Ranked-Choice Voting for Board Elections: A Practical Guide for Member Organizations

Most nonprofit board elections still use the simplest method there is: list the candidates, let members check a few boxes, and seat whoever gets the most votes. It is easy to run, but it can produce boards that represent only part of the membership. Ranked-choice voting offers a fairer alternative, and San Franciscans already know how to use it.

What ranked-choice voting is

On a ranked-choice ballot, voters mark a first choice, a second choice, a third choice, and so on. Ranking is optional beyond the first choice: a voter can rank every candidate or just one. The ballots are then counted in rounds. Whenever a voter's top-ranked candidate is out of the running, that ballot counts for the voter's next choice who is still in contention.

The core idea is simple. Your vote is never wasted on a candidate who cannot win, and you never have to worry that voting for your favorite will help the candidate you like least. That removes much of the strategic second-guessing that plurality elections force on voters.

"Ranked-choice voting" is an umbrella term for the ballot format. The counting rules differ depending on how many seats are being filled:

• Single-winner RCV, often called instant-runoff voting (IRV), is used to elect one person, such as a board chair or a mayor.
• Multi-winner RCV, known as the single transferable vote (STV), is used to fill several seats at once, such as a board of directors. It produces results that are roughly proportional to the support each group of voters has.

How instant-runoff counting works

Instant runoff is the easiest version to understand, and it is the foundation for the multi-seat method. The count follows a fixed loop:

1. Count first choices. Every ballot counts for its top-ranked candidate.
2. Check for a majority. If a candidate has more than half of the continuing ballots, that candidate wins.
3. Eliminate the last-place candidate. If no one has a majority, the candidate with the fewest votes is eliminated.
4. Transfer those ballots. Each ballot for the eliminated candidate moves to the next-ranked candidate on that ballot who is still in the race. If a ballot has no remaining choices, it becomes "exhausted" and stops counting.
5. Repeat until someone has a majority of the continuing ballots.

"Continuing ballots" means ballots that still rank at least one candidate in the race. Because exhausted ballots drop out, the majority threshold can fall slightly from round to round.

A worked example

Imagine a fictional coalition electing a board chair. One hundred members return ballots ranking four candidates. Here is how the count unfolds:

In Round 1, no one has a majority, so Candidate D, in last place with 12 votes, is eliminated. Of those 12 ballots, 3 list Candidate A next, 5 list Candidate B, 3 list Candidate C, and 1 ranks no one else and is exhausted. In Round 2, there is still no majority, so Candidate C is eliminated. Of C's 23 ballots, 4 move to A, 16 move to B, and 3 are exhausted. In Round 3, Candidate B has 51 of the 96 continuing ballots, a clear majority, and wins.

Notice what happened: Candidate A led on first choices, but most supporters of C and D preferred B to A. Under plurality voting, A would have won with only 38 percent. Instant runoff found the candidate with broader support. Every round adds up to 100 ballots, which is a good habit to check in any published count.

Filling several seats: the single transferable vote

A board election usually fills several seats at once. Running instant runoff repeatedly would just hand every seat to the majority's favorites. The single transferable vote solves this by aiming for proportional results: if a group makes up about a quarter of the voters, it should win about a quarter of the seats.

The Droop quota, in plain words

Instead of a majority, STV uses a quota: the number of votes a candidate needs to be guaranteed a seat. The most common is the Droop quota, which is the smallest whole number of votes that no more candidates than there are seats could all reach at once. The formula is:

Quota = (number of valid ballots ÷ (number of seats + 1)), rounded down, plus 1

For example, with 900 valid ballots and 8 seats, divide 900 by 9 to get 100, then add 1. The quota is 101. Nine candidates cannot all get 101 votes out of 900, so anyone who reaches 101 has earned a seat. For a single seat, the same formula gives a simple majority, which is why instant runoff is really just STV with one seat.

How the count proceeds

1. Count first choices and elect anyone who reaches the quota.
2. Transfer surpluses. If an elected candidate has more votes than the quota, the extra votes are passed on to their voters' next choices. Most modern rules do this fractionally: if a candidate has 151 votes against a quota of 101, the surplus is 50, so each of their ballots moves on at a value of 50 ÷ 151, about one-third of a vote. This treats every supporter equally rather than picking ballots at random.
3. Eliminate the lowest candidate if no one else reaches the quota, and transfer those ballots at their current value.
4. Repeat until all seats are filled, or until the number of remaining candidates equals the number of open seats.

Organizations adopting STV should write down the exact variant they use, including how surpluses are transferred and how ties are broken, so that anyone can reproduce the result.

Why "vote for 8 of N" elections fall short

In a typical plurality board election, each member may vote for up to as many candidates as there are open seats, and the top vote-getters win. This method, often called block voting or at-large plurality, has two well-known flaws.

A cohesive bloc can win every seat

Suppose 900 members vote for 8 seats, and a coordinated group of 480 members all vote for the same 8 candidates. Those 8 candidates each receive at least 480 votes. The other 420 members cannot beat that, no matter how they vote, because even if they all agree they top out at 420. A 53 percent majority wins 100 percent of the seats. Under STV with a quota of 101, the same electorate would elect roughly four candidates favored by each group, because 480 votes are worth about four and three-quarter quotas and 420 votes about four and one-sixth.

Vote-splitting and wasted votes

When several similar candidates run, their supporters' votes divide among them, and a less popular but more unified slate can win. Votes for candidates who finish far out of the running, and votes far in excess of what a winner needed, have no effect on the outcome. Members learn this quickly. They respond by voting strategically, discouraging like-minded people from running, or simply not voting.

San Francisco already knows how to rank

San Francisco has used ranked-choice voting for municipal elections since 2004. Residents rank candidates for offices like mayor and the Board of Supervisors, and the city publishes round-by-round results. That matters for a local membership organization: most of your members have already filled out a ranked ballot, understand the idea of backup choices, and have watched results update round by round. You can learn more about city elections at sf.gov.

The city's elections use the single-winner form. A multi-seat board election would use STV, which follows the same ranking logic with a quota instead of a majority. From the voter's side, the experience is nearly identical: rank as many candidates as you like, in order.

How a nonprofit can adopt RCV, step by step

Switching methods is a governance change, and it should be handled with the same care as any bylaw amendment. Our article on member voting rights explains why process matters as much as outcome.

1. Amend the bylaws

Draft an amendment that names the method (instant runoff for single offices, STV for multi-seat elections) and either spells out the counting rules or authorizes the board to adopt detailed written election rules that must be published before each election. Because a change to how directors are elected can affect members' voting rights, members should generally be given a clear, fair vote on it. Check your bylaws and consult a qualified attorney; this is not legal advice.

2. Write precise election rules

Specify the quota, the surplus transfer method, how exhausted ballots are handled, how ties are broken, and how many rankings a voter may mark. Precision prevents disputes after the fact.

3. Design a clear ballot

• Use a grid with candidates in rows and rankings in columns, or a drag-to-rank interface for online voting.
• State plainly that voters may rank as many or as few candidates as they wish.
• Explain that ranking a second choice never hurts the first choice.
• Rotate or randomize candidate order if your rules allow, to reduce the advantage of being listed first.

4. Choose trustworthy software

Several online voting services support ranked ballots and STV tabulation, and well-documented open-source tabulators exist that anyone can run. Whatever you choose, confirm that it implements the exact rules in your bylaws, that it can export an anonymized record of every ballot's rankings (often called a cast vote record), and that it protects ballot secrecy.

5. Make tabulation transparent

Assign oversight to an election committee or an independent third party rather than staff or sitting directors who are on the ballot. Publish the anonymized ballot data so any member can rerun the count and get the same result.

6. Publish round-by-round results

Report every round: each candidate's total, who was elected or eliminated, how many votes transferred, and how many ballots were exhausted. Totals that add up in every round are the best proof that the count was done right.

Common objections, and honest answers

"It's too complicated for voters." Ranking is intuitive, and San Francisco voters have done it for two decades. The counting is more involved than plurality, but voters do not need to understand the math to cast a ballot, just as they do not need to understand optical scanners.

"The count takes longer." With software, an STV count for a nonprofit board takes moments. The extra work is in publishing results clearly, which is worth doing anyway.

"Some ballots get exhausted." True. A voter who ranks only one or two candidates may see their ballot stop counting in later rounds. Good voter education and generous ranking limits reduce this. Plurality elections waste far more votes.

"RCV has mathematical quirks." Every voting method has edge cases; voting theorists have shown that no ranked method can satisfy every desirable fairness property at once. RCV's quirks are rare in practice, and its benefits in reducing vote-splitting and improving representation are substantial.

"Our current board works fine." Perhaps. But a fair method protects the organization against future factions too, including ones current leaders would not welcome. Good rules are insurance. You can see how RCV fits our broader governance goals on our platform page and in our FAQ.

Frequently asked questions