Within the frame of the robocall scandal, pundits across Canada seem to be talking a lot about “microtargeting.” Political parties are compiling information about us and using this information to segment their political markets and surgically deliver their communications and organization. They know what kind of cars we drive and therefore they know how we’re going to vote. This – the pundits say – is standard operating procedure, these days.
This description, while eerily familiar, is actually about one part truth and two parts fiction. Although research and sociodemographic selection are hallmarks of American political campaigns, the principles of research-based campaigning are not always applied, applied inconsistently, or applied with varying degrees of success in Canada, especially at the riding level. This is due, in part, to the limited resources available to our local teams. Riding-level campaigns generally have less capacity than their central command centres. Holding a clear recognition of this challenge, political parties have pushed to empower campaigners with new and innovative tools to deliver higher performance campaigns. One highly effective solution that has become a standard tool in the Canadian campaign toolkit is a locally accessible, centrally integrated and managed database. Data is information and information is becoming the foundation of today’s increasingly intelligent campaigns.
Within a simplified version of the current voter identification process, constituents’ party support is recorded “on the ground” through phone and door-to-door canvasses, administered by or on behalf of the riding-level campaign. The data is then inputted, aggregated centrally, and used to focus local and central campaign communications, and get-out-the-vote resources during advance-poll voting periods and election day (and possibly other purposes – like fundraising – during interim campaign periods). This process is the most important operational function of a political campaign. It’s often quite complex and managing voter identification data is both an art and a science. To a certain degree, Conservatives now know who the Liberals are in a riding and Liberals know where to find the Conservatives, as well as the NDPs and Greens. Parties know where the non-voters are located, as well. They won’t tell you this but they ignore those folks as much as possible.
The term microtargeting, however, means something completely different to political consultants and other campaign professionals than simply indentifying supporters. It includes the practice of applying mathematical analysis over top of large datasets to predict individual voter behaviour. Microtargeting analyzes the responses to survey questions collected from a sample population and builds statistical models that accurately predict likely responses for people in the electorate who haven’t been contacted.
Microtargeting models quantify the chance of an elector’s response to questions such as “Which party or candidate do you support?” This likelihood is calculated as a score and is used to rank electors in a riding. Obviously, very few ridings have the resources to microtarget potential supporters given the level of sophistication required to administer the process. Such efforts are usually managed out of our capital cities, nationally and provincially, by central campaigns.
More often, today’s traditional door-to-door and phone canvasses are being complemented or even supplemented by centrally-led microtargeting operations, providing local, user-friendly data through fully integrated, online databases. Nevertheless, neither the ubiquity nor the effectiveness of microtargeting should be assumed. Parties and campaign teams may be attempting to increase the capacity of their voter contact programs to include microtargeting techniques but it’s still a relatively new process. Campaigns are making mistakes, are still learning, and much of the expensive knowledge transfer is coming from south of the border. There are different statistical models used by many different teams, results have been varied, and candidates – especially veterans – are often slow to warm to such a hands-off approach to voter contact. The commitment from party insiders, on the other hand, seems to be rock solid.
To help manage and microtarget Canadian political markets, today’s campaigns are indeed becoming research-centric. This should not be a surprise to any Canadian. Each one of us are in multiple databases based on our own socio-demographic particularities. Google and Facebook monitor and sells us. Our smartphones geo-locate us. And political campaigns identify us and now increasingly analyze and microtarget us.
Although still far from standard operating procedure, we are getting closer and closer to experiencing political communications and organization that is as intelligent and sophisticated as our American counterpart. But we’re not quite there – yet. Who knows, the robocall scandal may impact future Canadian elections positively. Reforms could force politicians to remove the calculus and actually connect with all voters. We may become targets the old fashioned way - without the math.
Original Article
Source: ipolitics
Author: Martin Canning
This description, while eerily familiar, is actually about one part truth and two parts fiction. Although research and sociodemographic selection are hallmarks of American political campaigns, the principles of research-based campaigning are not always applied, applied inconsistently, or applied with varying degrees of success in Canada, especially at the riding level. This is due, in part, to the limited resources available to our local teams. Riding-level campaigns generally have less capacity than their central command centres. Holding a clear recognition of this challenge, political parties have pushed to empower campaigners with new and innovative tools to deliver higher performance campaigns. One highly effective solution that has become a standard tool in the Canadian campaign toolkit is a locally accessible, centrally integrated and managed database. Data is information and information is becoming the foundation of today’s increasingly intelligent campaigns.
Within a simplified version of the current voter identification process, constituents’ party support is recorded “on the ground” through phone and door-to-door canvasses, administered by or on behalf of the riding-level campaign. The data is then inputted, aggregated centrally, and used to focus local and central campaign communications, and get-out-the-vote resources during advance-poll voting periods and election day (and possibly other purposes – like fundraising – during interim campaign periods). This process is the most important operational function of a political campaign. It’s often quite complex and managing voter identification data is both an art and a science. To a certain degree, Conservatives now know who the Liberals are in a riding and Liberals know where to find the Conservatives, as well as the NDPs and Greens. Parties know where the non-voters are located, as well. They won’t tell you this but they ignore those folks as much as possible.
The term microtargeting, however, means something completely different to political consultants and other campaign professionals than simply indentifying supporters. It includes the practice of applying mathematical analysis over top of large datasets to predict individual voter behaviour. Microtargeting analyzes the responses to survey questions collected from a sample population and builds statistical models that accurately predict likely responses for people in the electorate who haven’t been contacted.
Microtargeting models quantify the chance of an elector’s response to questions such as “Which party or candidate do you support?” This likelihood is calculated as a score and is used to rank electors in a riding. Obviously, very few ridings have the resources to microtarget potential supporters given the level of sophistication required to administer the process. Such efforts are usually managed out of our capital cities, nationally and provincially, by central campaigns.
More often, today’s traditional door-to-door and phone canvasses are being complemented or even supplemented by centrally-led microtargeting operations, providing local, user-friendly data through fully integrated, online databases. Nevertheless, neither the ubiquity nor the effectiveness of microtargeting should be assumed. Parties and campaign teams may be attempting to increase the capacity of their voter contact programs to include microtargeting techniques but it’s still a relatively new process. Campaigns are making mistakes, are still learning, and much of the expensive knowledge transfer is coming from south of the border. There are different statistical models used by many different teams, results have been varied, and candidates – especially veterans – are often slow to warm to such a hands-off approach to voter contact. The commitment from party insiders, on the other hand, seems to be rock solid.
To help manage and microtarget Canadian political markets, today’s campaigns are indeed becoming research-centric. This should not be a surprise to any Canadian. Each one of us are in multiple databases based on our own socio-demographic particularities. Google and Facebook monitor and sells us. Our smartphones geo-locate us. And political campaigns identify us and now increasingly analyze and microtarget us.
Although still far from standard operating procedure, we are getting closer and closer to experiencing political communications and organization that is as intelligent and sophisticated as our American counterpart. But we’re not quite there – yet. Who knows, the robocall scandal may impact future Canadian elections positively. Reforms could force politicians to remove the calculus and actually connect with all voters. We may become targets the old fashioned way - without the math.
Original Article
Source: ipolitics
Author: Martin Canning
No comments:
Post a Comment