Organizations of all stripes are still analyzing their impact on the last midterm cycle to calibrate their efforts for 2020.
As these autopsies continue, the term cost per vote (CPV) is getting thrown around quite a bit. It’s a concept people put a lot of stock in without truly understanding what it means—and that can end up costing organizations.
CPV is a metric used to gauge the relative value of different tactics for investment and implementation. It’s one of the most useful ways to measure electoral work. It can be helpful in developing plans, measuring impact, and for evaluating results—if calculated correctly.
But in too many cases, CPV will be calculated incorrectly, justifying bad electoral funding decisions, promoting yet unproven civic engagement programs, and undercutting more expensive tactics.
Clearing up the misunderstandings around CPV calculations will help improve the way impact is assessed and investment decisions are made in the future. It will also help evaluate what it takes to win, rather than what the cheapest program is.
Estimating CPV is a bit more complicated than it appears. First and foremost, it’s the likelihood of changing an election outcome with a marginal vote, not the average cost of obtaining a single vote. This is rooted in the winner-take-all nature of elections. Think about CPV as a function of the likelihood that an election is won by a single vote and how much the last vote costs.
Here’s the calculation: Cost of changing election outcome = (marginal CPV) / Probability (win by 1 vote)
Thinking about it in terms of the cost of changing the election outcome, it’s easier to understand that CPV isn’t simply the number of Democratic votes divided by the total amount of money spent by Democrats. Rather, it’s calculated by comparing the votes generated by a group treated with a campaign tactic against an identical control group.
When done this way, it’s easier to understand the following.
Elections aren’t like commercial marketing.
In elections, the probability of winning by a single vote is extremely low — either because a candidate is extremely likely to win or extremely likely to lose. This is different from commercial marketing, where incremental gains matter. In commercial marketing, winning over the marginal customer has value, regardless of absolute market share. For instance, Coke and Pepsi still fight over customers despite Coke’s dominance.
Each marginal customer added is evaluated against their acquisition cost to determine if the acquisition was “worth it.” Compare this to elections, where adding a marginal point of support in a non-competitive race (e.g. increasing vote share from 10-percent-to-11 percent or 89-to-90 percent) has no impact on the outcome and as such is a waste of resources.
The last vote matters a lot.
One of the biggest problems with evaluating election tactics based on CPV is that it is rarely done by calculating how much the last vote costs. Obviously, money should be spent on the least expensive votes first. But stopping with the most cost-efficient tactics and voters will rarely secure a win. It’s more likely to get to 48 percent of the vote in a two-way race when 50-percent plus one is needed.
For example, one of the most efficient tactics is sending cheap official-looking social-pressure mail for turnout operations. This tactic typically costs $100-150 per net Democratic vote. This is great, but consider that there are a finite set of voters reachable through mail and that set of people narrows further when considering those who respond.
This limits the total population of people dramatically. Consider that 10 percent of any political geography are good targets for GOTV mail and the mail has a 1-percent turnout effect. This program will help generate a boost of 0.1 percent for Democrats, which will only make the difference in the very closest races.
To be clear, a 0.1-percent boost is nothing to sneeze at. But it isn’t enough. Changing election outcomes means going down the list of tactics, funding the cheapest new vote at each step. As a result, the cost of the marginal vote will go up with each tactic. This means that tactics with higher CPV should be evaluated by asking how they build upon cheaper tactics being used.
Not all votes are created equal.
When evaluating the true cost of an election, it’s important to recognize where there’s more bang for the buck. Consider three scenarios:
- An additional Democrat to vote in Manhattan;
- An additional Democrat to vote in a competitive congressional district; or
- An additional Democrat to vote in a competitive congressional district that is nested in a competitive Senate race.
Getting a Democrat to vote is a positive outcome anywhere, but getting them to vote in a competitive congressional district is even better, and nesting that inside of a competitive Senate district is best of all. Intuitively this makes sense, but it’s important to fit this into the framework above to evaluate the cost correctly.
When engaging voters in nested districts, the number of contests where that vote may decide the outcome is double or tripled, effectively cutting costs proportional to the likelihood that any race is decided by a single vote.
There’s a better way to use CPV.
An oversimplified analysis on CPV threatens election contests that are otherwise winnable—and failing to use CPV at all threatens election contests by wasting money that could go to good use. CPV must be analyzed within the existing
context: is the district layered with other important districts, is it a uniquely important contest, are there uniquely expensive-to-engage voters?
Second, uncertainty and variation in political data should be embraced with the knowledge that the numbers are estimates. Third, CPV should be used in the planning stages, ensuring that budgets are aiming to reach highest-return votes.
Finally, CPV should always be used to optimize the work, either maximizing the number of voters turned out or finding ways to save money.
Democrats are experiencing a wealth of resources, disproportionately powered by small donations on ActBlue. In this time of abundance, consultants and
organizers should avoid being penny-wise and pound-foolish by over-investing in low-cost, low-impact tactics. This is the time to engage in more testing and more building of long-term capacity that will continue to pay-off in leaner years. Now is the time to build.
Matt Singer is the founder of National Voter Registration Day, a veteran of numerous campaigns and civic engagement efforts, and a partner at Impactual LLC.
Matt Lackey is a political analyst with 15 years of experience helping progressives use data to be more effective in their campaigns.