diff --git a/CHANGELOG.md b/CHANGELOG.md
index c7eda1f0..bb1a5cde 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -6,7 +6,9 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
+
 - Deactivate Binder ([#340](https://github.com/Substra/substra-documentation/pull/340))
+- Added section about channels in the main concepts ([#344](https://github.com/Substra/substra-documentation/pull/344))
 
 ## [0.30.0]
 
diff --git a/docs/source/documentation/concepts.rst b/docs/source/documentation/concepts.rst
index 5e6b1a6a..099f1b2a 100644
--- a/docs/source/documentation/concepts.rst
+++ b/docs/source/documentation/concepts.rst
@@ -160,6 +160,32 @@ There are two ways to run several tasks in parallel on a same organization with
 .. TODO:: Explain what is substra tools
 
 
+Substra channels
+----------------
+
+If your Substra deployment has only one channel then you most likely never knew what a channel really is. A channel is a group of Substra `organizations <https://docs.substra.org/en/stable/additional/glossary.html#term-Organization>`_ which operates on a common set of assets.
+
+Having multiple channels is needed if you want to have multiple projects that are independent from each other. If you have a Substra network and need to start another independent FL project then you will have to create another channel.
+This is why many networks are “mono channel”, since they only have one FL project going on.
+
+In the example below you a Substra network with two channels:
+
+- Channel A with the three organizations
+- Channel B with only two organizations (A and B).
+
+.. image:: /documentation/images/channel_schema.png
+
+Users only belong to one channel
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+A user can belong to only one channel. A channel has its own users, its own datasets, compute plans, tasks and functions. A user from another channel has no visibility on this. This makes sense as someone from another project should not see any info on something he/she is not a part of.
+
+What happens if an organization gets multiple tasks from different channels?
+^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
+
+This is a FIFO (first in first out), if multiple tasks get sent to an org it will process those in the order it got them. 
+What is important to note is that if someone from channel A stacked up loads of tasks before you (a user from channel B) tasks will all be in waiting until the previous ones from channel A are done. This might be confusing as you will not see an obvious reason as to why your tasks are not being processed.
+
 Compute plan execution - deployed mode
 ---------------------------------------
 
diff --git a/docs/source/documentation/images/channel_schema.png b/docs/source/documentation/images/channel_schema.png
new file mode 100644
index 00000000..d77f81e7
Binary files /dev/null and b/docs/source/documentation/images/channel_schema.png differ