The Love-Hate relationship called Twitter
Diving into the (unnecessary complicated) Interaction Design of Micro-Blogging
In the last year or so — while a lot of people turned their back on Twitter claiming it has lost its early days spirit — I came to use the service more and more. My usage pattern expanded from pure link-blogging (what I did before on tumblr) to commenting / liking / retweeting (= interacting) with other people’s tweets and therefore opinions.
And that’s, I think, where the magic of Twitter is the strongest. In no other place it’s that easy to connect with likeminded folks or people you admire. (I’m not saying it’s a given that they will get back to you, but the chances are the highest on this particular social network.)
That fact makes it so frustrating to see Twitter struggling in so many places:
- There’s the issue around trolls / hate speech on a micro-level.
- There’s the issue with Trump personally (the chief troll of all).
- There’s the issue with 3rd party clients #BreakingMyTwitter.
- There’s the (underlying) issue of finding the right business model.
These topics have been discussed in length elsewhere. In here I want to add another puzzle piece by investigating their Interaction Design.
Nitpicking about UX decisions
Let’s make a deep dive on one of the features Twitter PM’s messed around with a lot.
Retweets are used by the community for more than a decade (!). It’s a simple use case: That tweet is so good, I will copy it to my followers with the least possible amount of work. Things go viral from there.
But since Likes (the little heart icon) are also pushed in my timeline no matter if I follow the content creator or not, I have an unnecessarily hard time deciding each time if I Like or Retweet something.
Quote Tweets
Furthermore, the Retweet with comment option (formerly known as Quote Tweet) is an extension of the Retweet with the option to add your own 2 cents.
It’s the best way to build upon an existing piece of content because it keeps the context to the original item and it makes sure that all followers will see the addition.
Where this feature falls apart?
- Logically the Retweet with comment is much closer to an @Reply, but it doesn’t initiate conversations as well. There is no obvious way to answer back.
- It’s “invisible” for everybody that doesn’t follow you (unlike the @Reply). On a specific tweet I can’t even see if other people retweeted or retweeted with a comment. (Why on earth should I want the user bio when I click on the Retweet count?)
- Visual hierarchy is flipped on its head. First the comment, then the piece that is commented on. Imagine a blog where you first see the comments and then the post.
- The original tweet is often truncated, with loss in formatting.
- The original tweet can be miss-understood as link card.
It could have been easy!
Why not make Reply’s per default public to all followers? (And not only to the subset of people that follow me and the person I replied to.)
- Since Twitter started to show the replied to-Tweet in my feed, context is secured and in the right order.
- Timeline would be more centered around conversations, not individual statements. (Isn’t that what Twitter should be for?)
- Timeline would be more permeable in the amount of people/opinions you get to see.
- Eliminates the need for a separate Retweet with comment interaction type & the .@Reply (with a dot) interaction type → paradox of choice, again
- By being optional, it would still allow to turn the “push to my feed” off — useful for e.g. support requests.
Tweetstorms
Before the Medium article I tried to convey my gripe with Retweets in this tweetstorm. It’s readability hell.
With some 3rd party support (also not the prettiest webpage around) threads are getting way better to follow visually — a Déjà-vu to days were Twitter users needed an external link shortener.
More UX issues
There’s more. Some opportunities I see to make Twitter more user-friendly:
- being able to edit a tweet with history (here’s a Chrome extension that works somewhat around this)
- being able to set some form of structure with formatting. Slacks leads the pack by integrating Markdown methods. (Of course Twitter is all about KISS, but then again, why are Emoji or very tall ASCII artworks supported?)
- being able to customize or hide the link card
- being able to attach a longform-text in a frictionless way:
It’s easy to create a mess using tweetstorms, embedded tweets from elsewhere and screenshotted texts, where even advanced users need several clicks and cognitive time to process what has been written, who wrote it and how it flows together.
The broader point
These little features are perfect examples showing what Twitter lacks in management skill: Navigating the tight line between incorporating user requests all while keeping the Twitter core simple.
This is the effect of Twitter being exposed to several circumstances in the beginning:
- Getting Product/Market fit pretty much right from the start
- Target demographic of strongly opinionated & outspoken “infovores”
- Launch date at the height of Web 2.0 mashups & open APIs
These first years shaped Twitter critically. In the absence of active management (being busy keeping the FailWhale from sinking the ship), early adopters (hi Chris Messina) created the basic additions like @Reply or App frontends. Because of this authority vacuum, Twitter Management and hardcore users built different viewpoints on what Twitter exactly is:
A centralized service run by a public listed company (that needs to evolve to meet Wall Street expectations) OR a decentralized, open & stable ‘protocol’ partially owned and run by the community.
As I laid out, good Product Management can try its best to mitigate between both camps — but it (even more in the case of Twitter) needs to work a lot closer alongside and for its users. Otherwise there’s no way to bridge the gap that history opened accidentally.