5 subtle A/B tests that Threads have been running since launch

Threads has been out for 2 weeks now, it's been a mixture of:

"Woah this grew so fast, Twitter is dead"

and

"Oh boy do we need a followers-only feed... Twitter rules".

As an IndieHacker, you'll already be aware that A/B testing, and incrementally improving specific engagement metrics (such as posts, or likes), can compound to create meaningful growth.

(That's one reason why we're all here—me included).

And it's inspiring to see a product team grow an app to 150 million users so quickly, and still have the headspace to run a considerable number of A/B tests, on even the most basic mechanics.

For what it's worth, I've added my own opinion about the possible rationale behind each test—although I could totally be wrong. It's just my own gut feel.

1. Hiding likes

The change: They've experimented with showing likes in 3 variations, two of which mask the total number of likes.

Possible reason: At launch, they wouldn't have known how strong engagement would be, so they hid them (celebs might not like seeing "5 likes" on their first post). But now, they're possibly experimenting on the impact of engagement when these are visible / hidden.

i.e., does seeing the number of likes on a post encourage, or discourage, you to post / comment more.

--

2. Hiding the 'Threads badge' on Instagram

The change: On launch, every Instagram profile was automatically updated to include a numbered badge (i.e., your unique ID). This created an incentive for people to sign up early (to get a low number).
Now they're experimenting with showing it (in two variations), or instead having a badge in the top right.

Possible reason: Perhaps to start, a low number was appealing. But after 100 million users, your unique ID would be 104787231... which is hardly something to show off about. It may have even turned influencers away.

So now they're trying to work out which method of display creates a better click-through-rate, pressumably from Insta into Threads. They may also be tracking referals (i.e., what creates the most further sign-ups).

--

3. Thread prompt

The change: Quite early on, they added a prompt to 'add to thread', when posting something.

Possible reason: I imagine that they were monitoring the number of people who created a thread (series of posts), without publishing one, and then publishing another. i.e., creating a whole thread, and then posting all at once.

Perhaps not many people were doing this, and instead all of the posts were singular "hey, I'm on Threads now", and they wanted to encourage more of the long-form content.

--

4. Added a 'plus' to follow

The change: They added a plus icon to a profile, to allow you to follow an account from inside a Thread.

Possible reason: This may have been a bug, because the plus was on the main feed, but not on threads themselves. But, it's also possible that they made the 'follow interactions' more obvious throughout the app, because people were reading threads, but then not following anyone.

--

5. Font changes

The change: Throughout the app I noticed a bunch of text changes, particularly with smaller fonts.

Possible reason: Perhaps they were testing readability / accessibility, or maybe they were just tidying up their design system. I wouldn't be suprised if when the app went live, they tested it on a bunch of older phones, and the font was unreadable.

i.e., newer phone displays are way clearer with small text than older ones.

If the Threads devs are reading this, it's been inspiring to see a team release an app so stable (given they went from 0 - 150m users), and still iterate on the small stuff.

Kudos.

P.S., If you're wondering why I bothered doing this, it's because I was writing a seperate analysis on their launch strategy.