How to win elections using data

European Union.  European stats, charts, financial market
Data analytics has become essential in understanding and winning voters.

Oh boy. We enter the season when we are overwhelmed by political upheaval, eclectic memories, and revisionist history. A lot of overly generalized messaging is designed to get out of the base – those voters who always support the party’s candidates no matter what crazy things their candidate says or does. This is the way elections have been held for generations because they appeal to the simplest “us against them” instincts. This strategy is working well… for Al Qaeda.

But something different is needed to win elections today, which is the ability to attract the rapidly growing number of independent or “swingers” voters. according to “The rise of independents‘, the number of independent voters – votes not associated with one of the main political parties – is increasing:

Prior to World War II, more than 80% of voters belonged to one of the major parties. Today, the plurality of American voters – in some polls, more than 45% – defined as a freelancer. “

This is correct. Prior to World War II, less than 20% of voters classified themselves as independent. Today this figure is greater than 45%.

As politics fails to attract more Americans, politicians will need to adopt a different approach to appealing to those important “swinging” independent voters who will decide our political leaders. So how does the modern political campaign triumph in this changing world?

Nobody is average.

As discussed on my blog”The risk of making decisions based on averages“, The roughness principle He asserts that by measuring a range of traits across a sufficiently large number of individuals, approximately half of the individuals will be above average, and nearly half will be below average for any given trait. And that of all traits, few (if any) will be “average”.

Every person is unique. To influence these important swing voters, you will need to appeal to each of them individually. Here comes the role of nanoeconomics. Nanoeconomics is the economics of expected human behavioral tendencies and performance. Campaigns must embrace nanoeconomics to create individual messages if they are to successfully influence these swing voters.

It is very easy to think that there is one important political metric – to collect more votes than your opponent. Very black and white. You’ve got winners and losers. game over. Let’s shoot “Dragon House”.

But politics is about more than just the current elections. It’s about developing “followers” who share similar (but not always the same) beliefs and principles. Thus, one needs to think more broadly about identifying, validating, and prioritizing KPIs and metrics against which political and campaign success can be measured (Figure 1).

2-1 . slice

Figure 1: Completed political campaign Value Creation Design Template

example political campaign The design diagram in Figure 1 ensures that there is consensus and alignment among the major components on how the party will define its value creation efforts and identify key performance indicators and metrics against which the party will measure value creation and the effectiveness of the political party’s goals.

There is a lot of publicly available and third party data about each individual voter. see blogHere data brokers are quietly buying and selling your personal information“For examples of personal data that anyone can purchase from third-party data brokers. They will surprise and terrify you (Fig. 2).

slice 1-2

appearance 2Your personal data for sale:

Third-party data provides a good start in creating individual messages and actions…but it’s just a start. Your organization needs to expand its voter data capture and policy specification to improve the predictive accuracy of your voter tendency models. This is where empowering frontline workers (investigators) will be invaluable.

Intelligent organizations (not just political parties) seek to empower frontline workers — those employees who are on the front lines of customers and operational engagement — to engage their natural human intuition to identify those tiny, subtle insights that may be better predictors of behavior. Editing main ideas, design-based mindset supports”Think Like a Data Scientistmethodology (Fig. 3).

Slide3-1

appearance 3The Art of Thinking Like a Data Scientist

Front line employees play a key role in data science and the value creation process by:

  • Define and validate key performance indicators and metrics against which the effectiveness of value creation will be measured.
  • Brainstorm the features that ML models will seek to improve in creating key analytic tendency scores.
  • Identify and measure costs associated with false positives and false negatives ML models (a step many organizations tend to skip).

Analytical A score is a mathematically modified number that predicts a particular outcome or the probability (or tendency) of a particular action. Analytical scores are used to track performance and make decisions industry

An analytical score can be created for each major business or operational decision to improve the accuracy, relevance, and effectiveness of decision making. The most common example of an analytic score is a credit score that predicts an individual’s likelihood, or propensity, to repay a loan (we all have unique credit scores).

In our political example, we want to create an analytic “probability vote for a candidate” score that measures the probability or probability that someone will vote for that candidate.

Ultimately, we want these analytical scores to be acquired, shared, and continually improved so that they can be used to guide messages and actions across multiple political uses. We will store these analytic scores in an analytics profile to facilitate this.

Analytical features An invaluable data model for capturing, recording, sharing, and refining individual analytical scores (expected behavioral tendencies and performance). Analytical profiles enable the application of analytical scores to drive accurate actions and decisions in political uses, including voter targeting, campaign messages, voter donations, voter volunteering, and other political outreach programs and activities (Fig. 4).

4-1 . slice

appearance 4: Analytical Profile of Bill Schmarzo Voter…This is just an example…

This is really where the rubber meets the road. This is where you can apply your Likelihood to Vote (LTV) and Likelihood to Vote for Your Candidate (LTVC) scores to target your campaign efforts, messages, and outreach programs to swing voters who can be persuaded to vote for your candidate and turn up to vote (Figure 5).

5 . slice

appearance 5Focus resources on the most influential voters

Three observations about voters are represented in Figure 5:

  • imprisoned votersVoters with more than 80% “likelihood of voting” and “likelihood of voting for your candidate” are probably worth only a small investment. They can only vote once (legally), and unless you’re trying to increase their donation and volunteer (which requires a different set of analytical findings), it’s best to focus resources elsewhere.
  • Missing votersVoters with less than a 30% “likelihood to vote for your candidate” result are probably not worth the investment. You won’t change their mind (remember they are the other party’s rule, and the rule can be very tolerant of their candidate’s deceptions). Remember, not many Americans can be convinced that the Earth is not flat…
  • You should focus most of your time and effort on these swing voters. Getting analytical results on every major political issue (eg women’s rights, gun rights, immigration, jobs, education, climate change) is critical to further improving messaging and efforts to reach swing voters with “correct” beliefs and attitudes from swing to your filter.

Final note: Never rely on one score to make your decision. In our political example, a high “likelihood to vote for a candidate” score doesn’t mean much if the same person has a low “likelihood to vote” score.

As more Americans become political independents, the old campaign strategies of shooting at your base will not suffice. “Playing on the base” may be enough to get you nominated by your party but unlikely enough that you will be elected in a general election.

Thus, it is time for political campaigns to embrace the best practices of data science (nanoeconomics, analytic outcomes, analytic filings, and Think Like a Data Scientist methodology) that are used by today’s leading business organizations to achieve more meaningful business and operational outcomes.

This means treating each voter as an individual (not an overly general group) and investing time and effort to learn more about their individual interests and aspirations. Now, isn’t that what politics is supposed to be?

Leave a Comment