Twitter suggestions
With Elon Musk attempting to buy Twitter, I figured I would pull together all my thoughts on Twitter from prior articles into one overarching set of recommendations.
Focal Points
The focal points for a revitalized Twitter should be
The interest graph
Direct Messaging and communities
Ending harassment
Ending the bots
The interest graph is the core of the unlocked value of Twitter. The related product that is stagnant is ‘Lists’. Baby steps have been taken here, such as being able to follow a topic. Oddly, I mainly see cryptocurrency as the topic I should follow. This has been particularly bad for onboarding (see my blue sky proposal below). An interest and list-focused onboarding would enable a much broader user base
Direct Messaging is another area where Twitter and LinkedIn have left billions of dollars of value on the table. I believe I read that until a few years ago, Twitter had half of a product manager’s time on the DM product. Every message that begins on Twitter and ends up on another network weakens Twitter. DMs natural scale from 1:1 to small groups to communities. Twitter should be able to handle all of these conversations on the platform.
If we are going to let it rip on speech, one thing that must go along with it is harassment filters. Twitter should buy BlockParty and appoint the founder to lead harassment tools. These tools shouldn’t restrict what someone says, but rather prevent end-users from seeing the filth. There is no requirement that someone reads a harassing comment.
One of the best reasons to take Twitter private is that the metrics are juiced by bots. Crypto scams are major detractors of the product. Nuke all of the bots once and for all. This would lead to a lot of harassment and misinformation reduction as well. Instagram and Facebook do have issues here, but Twitter stands alone in a negative way on this issue.
Blue sky ideas
One company slowly building an interest graph with flawless onboarding is TikTok. If Twitter really wants to go for it, it should be possible to use the product without following anyone. Just open the app and scroll. Like TikTok, the product should learn your interests and feed you great, relevant content. For the current user base, we can keep the timeline and follower dynamic. To grow past the stagnant present, Twitter could do worse than by mimicking TikTok. Whether this is a personal newspaper or an algorithmic time-waster, we can figure out the details in testing.
A TikTok-style feed could also show 1 piece of content at a time, opening up the door to a Twitter focused on all forms of content. Text clips could live alongside video clips, audio, and pictures. Yes, this is ending some of what makes Twitter unique. However, Twitter’s current business is a clunker. Video and pictures would certainly monetize better for ads.
Dead Ends
I see BlueSky (the decentralized blockchain Twitter, not my previous section) as largely a distraction. The proposal is both technologically years away, as well as a bad vehicle for generating value. It is a noble idea to preserve free speech, but this can proceed on its own. Blockchain struggle to scale at orders of magnitude fewer data and usage than a Twitter app would need, this is a red herring for now.
Open sourcing the Twitter algorithm is another meaningless idea. This sounds great in theory. In practice, machine learning algorithms are not explainable in any meaningful, human comprehendible way. A Twitter algorithm that is simple enough to be open-sourced and explained would be rules-based and essentially chronological or thematic. If we want to argue for that, thats fine. That is likely a financial dead-end however.
Subscription social media products don’t sit well with me. Revitalizing Twitter would require injecting new users and life into the platform. Subscriptions will play to a very vocal minority of users. Twitter already plays heavily to this user base, and the product clearly struggles to expand beyond that pool.
This is a pet issue: with all of the platform troubles, I find it very hard to justify engineering effort on NFT profile pictures. This feels non-strategic, to say the least.