THE UNTOLD STORY OF POLITICAL TECHNOLOGY IN 2004
Benjamin A. Katz
Politics & the Internet in 2004
In the 2004 election cycle, websites, email and online fundraising assumed a growing prominence. In each of these areas, new high marks were established in both volume and audience-reach.
However, 2004 also saw the emergence of a powerful new set of web-based tools that I have come to call Distributed Campaigning. While these second-generation Internet tools could eventually prove even more valuable than on-line fundraising, they also have the potential to inflict serious damage to a campaign. What is Distributed Campaigning?
“Distributed campaigning” is an adaptation of the term “distributed computing,” which is defined as “the use of multiple computers networked throughout a wide geographical area to solve a single problem.” Distributed Computing was popularized by the SETI@Home
Project in which at-home users could set up their computers to assist the SETI project when they were not otherwise in use.
The home computers would automatically download sets of data to work on, process them and send the result back to the main SETI servers. Using this distributed computing method, the SETI@Home
Project has effectively built the largest and most powerful supercomputer in the world.
In Distributed Campaigning, political campaigns likewise maximize their power and impact through Internet-based tools that allow supporters to easily assist the campaign from their homes. For example, via the web, volunteers could download lists of voters, call them, and send the results back to the campaigns’ website. Uses of Distributed Campaigning in 2004
With the requisite financial resources available to invest in new technologies, the presidential campaigns were generally the first to adopt and use distributed campaigning as a tool that allowed them to reach out to a nationwide audience and volunteer base. The presidential campaigns were also quick to adopt the best ideas from their competitors, so most of the key distributed practices were seen in multiple campaigns.
The most common distributed technique was also the simplest – using the web as an easy way to create and distribute campaign literature. Rather than shipping campaign materials and paraphernalia throughout the country, the campaigns provided posters and campaign literature (in a digital format, such as PDF) that could be easily downloaded and printed by supporters from their homes. This kept costs down and made campaign literature far more available than was possible in previous campaigns. The Bush-Cheney campaign took this a step further by allowing visitors to their site to create and print a customized campaign sign, such as “Texans for Bush-Cheney.”
While providing literature was a key first step in distributed campaigning, most of the presidential campaigns expanded on the idea by creating kits for house parties that could either be downloaded or requested from the website. In addition to the literature and paraphernalia, these kits included signup sheets and contribution envelopes that could be completed and sent back to the campaign headquarters.
However, the most ambitious distributed campaign effort undertaken by the presidential campaigns was in the use of distributed phone banks. First implemented during the democratic primaries, these web-based systems provided campaign volunteers with short lists of registered voters, their phone numbers, and a call script. Volunteers could login to the system from their homes, make calls, and immediately enter the results.
The benefits to the campaigns were tremendous. A traditional phone bank is limited by geographic location and the number of phone lines. Data entry is limited by the number of computers. These distributed phone banks were effectively unlimited. There was also a huge cost savings. Rather than having to pay for additional office space, phone lines and computers, the campaigns only had the relatively minor costs of the servers and the bandwidth. Campaigns implementing distributed phone banks also saw a value in immediate data entry. As volunteers made calls, the campaigns had immediate access to the results and could use them for analysis and to prepare follow-up materials. Down-Ballot Distributed Campaigning
While the presidential campaigns were the most visible users of distributed campaigning, their efforts still relied on traditional methods. On the other hand, many candidates in Congressional and local races, with fewer resources to spare, ran campaigns that were built entirely upon distributed campaign strategies.
In 2004, my company assisted several countywide political organizations in setting up web-based voter tracking systems. These systems allowed the easy creation of satellite offices. With minimal overhead, new offices could be set up in borrowed or rented locations. Just as soon as the location had Internet access, they could begin printing walk sheets and call sheets and doing data entry from these locations. In many cases, key volunteers’ homes became de-facto campaign offices where other volunteers could meet for campaign walks and pick up campaign walk sheets and kits.
In Orange County, California, a voter outreach PAC took this a step further. Jennifer Turner, the PAC’s executive director, explained how they coped as the need for data entry increased substantially in the final weeks of the campaign. While the campaign office would generally close around midnight, trusted volunteers would take data home and work on it all night, so that it would be available for use in follow-up the next day.
Even smaller campaigns took distributed campaigning still further--placing key campaign roles and data in the hands of volunteers working from home. Unlike larger campaigns, many city and other local campaigns cannot afford professional assistance for fundraising, accounting and volunteer coordination. In over 100 local races around the country, we provided our BackOffice system. BackOffice allowed these campaigns to share information securely between key staff and volunteers over the web, so that they all could work together despite the lack of a central campaign office.
While these campaigns were dealing with far fewer volunteers than the presidential campaigns, distributed campaigning was even more critical in their daily operations and in maximizing the value of the volunteers. Perils of Distributed Campaigning
Despite the exciting and powerful impact of distributed campaigning, it is not without risks. Poorly implemented distributed campaigning can damage a campaign.
The first and biggest area of concern for a distributed campaign is the lack of a verification method. For example, in a traditional phone bank, the campaign staff can easily listen in on volunteers’ calls to ensure the proper message. However, with distributed campaigning there is no way to check if volunteers are staying “on message.” There is no way to determine if they are giving their own well intentioned but damaging message; actually calling on behalf of the opposition or not making the calls at all.
A key technical staff member for one of the Democratic presidential campaigns admitted (after the primaries) that they knew that the other campaigns had infiltrated their distributed phone bank. Other than discounting the accuracy of the data entry, there was little they could do about it.
In a similar abuse of a distributed method, the Bush-Cheney campaign customized campaign sign tool was abused by Democrats who created signs with messages such as “Zillionaires for Bush-Cheney.”
Another common problem with distributed campaigning in 2004 was the duplication of efforts. With volunteers working from a wide-variety of locations, it was often difficult for the campaigns to ensure that the same voters were not accidentally contacted multiple times. Interestingly, distributed computing is generally designed with duplication built-in to ensure accuracy of result and to avoid gaps in the result set. While this works very well with data, it does not work with voter contact and can have negative impacts. For example, I personally received 3 calls in a single evening from a campaign that was presumably using a flawed distributed phone bank.
Finally, campaigns using distributed campaigning face new security risks. The more information they provide to supporters outside the campaign office and the more volunteers they allow to access that information, the greater the peril that the opposition can access, corrupt or delete key information. While this risk is not specific to distributed campaigns, it is amplified by the lack of physical location access controls. Best Practices
While distributed campaigning is still a relatively new methodology, a set of best practices has begun to emerge. These practices allow the greatest value of distributed campaigning while minimizing, if not eliminating the associated risks.
The highest priority in ensuring a high-quality and secure distributed campaign is the proper control of access. Home access should, whenever possible, be even tighter than what would be allowed in the campaign office. Justin Meyer, the campaign manager for a highly contested Congressional primary explained that they would regularly have trusted volunteers do data entry from home. “However, because of the highly sensitive nature of our financial data we were hesitant to allow people to access the data from home. [We] set control limits as to what information each user could access . . . and were able to take precautions without wasting any able volunteers.
In fact, in many cases, properly implemented distributed campaigning can offer a better level of security than traditional campaigns. Placing key campaign information on a properly secured, password-protected and access-managed server can ensure a better level of security than the traditional Microsoft Access database or Excel spreadsheet that is sitting unguarded in the campaign headquarters.
The second key step in implementing a distributed campaign is to recognize its limits. For nearly all campaigns, allowing the message to be shaped in the field by distributed-campaign volunteers is harmful. The Bush-Cheney campaign learned this lesson with their custom sign creator, eventually replacing it with a much more limited version. Democrats saw a similar problem with MoveOn’s “Bush in 30 seconds” ad contest. While some very powerful ads were created and activists were excited by this unconventional distributed message generation, Republicans latched onto the couple of ads that compared Bush to Hitler. In many ways, “Bush in 30 seconds” harmed the Democratic movement more than it helped. Apart from the message, the most successful campaigns we’ve seen limit distributed campaign to either very specific people or very specific tasks where inappropriate use cannot cause significant damage.
Finally, a key component of a successful distributed campaigning strategy is choosing the right technology partner. One large voter registration and GOTV effort in 2004 went through a lengthy vendor selection process. At the end of the process, they decided upon a company with substantial experience building and managing large web-based systems in the private sector. However, this company had no experience with political campaigns. In the final weeks before the election, the system was unable to handle the volume of traffic it was under and effectively shut down. The company was completely unprepared for both the increase in use and the immediacy of a campaign’s needs. While campaigns that used vendors specializing in the political space also had problems, generally these vendors were more prepared to understand and address the unique challenges of campaigns. What to expect in 2006 and beyond
As on-line fundraising was born in 1996 and grew in both volume and complexity, distributed campaigning will grow in 2006. We can expect a greater number of campaigns to be using it, and using it for more of their efforts. Campaigns that fail to take advantage of the power of distributed campaigning will find themselves at a competitive disadvantage. By the end of the decade, distributed campaigning will undoubtedly be the norm--not the exception.