Splitting Survey Respondents Across Variations

The scenario

You're running a study that requires multiple survey versions — different question orders, different wording, a treatment group and a control group. You need respondents to be randomly assigned to one version, but you only want to share a single link. You don't want to manually divide your participant list or manage separate URLs for each group.

How to set it up

1. Create your survey variations

Build your survey versions in whatever tool you use — Google Forms, Qualtrics, Typeform, SurveyMonkey, or any other platform. Each version gets its own URL.

2. Create a Redirect Randomly link

Go to the Redirect Randomly tool and enter each survey URL as a destination:

• https://forms.google.com/d/e/survey-v1 (Treatment A)
• https://forms.google.com/d/e/survey-v2 (Treatment B)
• https://forms.google.com/d/e/survey-v3 (Control)

Each respondent who clicks the link is randomly assigned to one of the variations.

3. Share one link

Distribute a single link — nimble.li/survey or whatever slug you choose — to your participants via email, course announcements, social media, or recruitment platforms like Prolific and MTurk. Every participant uses the same link, but each one lands on a randomly selected survey version.

What you get

• Random assignment — Respondents are assigned to conditions randomly, which is the foundation of experimental design. No manual group splitting required.
• One link for all conditions — You don't need to manage separate links or distribute different URLs to different participants. This eliminates the risk of accidentally sending the wrong link to the wrong group.
• Condition tracking — The analytics dashboard shows how many respondents were sent to each variation via the Destinations breakdown.
• Platform-agnostic — Works with any survey tool that has a shareable URL. You can even mix platforms — one variation on Google Forms, another on Qualtrics.

Variations

Even distribution with Redirect Evenly

If you want guaranteed equal distribution rather than random assignment, use Redirect Evenly instead. It cycles through destinations in order — the first click goes to Survey v1, the second to v2, the third to v3, then back to v1. Over time, each variation receives exactly the same number of respondents.

Weighted assignment

If your study design requires unequal group sizes (e.g., 70% treatment, 30% control), use Redirect by Weight instead of Redirect Randomly. You set the exact percentage each variation receives.

Passing a condition identifier

For tools that support capturing URL parameters (like Qualtrics with embedded data fields), enable parameter forwarding and append a condition tag:

nimble.li/survey?study=fall2026

The parameter passes through to the survey URL and gets recorded alongside each response, letting you tag cohorts or recruitment sources without asking participants to self-report.

More than three conditions

You're not limited to three variations. Add as many survey URLs as your study design requires — Redirect Randomly distributes across all of them equally.