Frequently Asked Questions¶
Can't find what you're looking for? Browse these common questions and answers.
General¶
What is FlagPal?¶
FlagPal is a feature management and experimentation platform. It lets you control which features your users see, roll them out gradually, run A/B tests, and measure results — all without deploying new code.
Do I need to be a developer to use FlagPal?¶
No! The FlagPal dashboard is designed for the whole team — product managers, designers, marketers, and analysts. You don't need to write code to create feature flags, set up experiences, run experiments, or view metrics.
FlagPal integrates with your existing application within a few minutes through our SDKs (plugins). If you're a developer, you can also use the API to integrate with your application manually. You may refer to our API reference for more details.
What's the difference between an Experience and an Experiment?¶
| Experience | Experiment | |
|---|---|---|
| Purpose | Roll out a feature | A/B test two+ variants |
| Variants | Single | Multiple |
| Needs metrics? | No | Yes |
Use an Experience when you want to turn a feature on/off for specific users. Use an Experiment when you want to test which version of something performs better.
What types of Feature Flags are there?¶
FlagPal supports five types:
- Boolean — true/false (on/off)
- String — text values
- Integer — whole numbers
- Array — lists of values
- Date — date values
Learn more about Feature Flag types →
Can I change a feature flag's name after creating it?¶
No — feature flag names are permanent. This is because your developers use the name in their code. If you could change the name, it would break the integration.
Choose descriptive names carefully when creating flags.
Instead of changing the name, consider creating a new flag with the desired name and migrating your users to the new flag.
Experiments¶
How long should I run an experiment?¶
At minimum, 1-2 weeks. Ideally longer, until each variant has had at least several hundred (preferably thousands) of distinct users.
Running an experiment for too short a time leads to unreliable results — you might see a strong early signal that disappears with more data.
The dashboard shows if your results are statistically significant for every experiment, so you can make an informed decision.
How do I know when an experiment has enough data?¶
There's no single rule, but a common guideline is:
- Each variant should have at least 100 users (more is better)
- The experiment should have run for at least one full week (to account for day-of-week effects)
- The conversion rate should have stabilised — it's not jumping around wildly anymore
The dashboard shows if your results are statistically significant for every experiment, so you can make an informed decision.
Can I run multiple experiments at the same time?¶
Yes, but be careful. If two experiments involve the same feature or the same users, their results can interfere with each other. It's best to ensure experiments target non-overlapping user segments when possible.
What does "weight" mean in an experiment variant?¶
Weight controls how traffic is split between variants. Equal weights = equal split. For example:
- Variant A weight 1, Variant B weight 1 → 50/50 split
- Variant A weight 9, Variant B weight 1 → 90/10 split
Use unequal weights when you want to cautiously test a new variant on a small percentage of users first.
Can I add metrics to an experiment after it's already running?¶
You can add metrics to a running experiment. However, metric data is only collected from when the metric is attached — it won't retroactively count events that happened before the metric was added.
Metrics¶
What do the metric types mean?¶
When creating a metric, you can choose from the following types:
- Boolean — did something happen? (e.g., purchase completed, signed up).
- Count — how many times? (e.g., clicks, page views).
- Money — how much revenue?
These types are designed to capture the most common use cases. They collect values in the same way and are only used to properly render the metric when analysing it in the dashboard.
How does metric data get into FlagPal?¶
Our SDK (or your development team) sends metric events to the FlagPal API whenever something measurable happens in your app. For example, when a user completes a purchase, your app sends a metric event to FlagPal.
Every recorded metric event is associated with a specific variant of an experiment. This way you can see the exact number of metrics for each variant of an experiment.
My metric shows 0 conversions. What's wrong?¶
Common reasons:
- The metric hasn't been attached to any active experiment
- Your developers haven't implemented the metric event yet
- The experiment is new and hasn't had enough traffic
Check with your developer that metric events are being sent correctly.
Projects & Teams¶
Can I have multiple projects?¶
Yes! Your subscription allows you to create and manage multiple projects. Each project is an isolated workspace.
How do I invite someone to my project?¶
Go to Project Settings → Invitations → Invite a User. Enter their email and they'll receive an invitation link.
What happens when I deactivate an Experience or Experiment?¶
New users will no longer be served that Experience's feature values and fall back to the defaults (or another matching Experience/Experiment). This happens in real time with no deployment required.
If your Experiment was successful, you should create a new Exprience that activates the new variant with your desired targeting rules.
I accidentally deactivated an Experience and users are seeing the wrong thing. What do I do?¶
Simply reactivate the Experience. Navigate to Experiences, find the one you deactivated, and toggle it back on. The change takes effect immediately.
Can I delete a feature flag?¶
Feature flags can be deleted from the dashboard. However, be careful — if your developers are still using the flag in your application code, deleting it will cause your app to receive a null or default value for that flag, which may cause unexpected behaviour.
Always coordinate with your development team before deleting flags.
Technical¶
What is an Actor?¶
An Actor represents an entity in your application that receives feature flag values. In most cases, an Actor is a person — but it can also be an organisation, an AI agent, an API client, or any other entity that your app uses. It is an optional feature of FlagPal: If you're not storing your actor Feature Flags on your platform (locally), you can use our Actor feature to store them remotely on our servers. Imagine this a remote database that saves your provided list of features together with your user's identifier.
What's an API Token?¶
An API Token is a secret key that your application uses to communicate with FlagPal. It identifies which project the requests belong to.
Still Have Questions?¶
If you didn't find the answer here, reach out to our support team at flagpal.com.