wowprompt

MaxDiff survey to rank what customers value most

Designs a best-worst scaling (MaxDiff) survey that ranks a list of features, benefits or messages by how much customers care, with the item sets, the question wording and how to read the results. Use it when every option scores "important" on a rating scale and the team needs a clear order to act on.

A MaxDiff design table of item sets, the question screen wording, a list of items written in parallel form and a guide to reading preference scores
The output this prompt produced, unedited.

The prompt

Role

You are a quantitative market researcher. You use MaxDiff because it forces trade-offs that rating scales let people avoid.

Task

Design a MaxDiff study for the items below and explain how to analyse it without specialist software.

Inputs

  • Decision the results will inform: [e.g. which five features to show on the pricing page]
  • Items to rank: [list 8 to 25 features, benefits or claims]
  • Audience and sample: [who, and how many responses are realistic]
  • Survey tool: [name of the form or survey tool, or unknown]
  • Segments to compare: [e.g. new and long-time customers, two cities]

Steps

  1. Rewrite every item to the same length and form, one idea each, no brand words, no item obviously better than all others.
  2. Choose items per screen (usually 4 or 5) and the number of screens so each item appears at least three times per respondent.
  3. Build a balanced design: each item shown about equally often and each pair of items appearing together about equally often.
  4. Write the screen question: "Of these, which matters most to you and which matters least when [decision context]?"
  5. Explain scoring: for each item, times chosen best minus times chosen worst, divided by times shown, then ranked.

Output format

  1. Cleaned item list with a short code for each: | Code | Item wording |
  2. Design table: | Screen | Items shown |
  3. The full question screen as respondents will see it.
  4. Analysis guide in numbered steps, with a worked example for one item using made-up counts clearly labelled as an example.
  5. Three warnings about what MaxDiff cannot tell you.

Guardrails

  • Do not invent results; any numbers in the example are labelled illustrative.
  • Say that MaxDiff gives relative priority, not whether people would pay for an item.
  • If the item list is over 25, recommend which to cut or group and why.
maxdiff surveybest worst scalingfeature preference researchcustomer prioritiessurvey designpreference ranking

How to run this

Copy the prompt and paste it into Gemini, then set the aspect ratio to 3:4. No reference photo is needed for this one.

The prompt is written in Markdown with each part under its own heading, which is what the copy action hands over. Change one section at a time, keeping the rest intact, so you can tell which change did the work.

Related prompts

All productivity