Following numerous Dynamic Logic cross-media studies and from studying the budget allocations of major US advertisers, I've always advocated spending about 4% or 5% of a campaign's overall media budget on the Web. Most people who have heard me say this think I'm crazy, but there hasn't been a single published study I've seen that has advocated anything less. Here’s a chart showing online spending as a percentage of offline budgets of all online advertisers, as compiled by Jupiter Research:

I couldn't make this stuff up if I tried!

Hart, thanks again for your questions. Sorry there isn't more specific data on the political sector. We're working on it!

How did blog ads perform compared to other sites? Which systems did you use? What part of the spend went specifically to blog placement?

Were text ads like Google AdWords used too? How did they do? What strategies worked best?

Inside the primary campaign, how did the conversation go? Who had to buy in? What objections were raised and overcome with regard to the online spend and creative? How did the online team get buy-in?