What is MaxDiff?
MaxDiff, short for Maximum Difference Scaling, is a robust research methodology used to determine people's preferences for different options (attributes, features, benefits, etc.). Instead of asking respondents to rate each item individually, MaxDiff presents sets of options and asks them to choose the best (most preferred) and worst (least preferred) from each set.
This approach provides a more accurate understanding of true preferences rather than traditional rating scales, as it forces respondents to trade off individual options against each other.
Why use MaxDiff?
MaxDiff is particularly useful when you need to:
Prioritise a list of features or benefits: Understand which aspects of your product or service resonate most with your target audience and understand the perceived.
Evaluate different product concepts: Determine which product variations are most appealing.
Assess brand attributes: Discover which brand attributes your audience prioritises in your category.
Identify the most impactful messaging: Find out which claims or slogans are most persuasive.
Setting up your MaxDiff survey
1. Designing your survey
MaxDiff questions can be embedded into any survey, by simply clicking “Add Question” > “Max Diff “ in the editor.
This gives you the flexibility to include other question types before and after the MaxDiff question.
You can add 1 Max Diff Question per audience in your survey.
2. Options
These are the individual options you want respondents to evaluate. They could be:
Product features (e.g., "Long battery life," "Memory" etc.)
Brand attributes (e.g., "Innovative," "Trustworthy")
Marketing messages (e.g., "Save money instantly," "Experience ultimate comfort")
We recommend 8 - 15 options for optimal results; but you can have up to 25 options.
We currently don’t support the use of images in your options.
3. Question sets
This is how many times a respondent will be asked to choose "best" and "worst."
The platform will automatically generate a sufficient number of sets to ensure each item is shown roughly 3 times.
The respondent will always see 4 items per Question set.
(1 Question set is equivalent to 1 Credit, so 1 MaxDiff question, with 8 question sets will total 8 credits)
4. Preview
You can preview your survey to understand how respondents will go take the survey.
Select “Preview” then “Test Survey”
Respondents will see a short introduction explaining the task before they go on to select their best and worst options.
Understanding your MaxDiff results
Once your survey is complete, you can see the results of your survey on the Results dashboard and dive deeper into the analysis.
We will show three key metrics for each option:
% Best - The percentage of respondents who selected the item as most important at least once
% Worst – The percentage who selected it as least important at least once.
MaxDiff Score – A summary score that reflects the relative preference for each item.
You can view these as charts (where you can toggle between showing MaxDiff score, % Best and % worst) or in crosstab view.
How is the MaxDiff score calculated?
We use a respondent-based approach to calculate the MaxDiff score. This means each respondent contributes equally to the final score, regardless of how often items appeared.
For each respondent, we calculate a score for each option:
We count how many times they picked the item as best and how many times as worst
Then we subtract worst from best
Finally, we divide by the number of times they saw the item — this step normalises their score
Once we have each respondent’s score for an option, we average all those scores to get the final MaxDiff score for that option.
This approach helps correct for any differences in item exposure due to random set generation.
Filter and compare your results
You can apply filters to see the MaxDiff results for different groups of respondents, or apply splits if you want to compare certain segments against each other. You can view this in a chart format (where again you can choose between MaxDiff score, % best, %worst) or in crosstab format where you can see all metrics in one place. We also show which differences are statistically significant, but only for % Best and % Worst, not for the MaxDiff Score, as it is a normalised average that isn’t compatible with standard significance tests like Fisher’s Exact Test.
Save to a board
You can easily save your MaxDiff charts to a board to start building a story that you can share with stakeholders. You can add your commentary here or get help from AI to summarise the key findings. Learn more about boards here.
Export your data
If you want to export your data, you can do this by clicking on the three dots next to the chart or crosstab and choosing “Copy data”. This will copy the data to your clipboard, allowing you to paste it in e.g. a spreadsheet. We do not support MaxDiff question export in the existing Excel or PPT exports.