Question type

Constant sum scale

A constant sum scale asks respondents to divide a fixed total, usually 100 points, between several options according to how important each one is. Because the total cannot grow, giving more to one option means taking from another. The result shows both the order of priorities and the distance between them.

Available on the free plan

Build a survey with this question Survey templates

Constant sum question examples

You have 100 points. Split them by how much each factor affects your choice of supplier.
If your training budget were 100%, how would you split it between these formats?
Distribute 10 votes between the proposed projects for next quarter.
Out of 100 minutes of a typical support call, how much goes to each of these steps?

Why a fixed total changes the answers

On a rating scale, a respondent can call price, quality and speed all “very important”. With 100 points, they cannot. They have to decide that quality deserves 50 and speed 15, and those numbers carry information a scale never captures: quality is not just ahead, it is more than three times as important.

How to analyse constant sum data

share of option = points given to it ÷ total × 100%
average share = mean of that share across respondents

The average shares of all options add up to 100%, which makes a stacked bar or a simple table easy to read. Look at the spread too: an option with an average of 20 points can mean everyone gave it 20, or half gave 40 and half gave nothing. Those are two different markets.

When to use it, and when not

Use a constant sum when weight matters: budget allocation, criteria for a decision, share of time. Keep the list to three to seven options; with twelve, people give up on arithmetic and spread points evenly.

Do not use it when a simple order is enough (ranking is less effort), or on audiences who find numbers stressful. And remember that the total is relative: an option that got zero points may still matter, just less than everything else on the list.

Mistakes with a constant sum scale

Unclear total: say it in the question and in the block (“100 points”).
Options that overlap, so points are split between two names for the same thing.
Too many options for the total: 10 options with 10 points forces zeros everywhere.
Asking on a phone without telling people they can use round numbers; nobody needs to type 17.

A worked example

Three customers split 100 points between price, quality and delivery speed: (30, 50, 20), (20, 60, 20) and (40, 40, 20). The average shares are 30% for price, 50% for quality and 20% for speed. Quality leads, and it leads by a lot: it weighs as much as price and speed combined. A ranking question would only have told you that quality came first.

Building it in Askyo

The Constant sum block starts with a total of 100 and the unit “points”; both can be changed, so you can ask for 10 votes or 100 percent. It needs at least two options. Each option gets a number field, and a footer shows the running total with “N left”, “N over” or “done”.

The answer counts only when the numbers add up exactly to the total; if a respondent goes over, they see a message saying the sum exceeds it. Analytics shows the average points per option and its share of the budget. The CSV lists “Option: N points” and skips zeros.

What you can set

the list of options
the total being shared out
unit of measure

Templates that already use it

«Distribute 100 points across the selection criteria» Customer expectations Research
«Distribute 100 importance points across the criteria» In-depth market research Research

Frequently asked questions

Constant sum scale: common questions

What is a constant sum scale?

It is a question where respondents split a fixed number of points, usually 100, between options by importance. The fixed total forces trade-offs, so the answers show how much more one option matters than another, not only which comes first.

Constant sum or ranking?

Ranking gives the order; constant sum gives the order and the distance between items. Choose ranking for quick prioritisation with little effort, and constant sum when you need weights, for example to allocate a budget or weight decision criteria.

What if a respondent’s points do not add up to the total?

In Askyo the block shows how many points are left or over while the respondent types, and the answer is accepted only when the sum matches the total exactly. There is no partial answer to clean up later.