Your Facebook Messages Are “Read”!

Yesterday, I came to know something new about Facebook Messenger and Facebook’s data collection capabilities, and I am sure many people (except developers apparently) do not know this:

Facebook reads/scans all your private messages in your Facebook Messenger app (or Facebook Inbox), and guess what it does with it? Yes, you got it right: advertising relevant products.

Here is how it happens: You or your friend need a dentist, so you have a private talk about “dentist,” “implant,” “tooth,” etc, and just after a couple of hours, yes, again, you see an ad by a dentist :) That’s what happened yesterday that I started to look for more info about it.

Maybe earlier you searched for a laptop to buy, and later saw laptop ads everywhere, which is something we were very familiar with already. But understanding our intent from the very first moment is something dazzling.

Facebook does this with its artificial intelligence (AI)product called DeepText, which analyzes texts in more than 20 languages. Here is how Facebook defines it: “a deep learning-based text understanding engine that can understand with near-human accuracy the textual content of several thousands posts per second, spanning more than 20 languages.” I wanted to make the important parts bold, and all definition become bold!

Here are the implications for marketers:

  1. It is more than ever easier to market your product. Facebook takes care of you.
  2. It is easy to market your product to the right audience, but it also means way too high competition and consequently higher costs to show your ad.
  3. Facebook collects data related to purchase intent, which earlier was an option in the targeting methods, which is no longer available. And this may mean Facebook decided to show the ad to the right people itself, but the ad costs show that it is not a valid argument.

Implications for people?

Well, your messages are not end-to-end encrypted! And not only in Facebook Messenger, but apparently in WhatsApp and Telegram, as well. A new recommendation is open-source Signal. Read more about it here: