Suppose a leader wants a whole organisation to agree on its priorities. There are twenty goals on the table. How do you find out which matter most?
The usual answers both disappoint.
Ask everyone to rank all twenty. Ranking twenty things in order is hard. People get tired, fall back on the first few items, and the middle of the list becomes noise.
Run a survey with a score per goal. Almost everything scores "important". The result is a list where everything is close together and nothing is decided.
Two at a time
There is a simpler question: which of these two is better? Almost anyone can answer that, and the answer is honest because there is no way to give every item top marks.
In vision lab, you see two ideas at a time. The person picks one. Then they see another pair. Each pick nudges the winner up and the other down. After many people have voted, the best ideas sit at the top.
The method behind it comes from chess. The Elo system rates players by who beats whom, and it works with any number of players and any number of games. We use the same idea for ideas.
Why it holds up
- Easy for everyone. One choice at a time, so people keep going.
- Fair to late ideas. A new idea is tested the same way as an old one, so one posted late can still reach the top.
- Hard to push. Authors never see their own ideas in a vote, and vote caps and pacing limit anyone trying to steer the result.
- Clear confidence. Each idea shows how often it was chosen and how many comparisons that rests on, so you can see how solid a result is.
Where it fits
Pairs work best when you have many ideas and many people. For a handful of options and a small meeting, a show of hands is fine.
What a person sees
A voter sees two ideas, taps the one they prefer and moves on. They are not asked to score, rank or explain. If they want to add something, they can add an idea or a comment, but it is optional. Most people vote for a few minutes and leave, and that is enough, because many people's small contributions add up.
Reading the result
At the end you do not get a score out of ten. You get a ranked list, and next to each idea, how often it was picked and how many comparisons that rests on. Use the list to decide what to look at first, and use the counts to decide how much weight to give a close call. If two ideas are neck and neck, the comparison counts tell you whether to treat them as tied.
When not to use it
Pairs are not the right tool for everything. A choice between two options that a small group can discuss in the room does not need a vote. And a vote shows what people back. It does not make the decision. Someone accountable still does that.