Introduction to Running An Experiment

Got a research question in mind? Great! Start by entering your goal, whether it’s testing a new product, message, or campaign idea. Our platform will help you define the variables and audience characteristics to match your needs​​.

Step 1 - Why

Starting Your Experiment: Asking a Causal Question

To begin your experiment, start by asking a causal question about human behavior. Your question should aim to uncover the reasons behind actions, preferences, or attitudes. To create an effective question, first identify the behavior you’re interested in, then explore the factors influencing it.

For example:

  • You could ask, “How does fuel efficiency impact car buying?”
  • Or, take a broader approach with, “What causes car buying?”

Using the ‘Brainstorm Questions’ Feature

If you’re unsure how to phrase your causal question or need ideas, the Brainstorm Questions feature is a valuable tool for generating suggestions. Here’s how to use it:

  • Click on Brainstorm Questions without any input: This will provide example questions based on general human behavior, offering inspiration if you’re just starting out.
  • Enter a topic or idea before clicking Brainstorm Questions: If you already have a concept in mind, type it into the box (e.g., “cars” or “electric cars”). The tool will then generate relevant suggestions that you can refine and use as a starting point for your experiment.

Step 2 - When/Where

Define the Time and Place for Your Experiment

1

In this step, you’ll set the time and location for your audience. These options allow you to control the temporal and geographic context of your experiment.

  • When: Specify a particular year (e.g., 2024) or choose a broader timeframe like “the early 80s.” The system performs best when focusing on past or present time periods.

  • Where: Select the geographic location by choosing a specific country for your audience.

Step 3 - Who

Define Your Population for Conjoint Analysis

Conjoint analysis is a powerful survey-based technique used to understand how people make decisions and which attributes matter most when choosing between options. In this step, you will define the characteristics of the population for your conjoint analysis. The survey will be run using synthetic respondents, and the attributes you select will determine the participants in the experiment.

Choose the Characteristics

Select 2-10 traits that define your target population by clicking on the relevant cards. These traits can include attributes such as profession, income, age, and more. If you’re unsure which traits to select, you can choose a specialist (e.g., Economist, Political Scientist) to automatically configure the population based on their expertise.

Step 4 - What

Select Attributes for Your Experiment

In this step, you’ll define the key attributes for your conjoint analysis survey. These attributes represent crucial aspects of your study—whether it’s a product, service, or public policy. Each attribute will include multiple levels, representing variations that survey participants will evaluate.

Edit or Add New Attributes and Levels

To modify existing attributes or levels, simply click on the text to edit. To add a new attribute, click “Add New Attribute,” and corresponding levels will be generated automatically. Press Enter to save new levels. To exclude specific attributes from the experiment, uncheck the boxes next to them.