Platform · Methodology

The science of
changing minds.

Grow Progress uses Randomized Controlled Trials, the same gold-standard methodology as clinical drug trials, to measure what actually persuades people. Here's exactly how it works.

Gold standard
RCT methodology
app.growprogress.ai
Study Overview
Complete
1,400
Respondents
24h
Field period
95%
Confidence
RCT
Methodology
Random assignment
👥
Survey
panel
Treatment A
Treatment B
Control
+18 pts
+9 pts
,
Causal lift, not correlation
Statistically significant
+18 pts
causal lift vs. control group
Why RCTs

Most research measures what people think.
We measure what changes it.

A poll tells you 58% of voters support a policy, but it can't tell you which message drove them there, or whether a different message would have moved them more. That's the difference between correlation and causation. Traditional research describes what people already believe. It can't isolate what shifted those beliefs, because nothing was ever randomly assigned.

An RCT randomly assigns respondents to see a message (treatment group) or nothing (control group). The difference between the two groups is the causal effect of that message, isolated from everything else.

This is the same design used in clinical drug trials. It's the gold standard for proving causation in medicine, and it's the gold standard for proving causation in persuasion. When a treatment group moves 9 points more than a control group, you know the message did that, not ambient news, not the question order, not demographic drift.

The Process

From study design to
actionable insight.

Five steps. Twenty-four hours. Results you can act on.

01
Step One

Design the study

Define a success question, your messages (test a single message or compare several head-to-head), target audience, and required sample size. Grow Progress strategists work with you to frame the study so results are clean and actionable, not just statistically valid but practically useful.

Success questionMessage variantsAudience specSample sizing
02
Step Two

Build the instrument

GP builds a structured survey with screener questions to qualify respondents, randomized message exposure across treatment and control conditions, pre/post attitude measures calibrated to your success question, and a full demographic battery.

Screener questionsRandomized exposurePre/post measuresDemo battery
03
Step Three

Field the study

The survey fields to a targeted sample matched to your audience, sourced through trusted, vetted third-party providers. Typical studies run 1,000–2,000 respondents. Rigorous quality controls remove speeders, straight-liners, and inattentive respondents before analysis begins.

1,000–2,000 respondentsAudience-matched sampleQuality controls
04
Step Four

Analyze the results

Statistical analysis isolates lift: treatment group minus control group. Every study breaks results out by subgroup, party, gender, age, geography, education, and more, so you know not just what worked overall, but which audiences it moved, and which it didn't.

Lift calculationSubgroup breakdownsConfidence intervalsSignificance flags
05
Step Five

Deliver insights

You receive a clear results report, not a raw data dump. Lift scores, confidence intervals, subgroup breakdowns, and a written strategist recommendation explaining what the findings mean and how to act on them. Results are typically back within 24 hours of fielding.

Lift scoresStrategist recommendationSubgroup report24h delivery

A complete picture,
not a point-in-time snapshot.

The Grow Progress platform isn't a single tool, it's a connected intelligence cycle. Each product feeds the next, so your audience insight shapes your messages, your message tests validate what works, and your tracking catches when conditions shift.

1
Understand

Know your audience before you write a word

Map who your audience is, what they believe, and what frames resonate, before you ever test a message.

Audience Understanding →
2
Test

Prove causal lift, not just preference

Run Randomized Controlled Trials on your messages. Know exactly which frame moves people, and by how much.

Rapid Message Testing →
3
Benchmark

Build on what's already been proven

Search 500+ studies to find what has worked for similar audiences and issues, so you're not starting from zero.

Persuasion Library →
4
Track

Know when opinion shifts, and react fast

Monitor how your audience's views evolve over time, and re-enter the cycle the moment conditions change.

Pulse Surveys →

Results that are causal, not correlational. Unlike traditional surveys or A/B tests that measure preference, Grow Progress's RCT methodology produces evidence that your message caused a shift in support, the kind of result you can defend in front of a board, a funder, or a campaign manager under pressure.

1,000–2,000
Respondents per study
24h
Field-to-results turnaround
95%
Confidence threshold
500+
Studies run
What Makes It Different

Why RCTs beat every other
approach to message research.

Causal isolation

Unlike A/B testing on live campaigns or focus groups with self-selected participants, RCTs give you causal proof of what moved the needle. Because respondents are randomly assigned, the difference between groups can only be the message, not the news cycle, question order, or who happened to land in each group. You don't just know that a message performed, you know it caused the result.

Speed

Traditional research takes weeks and costs $50–100k per study. Grow Progress takes 24 hours at a fraction of the price. That's the difference between data you can act on today and data that arrives after the decision has already been made.

Subgroup precision

Every study breaks out results by key demographics, party, gender, age, geography, education, and more. You know not just what works overall, but who it works on. The message that moves suburban women may do nothing for rural independents.

FAQ

Common questions about
the methodology.

What's the minimum sample size?
Most studies use 1,000–2,000 respondents, split between treatment and control. Sample size is set to achieve 80%+ statistical power for detecting meaningful effect sizes. Grow Progress strategists help you size the study correctly based on your expected lift and number of variants.
How many message variants can I test?
From a single message to several at once. You can test one message on its own or compare a slate of them head-to-head in the same study. Each additional message requires proportionally more respondents to maintain statistical power, so Grow Progress will recommend the right number based on your questions, audience, and budget.
How do you recruit respondents?
GP sources respondents through trusted, vetted third-party sample providers, matched to your target audience, national, district-level, or custom. Every sample runs through rigorous quality controls before analysis. For political studies, samples can be matched to registered voter files for precise audience targeting.
What does the results report include?
Lift scores (treatment minus control), confidence intervals, statistical significance flags, subgroup breakdowns by party, gender, age, geography, and education, plus a written strategist recommendation. You get a clear narrative on what the findings mean for your next decision, not a raw data dump, including which messages to lead with and which to drop. A message that produced no effect, or backfired, is as important to know as one that worked.
Can I test video ads?
Yes. Grow Progress supports text, image, video, and audio message formats. If your organization uses it to persuade people, it can be tested in a Grow Progress study.
See It In Action

See the methodology work
on a real question.

Talk to our team. We'll walk you through a real study example and show exactly how an RCT would answer your most pressing message question.