in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc Oregon – 4th Congressional District Summary   For the Art Robinson for Congress campaign, Cambridge Analytica SCL [CA-SCL] assumed a comprehensive set of responsibilities and effectively managed the campaign in its entirety, with strategic advice channeled through US nationals on the CA-SCL team. Based on a subjective analysis of the district on our arrival, Dr. Robinson was starting from a position of significant reputational deficit compared to his previous two campaigns. Two previous unsuccessful campaigns hindered the team’s fundraising efforts; roughly half of the financial resources raised in 2010 and in 2012 were generated. This impacted the campaign’s ability to communicate with voters seriously, and particularly in the critical final weeks of the campaign. Qualitative and anecdotal evidence gathered during 2014 canvassing activities and from Dr Robinson showed clearly that the team successfully shifted voter perception of the candidate; voters became more receptive to messaging. Limited funds and the candidate’s preference for less personalized (and less effective) direct mail outreach the campaign lacked the volunteers, funds and leadtime to effectively deliver issue-based messaging to newly receptive voters. Had CA-SCL been involved and deployed earlier in the campaign cycle, ideally 6 to 9 months before November, it may have been possible to have a greater impact on the campaign and address external risks more effectively. Ultimately, the lack of funds and volunteers to undertake paid or volunteer canvassing resulted in limited capacity available to deliver messages and persuasion scripts to the voters via face-to-face direct voter contact – the most effective medium for persuasion – which could have increased voter interest in Dr. Robinson as a candidate. Overall the result was less than positive, but this race was effectively unwinnable due to the factors outlined above and fundamental electoral math. This does raise questions about whether Western Oregon and CD4 is a viable target for Republican and/or conservative candidates. in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc     Background   CA-SCL’s engagement with the Robinson for Congress campaign had been designed in advance as a supportive intervention to augment an existing campaign infrastructure. Once on the ground, it became clear that no such professional ‘campaign team’ existed, despite sterling efforts made by the Robinson family and supporters. The campaign’s initial efficacy was impeded by a number of external factors: • Dr. Robinson’s support base is predominantly rural or in small towns making it difficult to mobilise volunteers in population centers. • The Oregon Republican Party is severely lacking in human, financial and infrastructural resources and riven by factional infighting. There was little coordination between campaigns, open animosity between staff on the Wehby and Richardson campaigns, and between both of these and Dr Robinson’s supporters. • Although his uncontested drive for the nomination saved financial resources, the lack of a primary meant that Dr. Robinson lacked the exposure that would have come from a longer public campaign. • Candidate was again endorsed by the Constitution Party, but not the Independent or Libertarian Party as he had been in previous years. This likely cost him several thousand votes which could have narrowed the margin of loss. • The candidate’s reputation with the voters was toxic. o Damage done by the incumbent’s messaging in the previous two campaigns had never been effectively countered. Consequently Dr. Robinson’s personal reputation was significantly tarnished. o Dr. Robinson’s decision to solicit urine samples by direct mail under his own name less than a year before the election weakened his positives yet further. o The first night that the canvass vendor canvassed in Eugene, the canvassers had a refusal rate in excess of 80% amongst registered Republicans. o There was also a large quantity of negative content on the internet to overcome. Items such as the Rachel Maddow MSNBC interview, previous attack ads, and materials of the Robinsons’ own creation (the curriculum site/video, for example), which contained numerous unfashionable views, in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc • made it difficult to get traction with target segments of the electorate in CD4. The data showed an unfavourable pre-existing voter intention playing field: o 152,285 registered Democrats o 130,627 registered Republicans o 121,607 independent voters § 104,141 independent voters below .55 partisanship (likely Demleaning) § 17,466 independent voters above .55 partisanship (likely Repleaning) Design  of  Communications  Strategy   The plan developed by the team was aimed at rehabilitating Dr. Robinson’s image with voters by presenting him as a sympathetic family man and serious scientist rather than as the extremely right-wing, unstable ‘mad scientist’ caricature created by the opposition over the previous two campaigns. Using direct mail, digital ads and signs, the team developed and executed a communications strategy designed to rehabilitate Dr. Robinson’s image with voters. This focused on four main themes that were informed by CA-CA-SCL’s research and the candidate’s relative strengths and interests: • • • • education policy economic policy and job creation national security Dr Robinson’s personal history as an accomplished scientist and single parent to 6 children. In addition, a significant and extensive door canvass campaign was designed to contact voters. This was running for about two weeks before the candidate decided that the expense associated with this door-to-door effort was too great relative to their fundraising expectations. This was despite the campaign team’s best efforts at attempting to convince them of the efficacy of direct voter contact, and the canvass subcontractors were contacted directly by the Robinson family and asked to cease operations. Before the canvass was suspended, the canvassers reported that the rehabilitation message was working, and that voters who they believed would have previously been hostile or refused to participate were remarking on the direct mail and digital ads while taking the survey. The team designed a program of public forums; events at which local voters could hear Art give a short fifteen-minute presentation on a specific theme (jobs, national security, or education) and ask questions. These were also intended to act as in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc canvass anchors; something that canvassers could talk to local voters about. In addition to these events, the candidate was also deployed to meet with voters in small cafes and fast food restaurants across the district. The candidate was asked to spend his remaining time meeting voters either through mainstreeting (talking to voters and business owners in retail business districts) or knocking on doors in residential areas. Deliverables   The campaign team provided the following deliverables: • • • • • • • • • • • Creation of a new ‘Art for Congress’ website and integration with CRM platform and social media Media relations management Provision of communications strategy, including key topics and slogans Provision of talking points, speeches, debate prep Logistical support and planning for events and candidate travels Design of printed campaign materials including direct mail, leaflets, and billboards Fundraising via email and direct mail that raised c. 500k Management of paid and volunteer telephone canvas Management of paid and volunteer door-to-door canvassing Planning and deployment of extensive robocalling GOTV effort Management of campaign social media platforms Lessons  Learned   During the months of September and October, a number of risk factors arose that likely had a detrimental impact on the campaign’s ability to increase vote share: • • • The campaign only managed to raise half of the amount of the previous campaigns, which was likely due in part to fatigue and waning interest amongst donors both in CD4 and across the United States following two previously unsuccessful campaigns. Due to the poor fundraising results, the canvass program had to be suspended after just a few weeks. Not only did this eliminate most of the direct voter contact the campaign was planning to undertake, but it also meant that a campaign strategy organized around events and public appearances that were not optimal without a direct voter contact effort could no longer be pursued as planned. The candidate preferred visiting businesses to visiting residential addresses. While this likely had limited impact on the campaign, it did limit the contact the candidate had with some demographics of voters. in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc • • • The infighting and backstabbing in the Oregon Republican Party prevented us from being able to recruit many young Republican volunteers. Several volunteers were ‘poached’ by candidates for statewide office. The majority of people who were interested in volunteering for Art did not want to canvass doors or make phone calls. This further limited the campaign’s ability to engage in a meaningful ground game. Timing issues and problems with communication between the campaign team and Dr Robinson’s local core supporters resulted in an inconsistent branding of signs. Many signs installed in the district made use of old campaign branding and messaging from previous campaigns, which diminished brand consistency and did not effectively communicate the 2014 campaign messaging. The new sign designs went through multiple iterations at the behest of the Robinson family, all of which were posted at roadsides, resulting in new signs that lacked branding consistency even though they conformed to the approved messaging. Given the limited financial resources available to the team, the GOTV effort consisted mostly of paid calling and attempting to mobilize some volunteers to make persuasion calls. • • • • • Dr. Robinson won the first debate and tied the second, but since very few voters watched the debates and local media provided limited coverage, it had no discernable impact on the campaign. The campaign suffered reputational damage as the result of the opposition publicizing the involvement of PACs in the campaign, which presented a particular challenge because the campaign had previously portrayed Dr Robinson as a political outsider not beholden to special interests. This reputational damage was magnified, however, by the fact that the work of these independent expenditures was not visible and seemed to have little impact. This undid much of the disaggregation messaging the campaign had been able to communicate to registered Democrats and non-affiliated voters, while seemingly yielding very little in increased voter turnout. Our opponent raised more than twice as much money as our campaign. Our opponent saved the majority of his financial resources till the last 3 weeks – at which point the Robinson campaign had exhausted most of our money. Since we had no choice but to rehabilitate Dr. Robinson’s image early in the campaign, the team had little ability to counter the negative, well-funded attacks in the last 3 weeks of the campaign. Based on the nature of our opponent’s attacks in previous cycles, the campaign was able to pre-empt attack messaging by addressing these issues first. We successfully inoculated voters with regards to Dr. Robinson’s reputation as a scientist and researcher, and our messaging on Dr Robinson’s previous thoughts on education seemed to have some (albeit limited) impact. The socially-conservative nature of Dr Robinson’s core support and the socially-liberal disposition of large in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc segments of the electorate (particularly in Eugene and Lane County) made it almost impossible to successfully tackle life/choice issues, however, and the campaign chose not to message on these. In the end, this formed a significant portion of Rep. DeFazio’s late-campaign GOTV messaging, though it is very difficult to see how this could have been pre-empted. Conclusions   • • • • • • • • The Art Robinson for Congress campaign delivered similar results as the 2012 campaign while spending less than half the money. The candidates pre-existing reputation put the campaign in a hole that was simply too hard to climb out of. Research clearly shows that direct, face-to-face voter contact is the most effective form of persuasion and get out the vote efforts. Without a significant canvass, the air war, which was itself limited by lack of funds, simply could not overcome the reputational deficit. The IE gave a political weapon to our opponent while producing little if any impact on the voters. The campaign needed to be able to respond to the DeFazio messaging in the final 3 weeks in both the air and on the ground, but lacked the funds to do so. High quality data and a well-run campaign cannot invent voters who don’t exist. The predictive data always showed that there were more voters naturally inclined to vote for the Democrats than for Dr. Robinson. To overcome this kind of voter inclination deficit, even a candidate without pre-existing negatives would have required a budget at least equal to, if not significantly higher than, that of a wellloved and respected incumbent. With Dr. Robinson’s pre-existing reputational deficit, it would have required far more time and a significantly increased budget to win this election. It is not clear whether Dr. Robinson or a member of his family are capable of winning an election in CD4. A qualitative examination of online message boards associated with coverage of this November’s election shows that the Anti-Robinson messaging developed by the Democratic campaigns still has significant traction amongst the electorate. Polling would offer some insight into whether a Robinson is electable, or whether Republican/Conservative political electoral interests would be better served by selecting a new candidate that the Robinsons could throw their support behind. The Robinson family still commands real devotion amongst their support base, and while these voters are not significant enough in number to win an election, victory without their support will be very difficult to achieve. In the opinion of CA-SCL, Western Oregon should be considered an essentially risky proposition for funders of Republican/conservative candidates. Evidence from all sources including voters’ party registrations, past electoral results, and modelled partisanship and issue scores indicate that the prospects for future Republican victories in the Western third of Oregon (including CD1, CD3, CD4 and CD5) are in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc • dim. To present but a few examples, Western Oregon has bucked national trends in each of the significant ‘Republican Waves’ of the last 20 years. In 1994 Republicans won only additional seat in Oregon, which was lost two years later. And in both 2010 and 2014, national Republican waves failed to cross the Cascade Mountains as all Democratic incumbents remained in place. If funders wish to continue supporting political candidates in Western Oregon, an immediate propriety should be to establish and resource a field operation ASAP to begin the process of in-person message delivery and volunteer recruitment to best prepare for the 2016 election. in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc Colorado – State Senate Districts 16, 19, 20, 22, 24 Executive  Summary   The CA-SCL team in Colorado supported several Colorado State Senate campaigns, and made a demonstrable contribution towards Republican victories in 3 of the 5 districts where CA-SCL staff were engaged. In Senate districts 16, 19, 20 and 24, the CA-SCL team worked with Colorado State Senator William Cadman and the Senate Majority Fund to support the Republican candidates. In district 22 CA-SCL created a pro-life independent expenditure campaign, the Centennial Coalition, promoting the candidacy of Republican candidate Tony Sanchez. Though there were some initial issues related to the practicalities of integrating the CA-SCL team into the pre-existing campaign structures, the venture was ultimately successful. The CA-SCL team produced substantial quantities of creative work including direct mail and digital communications products, door canvassing and phone scripts, lists of voters for direct mail and other contact mediums, and other products and services that made a substantial contribution to the three victorious Republican campaigns. These victories ultimately resulted in the GOP taking control of the Colorado State Senate. Background   Upon arrival, CA-SCL were asked to identify the four most winnable State Senate races from a list of seven provided by Senator Cadman and the Senate Majority Fund. Following a process that involved analyzing probable voting outcomes based on modeled partisanship, the CA-SCL team selected four races (SSDs 3, 16, 20 and 24), though SSD 3 was replaced by SSD 19 after Senator Cadman and his staff expressed a special interest in working with the candidate in that district due to her demonstrated aptitude for campaigning. There were initial difficulties in rolling out the campaign, however: • Local campaign operatives had limited understanding of what CA-SCL was providing, and this created confusion as all parties tried to work out how the teams would work together. It also seemed (at first) that the local Senate in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc • • • Majority Fund operatives were reticent to share their own voter ID data and strategy with the CA-SCL team. Senator Cadman and his team were initially unsatisfied with the choice of SSD 24 as one of the options for CA-SCL intervention, but we were eventually able to use the data to illustrate that it presented a better opportunity for achieving a Republican victory than other options. This was ultimately the correct choice, as Beth Martinez Humenik’s victory in SSD24 proved to be the margin that allowed the GOP to retake the State Senate. The relatively embryonic nature of the RIPON platform at the time of the team’s deployment meant that some of the activities early on in the campaign had to be undertaken with assistance from the data teams in London and Victoria, but these ‘teething problems’ were overcome by the time the campaign was fully underway. The Senate Majority Fund’s operatives had chosen their own printer for mail products, and the operatives’ requirement that all printing go through them generated delays, as did As previously mentioned, the CA-SCL team was also contracted to design and implement a pro-life independent expenditure campaign supporting the candidacy of Tony Sanchez in SSD 22. The modeling predicted a Democratic win by a significant margin (in the region of fifteen percent), and SCL designed and disseminated a complete package of communications products as well as implementing a program of telephone canvassing. Design  of  Communications  Strategy   In the districts where CA-SCL worked with the Senate Majority Fund, the team identified key issues likely to be salient with the target voting populations. These were: • • • • Economic issues Education policy 2nd Amendment rights National Security The team designed and wrote copy for issue-specific mailers to be disseminated to target populations according to modeled issue salience and psychographic profiles. For the first six weeks following deployment, voters would receive one issue-specific direct mail product per week, with the focus shifting towards the promotion of specific candidates and their individual selling points in the later stages of the campaign. in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc In the final two weeks of the campaign, the CA-SCL team prepared GOTV mail pieces and lists of voters identified as likely Republicans with moderate propensity to vote. The messaging on these GOTV mail products was also formulated for specific psychographic profiles identified amongst the target population. In total, the team designed over 80 pieces of targeted direct mail, working with operatives from the Senate Majority Fund to physically produce and disseminate these to the voting population. Informal feedback received from Senator Cadman indicated that he was very pleased with the literature produced by the CA-SCL team, which was described as some of the best he had ever seen. Deliverables   The campaign team provided the following deliverables: • • • • • • • • • • • • Strategic communications plan(s) Profiles of voters in modelled issue and personality clusters Plan for programme of persuasion canvassing GOTV communications plan 80+ pieces of direct mail (written and designed) Scripts for door-to-door and telephone canvassing Redesigned Senate Majority Fund website Communications products for social media Door hangers, designed for different personality clusters Issue-based fundraising emails Lists of modelled fundraising targets 446 lists of voters generated for targeted communications Lessons  Learned   While the project in Colorado was largely successful, some issues did present themselves. This was particularly the case in the early stages of CA-SCL’s engagement with local campaign staff: • • There was significant confusion amongst local staff regarding the role that CA-SCL operatives would be playing in the campaign. This generated problems early on, and although these were overcome, making our team’s role clearer from the outset should be a priority in future. One aspect of this integration process that arose was the vested interest that some local staff had in using certain vendors. In some cases this generated tensions, with in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc • local operatives seeking to restrict communication between the CA-SCL team and printers, for instance. These difficulties can probably be overcome by CA-SCL being clearer about our role and the supporting infrastructure required earlier in the campaign cycle. Colorado Republican staff are accustomed to seeing data visualized in formats that are quite different from those used by CA-SCL, and whilst they were able to understand the data once it was explained to them, it may be worth considering whether there are other ways to present data that could be more familiar to local staff.   Conclusions   The CA-SCL team in Colorado was able to make a major contribution to victories in three of the five State Senate races we engaged with, and results for Republican candidates in the other two were better than our modeling or client expectations would have predicted. Overall this is a very positive result, and one of the victories gave the GOP control over the Colorado State Senate. There were definitely issues that can be addressed in future, particularly concerning the relationship between CA-SCL and local teams and how CA-SCL integrates with local efforts. On the whole, however, these issues were overcome by additional effort and high-quality output that impressed local operatives, and with a clearer mandate on future projects and more planning regarding campaign integration it should be possible to have a similarly positive impact on other campaigns in future. in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc North Carolina – NCGOP Executive  Summary   The CA-SCL team in North Carolina was contracted by the North Carolina Republican Party to provide data and analysis to campaigns at several levels during the 2014 election cycle. The team was involved in providing data, particularly lists of voters that could be contacted by direct mail, canvassing or other means, as well as analysis of what CASCL’s data on partisanship, turnout likelihood and psychographics actually meant for campaigns. Day to day activities also involved checking text and graphics work against the psychographic profiles of target populations to ensure that the messaging was as targeted as possible. Cambridge Analytica’s research and analysis had determined that the most salient issues were the economy, immigration, issues related to firearms ownership, and national security issues. Throughout North Carolina, Republican campaigns had focused their messaging strategies on issues such as Obamacare that, in some cases, were not actually as salient as Republican operatives believed them to be. The CA-SCL team was able to use data and research analysis to prove, however, that the campaigns would be better served talking about the more important issues. The most significant campaign supported by CA-SCL’s data was Thom Tillis’ campaign for the US Senate, which was successful by a narrow margin. It is the opinion of the CA-SCL team that their work enabled crucial economies to be achieved through the successful microtargeting of messages, and that this made a meaningful contribution towards the GOP victory. Background   Upon arrival, the CA-SCL team was placed with the Raleigh branch of the Republican Victory programme, and began working to support several Republican campaigns at different levels. in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc In the early phases of the deployment, there was real confusion amongst local staff regarding CA-SCL’s role in the campaign structure. This issue continued throughout the campaign period, and this is something that should definitely be addressed in future. There was also confusion amongst both local campaign staff and the CA-SCL team regarding concrete deliverables, and precious time was lost in August after the team was deployed as both sides attempted to work out how the resource of CASCL’s data and analysis could best be integrated into and utilized by the campaigns. Design  of  Communications  Strategy   Many of the Republican campaigns in North Carolina, at least those for Federal office, were focusing on issues such as Obamacare that CA-SCL’s research had indicated were not the most salient in the minds of the target populations. According to the primary research and modeling performed by CA-SCL and Cambridge Analytica, the issues most likely to be salient to North Carolina’s target voter populations were: • • • • the economy immigration 2nd Amendment rights national security National security, for example, is was particularly topical and significant, and the Tillis campaign’s focus on National Security during the election period has been cited by several news outlets as one contributing factor to the traction the campaign’s messaging was able to achieve amongst the electorate. Using these issues data and the other modeled information in the voter database, the CA-SCL team put together a series of briefing packs for the different races that profiled target segments of the electorate and outlined which issues they would likely respond to. Also provided as part of each briefing pack were sample messages on these key issues, illustrating how the psychological profiling could be employed to better appeal to likely or possible Republican voters. Deliverables   The campaign team provided the following deliverables: in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc • • • Briefing packs for specific races, which each included lists and psychographic profiles of key voter groups (Certain/‘Hardcore’ Republican voters, Likely/‘Reliable’ Republican voters, priority persuasion targets, and Republican turnout targets) Ongoing support through the provision of lists of voters defined by the candidate, as well as advice on how direct mail and other interventions could be better targeted. In many cases, local campaign staff had ideas about how they wanted their target universes defined, but the CA-SCL team was able to use their knowledge of the data to suggest more effective targeting strategies. In many cases, these suggestions were accepted and the campaign staff acknowledged that the universes could be better defined according to parameters outlined by CA-SCL. In some cases, however, there were (occasionally persistent) difficulties in explaining CA-SCL’s data to local campaigns, since the CA-SCL universes had been defined using terminology and methodologies that the local staff were unfamiliar with. The CA-SCL team provided ongoing support on messaging, checking prepared creative such as mail pieces and digital graphics to ensure that the visual and text elements were as effective as they possibly could be for the intended target populations. Lessons  Learned   Data and Analytics: • • • In the event our turnout scoring was flawed in North Carolina, and our prediction for ballots cast was over 300,000 too high. We also need to separate early, absentee & E-day voting in our modelling and turnout predictions. We need to have ability to call up & manipulate turnout data as the GOP does, both for practical & teaching purposes. (I.E. 2/4 voters etc.) We need to be able to display data in the format GOP operatives expect to see it moving forward. This means that categories like 4/4 voters, 2/4 voters, registration rates, precinct-level results (both historical and live) all need to be easily displayed. Our work in moving campaigns to a new model requires we be able to translate operatives from how they see the world now to how we want them to see it. In some cases, it seems that the quality of data provided by vendors was not acceptable. Unfortunately this did not become clear until the campaign was already underway, and the team was able to work around this, but it is something that should be addressed earlier in the campaign cycle going forward. in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc Client Management: • • • • The lack of clarity about deliverables prior to our arrival on the ground bedevilled us throughout the campaign. Ultimately one of the key outcomes from this lessons-learned process will need to be a series of identifiable, appropriately priced, deliverables for the company. Despite the substantial spending totals most campaigns live a very ‘hand to mouth’ existence on the fundraising front, so the ability to atomize costs will be valuable. The focus of a campaign business is General Consultants and Independent Expenditure groups, not candidates. Vendor selection is almost never a candidate choice. There are rich opportunities for cost-sharing among clients, based on primary clients devolving elements of our billing on their other clients/partners. This, in combination with a quality deliverable and cost matrix, will make market penetration and alliance building much easier than it has been to this point. Ripon Platform: • • • There are a series of User Interface issues with Ripon, however, which needs to become much more intuitive and flexible in its search and group management features. The reporting functions in Ripon need to be richer and more interactive. Ultimately for both strategic planning and client-side issues we should phase out Tableau on the field side in favour of improved functionality for Ripon. We need greater speed in integrating data entered through Ripon, with Ripon reflecting a database/model that updates automatically and in real time.   Conclusions   CA-SCL’s engagement in North Carolina was a significant success, with Thom Tillis’ victory constituting a major win for the GOP nationally. CA-SCL’s operatives were also successful in developing good working relationships with Republican strategists in North Carolina, and despite some early difficulties involving integration into the local campaign teams, the feedback the team received was overwhelmingly positive. in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc BOLTON SUPER PAC – AR, NC, NH   Executive  Summary     CA-SCL was contracted to provide messaging and communications support to Ambassador John Bolton’s super PAC, which was supporting Republican candidates in Arkansas, North Carolina and New Hampshire. Primarily concerned with promoting the Ambassador’s agenda with regards to national defence and foreign policy, the PAC made use of significant input from SCL on messaging and target audiences, with positive results. CA-SCL contracted an external firm to conduct a phone survey to test attitudes amongst target populations, which also provided significant evidence of the efficacy of CA-SCL’s messaging. Relative to an uncontacted control group, respondents in the study who had been targeted with Bolton Super PAC messaging on foreign policy and national security showed a significant increase in their awareness of these subjects.   Background     The Bolton Super PAC campaign had three goals: 1) Elect Republican Senate candidates in Arkansas, North Carolina and New Hampshire, 2) Raise the issue of national security, 3) Raise Ambassador Bolton’s profile. To serve those ends, CA-SCL were contracted to work with existing PAC staff to craft and deploy a communications programme in the target states.     Design  of  Communications  Strategy     SCL worked closely with Bolton Super PAC, Campaign Solutions (the PAC’s comms vendor), and various media vendors to produce a campaign plan that included two rounds of advertising and many different ads, each targeted to a specific cluster of voters in target states. There were some hiccups in terms of data delays and limited functionalities with the online distributors, but overall the ad creation process went smoothly. SCL data provided invaluable insight into the voters it needed to persuade and turnout in order to be victorious in the three Senate races. The media teams took in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc direction well and worked with Harris MacLeod (SCL) to ensure each message was tailored in a way that would resonate with its target. Having an SCL operative on the ground enabled the Bolton campaign to dive further into the data and modify its approach for the second round of advertising. Whereas the first round of ads were targeted mainly on personality and were aimed at persuading swing voters, the second round of ads targeted larger groups of both persuasion and turnout targets, which were clustered together based on their common personality traits, demographics and priority issues. The end result was a suite of diverse, powerful and persuasive ads that were deployed online and using Direct TV.   Deliverables     • Microtargeting data for Arkansas, North Carolina and New Hampshire o Complete voter universes in each state, with voters grouped according to partisanship and turnout scores, as well as personality traits o A Libertarian/Third Party model was also created to address the strength of a third party candidate in North Carolina • Issue data o Each cluster was assigned a ranking of priority issues based on what we know about those voters o Each cluster also contained priority foreign policy issues based on testing we did after receiving feedback from Ambassador Bolton • Message testing o Each cluster was also tested on various images and political messages, and campaign creative teams were given further guidance based on which messages resonated most with target groups • Voter files o Work closely with Campaign Solutions, the firm that coordinated the Bolton campaign, and online ad distributors to supply necessary data to purchase ad space and deploy messages directly to voters • Guidance on interpreting data o Insights from focus groups conducted in target states to feed into language and themes used to address target voters on particular issues o Direction and feedback on all creative, including personality-targeted ads for online and Direct TV, and political emails to target voters in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc o o • Strategic grouping and regrouping of voter clusters for maximum efficiency, and to adapt and respond to the changing dynamics of the Senate campaigns in each state Work closely with Campaign Solutions to make judgments on how best to use data and deploy messages to voters Post-campaign study o SCL contracted an outside firm to conduct a phone survey of several groups of targeted and non-targeted voters in each of the three Bolton states in order to test effectiveness Lessons  Learned   • Communication between SCL, Campaign Solutions and Bolton Super PAC could have been better in the lead-up to Harris MacLeod and Mark Gettleson’s (SCL) arrival in Washington DC. o Data delays were compounded by SCL not fully understanding that voter files needed to available in a specific format in order to purchase and reserve online advertising slots. However, once Harris and Mark were on the ground SCL was able to move quickly to provide this in time for the ad buy and the campaign was not adversely affected. o The project could also have benefitted from SCL and Campaign Solutions working together more closely on formulating the media strategy and rollout plan for the campaign. This would have allowed SCL to tailor its data suite to the needs of the campaign at an earlier stage in the process, which would likely have led to higher quality advertising released on a schedule agreed upon by all stakeholders. • While the SCL data was excellent, both in terms of the insight it provided into voters and our ability to send them messages directly, the campaign would likely have benefited from some online polling data in the three target states in order to determine which candidates were the most ahead/behind at various stages of the campaign, and also to ask the electorate any specific questions to gage opinion about events in the news. Having this additional data point would have enabled us to tailor the campaign messages to an even greater degree, and to allocate resources to the tightest races where it could make the most difference. • The RIPON platform was extremely helpful in creating new groupings of voters to target for the second round of advertising in the Bolton campaign, but it could be improved further by making the program more user friendly, better at data visualization and clustering, and with more data points for each cluster (ex. more demographics, more consumer, and a more sophisticated description of the personality traits and interactions). in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc • AIQ (creator of the RIPON platform) would likely have been a better online ad and email distributor than the one used by Campaign Solutions for the Bolton campaign. o CS’s distributor claimed at the beginning of the campaign to have the functionality to track viewership of ads per individual voter, however it turned out they were only able to do this by cluster. AIQ proved in other campaigns where it acted as the primary ad distributor to have more advanced tracking abilities that were used to make the online ad campaign more adaptable based on which ads were performing best with particular voters. Conclusions   Overall, CA-SCL’s engagement with the Bolton PAC can be described as reasonably successful. Two of the three closely fought races (Arkansas and North Carolina) were won by Republicans, while in the third (New Hampshire) the Republican Senate candidate did not win but finished much stronger than expected. Furthermore, all three candidates made national security a central pillar of their campaign messaging, indicating Bolton Super PAC was successful in influencing the public conversation and raising the prominence of national security as a key issue with voters. The randomized control trial study undertaken after the election is also very valuable, and shows that the messaging deployed by the PAC was successful. In each state, 10 groups of voters were identified. Five of the groups in each state are comprised of voters who received online ads from Bolton Super PAC, while the other five are comprised of voters with similar personality traits and demographics, but who did not receive ads from the campaign. In total, over 6,589 voters were surveyed across the three target states. The results are impressive, and demonstrate that the campaign achieved its three goals of 1) Elevating national security as a key concern among voters; 2) Raising awareness of Ambassador Bolton and his organization; 3) Persuading target voters to vote for the Republican Senate candidate in their state. Below are a few examples of the most statistically significant changes in voter thinking due to Bolton Super PAC campaign messages. in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc DIGITAL - Jobs, Growth and Freedom Fund The creative ad units were tailored to SCL’s behavioural models and were primarily issue driven with strong calls to action reminding the audience members to vote. Target audience criteria was selected to help narrow our impressions to the desired audience characteristics including: demographics, zip code, DMAs, geolocation, likely partisanship, consumer habits (both observed and modeled), user agent, display retargeting and user matching. These narrowcast selections ensured the accuracy of advertising targets while at the same time allowing for spillover to occur which naturally caught those audience members who fell just outside of our optimum target universes. In all 320 creative pieces were created and loaded into the respective display networks and targeted using the filters outlined above. There were a total of 55,706,388 impressions served, for an average display frequency of 30 times over the lifetime of the campaign, or 3.5 impressions per audience member per day. The digital campaigns drove approximately 99,500 unique users from our target universe to the campaigns respective landing pages. The campaign budget was distributed across three main channels favoring web and mobile each at 35% of the total campaign budget with the remaining 30% being allocated to Facebook. This allowed us to capture the attention of our target audience whenever they were online - browsing the web on their cell phones or computers, in mobile applications, as well as on Facebook. The rational for this approach was driven by candidate awareness, primary issues, and the need to be ever-present in the primarily turn-out driven races that were focused on. in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc The top performing campaign banners based on click through rate follow below: New Hampshire: CTR: 0.22% CTR: 0.22% CTR: 0.21% Iowa: CTR: 0.38% CTR: 0.35% CTR: 0.33% Colorado: in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc CTR: 0.33% CTR: 0.33% CTR: 0.27% in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc DIGITAL – For America The creative ad units were tailored to SCL’s behavioural models and were primarily issue driven with strong calls to action reminding the audience members to vote. Target audience criteria was selected to help narrow our impressions to the desired audience characteristics including: demographics, zip code, DMAs, geolocation, likely partisanship, consumer habits (both observed and modeled), user agent, display retargeting and user matching. These narrowcast selections ensured the accuracy of advertising targets while at the same time allowing for spillover to occur which naturally caught those audience members who fell just outside of our optimum target universes. In all 434 creative pieces were created and loaded into the respective display networks and targeted using the filters outlined above. There were a total of 242,306,627 impressions served, for an average display frequency of 212 times over the life of the campaign, or 7 impressions per audience member per day. The digital campaigns drove approximately 289,443 unique users from our target universe to the campaigns landing pages. The campaign budget was distributed across three main channels favoring web and mobile each at 35% of the total campaign budget with the remaining 30% being allocated to Facebook. This allowed us to capture the attention of our target audience whenever they were online - browsing the web on their cell phones or computers, in mobile applications, as well as on Facebook. The rational for this approach was driven by candidate awareness, primary issues, and the need to be ever-present in the primarily turn-out driven races that were focused on. in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc Using detailed audience modeling techniques, SCL created banner advertisements based on the most likely issues in each target audience. Some were broad based issues, while others spoke to the individual candidate issues in each campaign. Examples of the highest CTR banners in each race are as follows: Nebraska CTR: 0.33% West Virginia CTR: 0.29% Minnesota CTR: 0.35% Washington CTR: 0.33% in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc Oregon CTR: 0.35% Colorado CTR: 0.33% Arkansas CTR: 0.34% Iowa CTR: 0.32%   in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc DIGITAL – Bolton PAC The creative ad units were tailored to SCL’s behavioural models and were primarily issue driven with strong calls to action reminding the audience members to vote. Target audience criteria was selected to help narrow our impressions to the desired audience characteristics including: demographics, zip code, DMAs, geolocation, likely partisanship, consumer habits (both observed and modeled), user agent, display retargeting and user matching. These narrowcast selections ensured the accuracy of advertising targets while at the same time allowing for spillover to occur which naturally caught those audience members who fell just outside of our optimum target universes. In all 188 creative pieces were created and loaded into the respective display networks and targeted using the filters outlined above. There were a total of 18,220,521 impressions served, for an average display frequency of 20.5 times over the life of the campaign, or 1.86 impressions per audience member per day. The digital campaigns drove approximately 34,000 unique users from our target universe to the campaigns respective landing pages. The campaign budget was distributed across three main channels favoring web and mobile each at 35% of the total campaign budget with the remaining 30% being allocated to Facebook. This allowed us to capture the attention of our target audience whenever they were online - browsing the web on their cell phone or computer, in mobile applications, as well as on Facebook. The rational for this approach was driven by candidate awareness, primary issues, and the need to be ever-present in the primarily turn-out driven races that were focused on. The top performing campaign banners based on click through rate follow below: New Hampshire: CTR: 0.38% CTR: 0.34% in Partnership with The News Corp Building, Suite 2703 1211 6th Ave, New York, NY 10036 (262)-617-2716 or (202)-509-1181 info@SCLgroup.cc www.SCL.cc CTR: 0.34% IOWA: CTR: 0.40% CTR: 0.38% CTR: 0.35%