What is G & R Block?

What is Group and Randomization?

Group and Randomization is a survey design technique used to control how different participants experience a study. Instead of showing every participant the same set of questions in the same order, it lets you divide participants into groups and assign different content to each group, randomise the order in which questions or groups appear, or control exactly how many participants are exposed to each version of the study.

It is used when you need to test multiple versions of a concept, message or creative, compare audience segments or eliminate order bias from your data.

CORE CONCEPTS

Grouping

Grouping means dividing participants into separate paths within the same study. Each group sees a different set of questions or content. Participants are assigned to a group either randomly or based on quota controls. This is useful when you want to test different concepts, messages or designs without running separate studies.

Randomization

Randomization shuffles the order in which groups or question blocks appear for each participant. This removes order bias, which occurs when the sequence in which questions are presented influences how participants respond. For example, if you are testing three product concepts, showing Concept A first every time may give it an unfair advantage because participants are freshest at the start of a study.

Quota Control

Quota control lets you define exactly what percentage of participants should be assigned to each group or block. For example, if you want 50% of participants to see Concept A and 50% to see Concept B, you can set a quota of 50% per group. The study stops assigning participants to a group once its quota is filled.

Note

You can use either Randomization or Quota control on a group, but not both at the same time.

When to Use Group and Randomization

• You are testing multiple versions of a product concept, message or creative and want each participant to see only one version
• You need to compare how different audience segments respond to the same content
• You want to eliminate order bias by randomising the sequence of questions or groups
• You need to control the number of respondents exposed to each version of the study
• You are running an A/B or multivariate test within a single survey

COMMON USE CASES

Concept testing

A brand has three new product concepts and wants to test each one with a separate group of participants. Using Group and Randomization, each participant sees only one concept, preventing exposure to other versions from influencing their response.

Ad and message testing

A marketing team wants to test two versions of an ad headline with equal numbers of participants. Quota control ensures exactly 50% of respondents see each version, giving balanced and comparable data.

Removing order bias

A study asks participants to evaluate five product features. If all participants see the features in the same order, the first and last items tend to receive more attention. Randomizing the block order ensures every feature gets equal exposure across the full sample.

Segmented research

A retailer wants to show different sets of questions to frequent buyers and occasional buyers within the same study. Groups are used to assign each segment a tailored question set without running two separate studies.

Multivariate testing

A product team wants to test three price points and two packaging designs simultaneously. Group and Randomization allows different combinations to be shown to different participant groups, generating data on each combination from a single study.