Help us unveil the secrets of Instagram’s algorithm

Install the add-on for Firefox or Chrome to help us monitor Instagram's newsfeed algorithm!

Instagram is the third largest social network in Europe, but compared to YouTube and Facebook, little was made to research its algorithm.

The project is now completed. Many thanks to all data donors for their help and their trust.

AlgorithmWatch wants to change that and calls on Firefox and Chrome users for their help in donating data. If you want to contribute, here’s what you can do:

Chrome users

  1. Install the Chrome extension from the Chrome Web Store.
  2. Follow the instructions. If you don’t have an Instagram account, you can create one for the experiment.
  3. Use your browser and your mobile normally. You do not need to be especially active or inactive on Instagram during the experiment.
  4. If you want to quit the experiment, you can remove the add-on at any time.

Firefox users

  1. Download the Firefox add-on here. Clicking on the link will prompt you to install the add-on. If nothing happens, right-click on the link and choose “Save link as…”.
  2. Install the add-on in your Firefox browser. You can drag-and-drop it in a new Firefox tab. Alternatively, you can go to your Add-ons by pressing Ctrl+Shift+A and select “Install add-on from file…” in the Settings menu.
  3. Follow the instructions. If you don’t have an Instagram account, you can create one for the experiment.
  4. Use your browser and your mobile normally. You do not need to be especially active or inactive on Instagram during the experiment.
  5. If you want to quit the experiment, you can remove the add-on at any time.

How it works

We follow a set of Instagram users who use the platform to earn an income. We want to know if Instagram’s algorithm favors some type of content over the rest.

The add-on asks data donors to follow three specific accounts. It then collects information about the pictures and videos that appear in their newsfeeds. It also collects information about some of the accounts that data donors follow. All the information is anonymized in a way that makes it impossible for us to re-identify data donors.

For any question about the add-on, contact Nicolas Kayser-Bril <>.

Frequently asked questions

What information does the add-on collect?

The add-on checks on your Instagram newsfeed at regular intervals. It opens it rapidly in a new window, so that you won’t see it happening. If your newsfeed displays a post from one of the accounts that you follow as part of the experiment, this information is passed on to the database.

The add-on also collects information about the accounts you follow. The names of the accounts are encrypted using a one-way hash function, so that we are not able to read them. (We collect this information to check if users who follow the same accounts are treated in a similar way by Instagram.)

My account is private. Can I participate?

Yes. Installing the add-on does not make your account public, even if you contribute data (and, because we don't know the names of people contributing, there's nothing we can find out on your account either).

I don’t use Instagram. Should I participate?

Yes. If you care to create an Instagram account for the sake of this experiment, your data will be very useful to understand what happens to a “clean” Instagram user.

Why does AlgorithmWatch does this?

AlgorithmWatch is a non-profit organization that sheds light on automated decision-making in Europe. Instagram’s is one of Europe’s largest social networks, where dozens of millions spend hours each day. Monitoring the algorithm of its newsfeed is part of AlgorithmWatch’s core missions.

What’s the point of the project?

One in 3 Europeans use Instagram. Among Europeans aged 18 to 14, Instagram use is close to 100%. What Instagram chooses to display or to hide shapes our perceptions and values. To understand Instagram’s power, we first need to have a better idea of its inner workings.

How did you pick the accounts to follow?

We currently run two projects. One focuses on small-scale entrepreneurs who use Instagram, especially in the fashion, sports, beauty and travel sector. We chose accounts that reflect the diversity of these profiles, by gender and geography. The other project, which is for Dutch-speaking users, focuses on Dutch politics and monitors the accounts of the members of the Dutch parliament.

Is it legal?

Yes. In the EU, article 3 of the 2019 directive on copyright allows for data-mining for research purposes. In the US, the 2020 Sandvig v. Barr case made clear that researching possible discrimination by platforms was not a violation of the Computer Fraud and Abuse Act.

How much time will it take?

For you, none. Once the add-on is installed, you don’t have to do anything. Your browser will open a new window, which you can minimize. If you want to quit the experiment, simply uninstall the add-on.

Should I install the add-on on my phone?

No. The add-on only works on computers. You can still use Instagram on your phone normally.

Is the add-on a security risk?

No. The add-on was developed entirely by AlgorithmWatch. We can give you access to the source code if you’re curious or want to audit it.

Can you post on my behalf?

No. The add-on simply checks the top items on your newsfeed at regular intervals when you use your browser.

Where is my data stored?

The data you contribute, which is fully pseudonymized, is stored in a database on Heroku, an online service. Only AlgorithmWatch personnel have access to the database. No one can know what data you contributed, because all usernames are pseudonymized using one-way hash functions.

Can I access the data I donated?

Yes. We don't know who our data donors are, but if you provide us with your Instagram username, we can look for the data associated to it. If you want to access your donated data, send an email to and a proof that you own the Instagram account in question (e.g. a direct message on Instagram).

Will my data be shared with third-parties?

No. Your data (which no one can identify as your data because it is pseudonymized using one-way hashes) will only be analyzed by AlgorithmWatch.

Sign up for our Community Newsletter

For more detailed information, please refer to our privacy policy.