Campaigns really DO Matter: Robocalls, Polling and the electoral impact of Big Data

Two weeks ago, British Columbians were preparing to cast their ballots in an election that was widely expected to be  the NDP’s to lose. You see, public opinion polls on voting intentions had been published showing an almost insurmountable gap between the NDP and the second place Liberals. The polls were mostly properly conducted, they reflected the stated intentions of representative, and sizeable samples of the BC Electorate. The condidtions for accurate polling were in place, but lo and behold, when the votes were counted, the Liberals had achieved a bare majority, apparently snatching victory from the jaws of defeat. This was not the first time that publicly released polls of voting intentions missed the boat big time. In the Alberta Provincial election, the Wildrose Party went from top of the polls, to a pretty sorry second place at the finish line. In the last Quebec election, the Liberals evaded a much ballyhooed oblivion with a strong last-minute finish. There has been a lot of gnashing of teeth over the ‘failure’ of the publicly released polls to accurately predict an election, and untold column inches spilled over analysis of how to compensate for turnouts, and what sampling methodology is failing and what is working. Well in my opinion, the answer is as simple as the headline from this (absolutely correct) CBC Article: “B.C. election proved campaigns matter more than ever”

It is worth noting that in all three Provincial Elections with surprising results, it was the incumbent, Governing Party that defied the odds. Is that a coincidence? IMHO, that is not coincidental. You see, it is the premise of this post that we are living in a New World of Electoral Politics. This world is where big databases are parsed and segmented according to tastes, preferences, affinities, and location. This is the world where elections are won or lost by closely targeting very specific groups of voters, crafting, testing, and verifying the message that can sway these micro-groups. Delivering that message effectively ON TARGET and building the support of a bare plurality, one riding at a time is the primary objective. This world is highly dynamic. It focuses on electors who are undecided, or relatively uncertain of their first choice in voting intentions. In such a world, it is incredibly important to understand what motivates small, select, and distinct groups of the electorate. A little appreciated fact of life in Canada is that Governments have nearly infinite resources for opinion polling, focus groups, and acquiring sophisticated datasets about target groups. It is not surprising to me that the results of untold millions of dollars worth of Government funded opinion research are considered a  State secret. In  nutshell, the incumbent will always have a massive advantage in any election, inasmuch as they can easily assemble a complete picture of who can be motivated by which message long before the writ is dropped, and the electorate is in play.

So what do Robocalls have to do with this post? Robocalls are a big data communications tool. As has been illustrated by the recent finding in the robocalls law suit, it is easy to abstract list of contacts based upon any number of variables (Like which Party they support). The mechanics of targeting and launching a robocall campaign, whether large or small is as simple as falling out of bed. And the cost of robocalling is so damned cheap, that virtually any political party that possesses a robust electoral database can flood the phone lines with negative messages for pennies a pop. It is not just robocalls. Robocalls are mildly irritating to the recipients, so their best use is to sway the OTHER sides voters with them. They are tailor-made for vote suppression tactics, like pointing out the opponents failings to his or her supporters. Email communications are an even more versatile tool, because they are even cheaper than robocalls, and they are good for positive, vote winning communications with reams of people on short notice. Then there are more conventional message delivery media, like direct mail, admail, live telephone calling, door to door foot canvas, and at the broadest level of all, targeted print and broadcast media buys. These all have one thing in common, that their efficacy improves dramatically when they are targeted based on issues research, and solid information on large numbers of electors.

Campaigns are dynamic events, and they DO matter. When the media commissions a voting intention poll, they are basically paying for a snapshot, at a moment in time, and the snapshot they are taking is intended to reflect an amalgam of the entire electorate. The individual campaigns couldn`t really care less about what the entire electorate is thinking. They are focusing on small subsets of the electorate, and directly influencing them to vote one way or another, or even not to vote at all. The effective campaigns have access to hard data on what messages will work with what segment of the electorate, and there are multiple campaigns each working on influencing small numbers of voters to tip one way or another. There is just no way for low-budget media bought voting intention polls to keep up, and accurately predict what is actually going to happen, because the decisions and actions that will determine the election result are happening under their radar. All this campaigning is happening in real-time, and it can be extremely effective. Campaigns are each working simultaneously on different population segments, so polling for, and capturing the significant movements is kind of like trying to model next weeks weather. There are just too many variables interacting to draw any useful conclusions. The media polls are not focused on the groups that are actually in play, and there is no way that a polling company can predict just what kind of voter suppression, and voter conversion campaigns are happening, and how effective they are. This is perfectly illustrated by a quote from this CBC Article interviewing the BC Liberals internal pollster, Dimitri Pantazopoulos:   `”At the end of the day the more important thing is understanding what motivates people to vote, and how to actually communicate with people and what are the underlying factors that will turn the undecideds one way or another.” And you know something? Dimitri was the one measuring the effectiveness of the BC Liberal Campaign, so I am perfectly willing to take his word for it.

Vote for this post at Progressive Bloggers!