SDKs
Client-side
Low-code

Running low-code experiments

FeaturesFlow allows you to run experiment in low-code environments such as Shopify or Webflow. Our platform allows you to achieve that by running split URL tests.

Split URL tests is a type of A/B test where users are sent to entirely different URLs rather than variations on the same page. It’s typically used to test major design or layout changes that require separate pages but can easily be used to compare variants with minor differences.

How it works

  1. Traffic Allocation – Users are randomly assigned to different versions of the test, each hosted on a different URL.
  2. Redirects – The test redirects a portion of visitors from the original page (example.com/old) to a new page (example.com/new).
  3. Tracking & Analysis – FeaturesFlow will forward events pushed to dataLayer so that you are able to analyze your experiment within our platform.

Getting started

  1. To get started, paste the following script tag into the <head> element of your website. We recommend pasting it in all pages of your website. In Shopify it can be achieved by pasting it in theme.liquid file.
<script src="https://unpkg.com/@featuresflow/sdk-low-code@1.0.0-beta.6/dist/index.umd.min.js" id='features-flow' authentication-key='YOUR-AUTH-KEY'></script>
  1. Go to app.featuresflow.com (opens in a new tab), navigate to experiment in your project and click the create experiment button.
  2. Give your experiment a name and hypothesis, then select URL Split as the way of splitting traffic and select your experiment runtime.
  3. You will then be asked to define your variants. Each variant has a name, URL and traffic. Make sure to input those fields. The baseline field corresponds to the default version of your website that you want to change.
  4. Now your experiment will start running from the start date.

Note

If that’s the first experiment that you are running an experiment, there will be no events in FeaturesFlow yet. We recommend you come back the next day once the data is already in place. Then you will be able to create metrics, you can read more about metrics here. If you have already run experiments with FeaturesFlow, you should already have your metrics in place. You can read more about creating metrics here.

  1. Now we can go ahead and attach relevant metrics to our experiment.
  2. The only thing left to do now is to wait for the experiment to finish running and conclude the results. FeaturesFlow will automatically stop the experiment once the runtime is over.