Ranking questions let participants drag a list of options into order, from most to least important. Where multiple choice tells you what was picked, ranking tells you what people prioritise when they have to choose between things.
That makes them well suited to preference research, where the gap between first and second place matters more than raw popularity.
When to use ranking questions
Prioritising features or ideas when you cannot build everything.
Comparing messages, concepts or designs head to head.
Understanding the trade-offs participants are willing to make.
Testing whether a favourite is a genuine favourite or just the least disliked.
Adding a ranking question
Open your study and go to the Build tab.
Click Add step.
Select Ranking, under Survey.
Type your question and add a description if needed.
Click on the choices and enter your text.
Hit Enter or click Add choice to add more options.
You can reorder choices by clicking and dragging. Every ranking step needs at least two options.
Additional options
Click Options on the step to access these settings:
Required step: Participants must rank at least one option before continuing.
Randomize options: Display the options in a random order each time the study is taken, which removes order bias.
Must rank all options: Participants must rank every option before they can continue. Leave this off to allow partial rankings.
How participants rank options
Participants see two panels: Unranked, holding all your options, and Ranked, which starts empty.
They drag options from Unranked into Ranked. The first one dropped in becomes 1, and each option after it takes the next position. They can drag options up and down within Ranked to change the order, or drag one back to Unranked to remove it.
Ranking is done by dragging, with a mouse or by touch. On mobile the two panels stack vertically.
Understanding your results
The Summary shows two tabs. Rank lists your options by average rank, best first, with the number of participants who ranked each one. Strength shows the percentage who put each option first, and the percentage who put it in their top three.
The two views can disagree, and that is usually the interesting part: a middling average with a high share of first places means opinion is split rather than lukewarm.
In the responses table, each participant's ranking appears in order, such as Cheese > Peas > Bread. Individual responses show a numbered list, and CSV exports separate the options with a pipe. Options that nobody ranked do not appear in the Summary.
Tips for writing good ranking questions
Aim for 4 to 7 options. The bottom of a long ranking is usually guesswork.
Keep options comparable, so participants are ranking like against like.
Say in your question wording if participants need to rank everything, since a locked continue button does not explain itself.
Next steps
If you only need to know what participants picked rather than the order they would put things in, use a multiple choice question. To score each option on its own instead of against the others, use a rating scale.
You can also use conditional logic with ranking steps to branch on whether a particular option was ranked, though not on the position it was given.