What Twitter Got Wrong and How To Fix it

The Current Plan

Elon Musk just announced the plan for Twitter Blue, which will cost $8/month. Here’s a list of proposed features:

  1. Account verification (similar to the current blue check)
  2. Boosting of their tweets in searches, replies to other accounts, and in user’s mentions.
  3. Ability to post longer videos and audio
  4. 50% less ads.

I think these features are misguided for a couple of reasons, which I’ll address individually.

They don’t solve Twitter’s problem with bots and scammers

Musk’s tweet suggests these features would be crucial to solving the bot problem, but I strongly disagree.

Why we can’t assume malicious users won’t purchase verification

Firstly, the current problem is not all future problems. Bots spray Twitter with insane volumes of low quality spam because that requires the least possible effort. But it’s not to say that is all they can do. If user verification made mass spam less viable, they’d pivot to more targeted alternatives.

I actually don’t think verification will even combat spam at all. Firstly, the kind of users falling for these scams are likely not thinking critically in the first place (a website that doubles your bitcoin? really?). But more importantly, I think it’s unlikely enough people will pay for verification that users will start to see unverified accounts as less trustworthy.

In fact, giving the option to boost accounts visibility could open the door to more problems. Scammers, spammers, and bots need only exceed $8/month worth of value from an account for verification to be worth the cost. Social media verification has always given an illusion of trustworthiness. it could certainly be worthwhile for bad actors to pivot to more convincing messages from verified accounts, rather than carpet bombing with burner ones.

Give that state-backed disinformation ops have always been a cost center, I can’t see any real effect there. In fact, being able to buy verification (even if the account gets banned after a week) would be highly beneficial. 1000 verified accounts for $8,000 a month is simply a rounding error for an influence op.

Hijacking of verified accounts

As someone verified on most social media platforms, I see a lot of account hijacking attempts. Some of these attacks are really clever, and I could see myself falling for them if I didn’t work in cybersecurity.

One such attack involves fake influencer marketing agencies contacting verified users, offering them thousands per month in sponsorship revenue. The income ranges are not unbelievable, often inline with what an influencer with my following could expect to make. The trick? Most services designed to enable brand collaboration can also be used to take over accounts. It’s very difficult for the average user to tell if they’re granting access to posting, metrics, or full account control.

Another common trick is one where you’ll get an email from ‘Verification Support’ or ‘Trust and Safety’. The email will usually be something along the lines of you having broken the rules and are at risk of losing verification. This is a big fear for many verified accounts because it is quite easy to actually lose verification. On some platforms it has often been possible for malicious users to get even major celebrity accounts banned via report bombing or insider access. The email will either contain a phishing link, or sometimes a link to a legitimate page which grants the attacker access via OAuth.

Twitter has already experienced massive issues with verified accounts being hijacked to promote scams. Creating even more accounts is most certainly going to make the problem worse, not better. That’s not to say verification couldn’t do with a rework, but it is not the solution to spam, scams, and influence ops.

Pay-to-win, user suppression, and the death of quality content

Yesterday I made a tweet saying boosting users who pay for Twitter Blue equates to suppressing those who don’t, which confused some people. Many saw priority tweets as akin to offering first class seats on a plane (i.e. an optional extra), but this assumption is wrong.

Twitter is a zero-sum game. There is a finite amount of attention, from a finite amount of people, reading a finite amount of tweets per day. If you boost the amount of attention one user gets by giving them ‘priority’, it stands that you are taking said attention from non-priority users. People’s belief seems to be that high quality users can afford to and will pay for Blue, whilst low quality users can’t and won’t. There is no correlation between good content and income, or bad content and lack of. In fact, some of the largest most profitable accounts are hate farms.

When I started out on Twitter, I was broke. I had enough money to afford a couple of video games per year but nothing more. My account grew because people liked my content, but my income didn’t change. I wasn’t doing Twitter for money, so I wasn’t trying to monetize my platform. Eventually, as a result of my content, I got a well paying job. I may be able to afford Twitter Blue now, but I’m not sure I’d be in the same position could I not have grown my audience organically and for free.

In fact, I can’t think of a single account I follow who is doing Twitter for profit. People making content for free aren’t going to pay for ‘exposure’. Not to mention, people looking to pay for reach can already do so via promoted tweets. Why turn the entire platform into one where every user is just an advertiser in a trench-coat. With this model, Twitter risks losing the many non-profit creators who make the site great.

The right way for Twitter to go

Musk has a lot of the right ideas already, but the implementation seems wrong.

Paying creators is absolutely the right way to go. This is what make every other social media platform work. Users are not customers, they can be either products or employees, but neither are paying for that privilege. Now, I’m certainly not saying there aren’t things user will pay for. Cosmetic items in games are popular (I’ll reluctantly admit I’ve spent hundreds on them). TikTok users love to buy virtual gifts to give to creators, who can then redeem them for money. There are plenty of ways to monetize users, but having them pay for basic usability is not one of them. Not to mention, even if every single user pays $8/month for Blue, that’s just a maximum of $8/month that can be paid to creators. Not very appealing.

Donations / Gifting

One great idea Musk previously touched on with Dogecoin: the tipping economy. Though I’m not sure Dogecoin is the right medium, the idea is solid. Instead of allowing people to directly pay money to creators, you can create an abstraction such as virtual tokens or gifts. By creating virtual tokens that can only be purchased on-platform, Twitter can take a commission. In fact, they can take a cut when users purchase token, then again when creators redeem them for real money. It’s very costly to set up and maintain a billing system, so a 5% cut for someone to do it for you is a no-brainer.

It felt like Twitter almost hit the mark with Super Follows, but the implementation was poor. Posts from people you super follow would just go to your normal feed, making it very hard to keep track. There was also no subscription tiers, no gifting of subscriptions, or any integration with external websites. The whole thing just felt awful from a user experience perspective, so I ended up moving to Patreon.

Videos

Let’s face it, videos are the future. Since every platform just rips of everyone else, there’s no shame in stealing ideas. Plus, doesn’t Twitter technically owns Vine, which created the new paradigm for video? I find that when I post video content, even on Twitter, I get more engagements. The problem being not everyone on Twitter wants to see video. Mixing them in with the regular feed isn’t ideal. A new tab in the app which mirrors the functionality of TikTok, Reels, Shorts, Vine (or whatever else), is most likely the way to go.

Video content is also a value ad for both creators & advertisers, which I’ll get to later.

Content tuning & the need for algorithms

Another great idea Musk touched on is the ability to customize content. I think a good mix of explicit and implicit (algorithmic) customization would be great.

Currently, Twitter feels like a cost center, and most that I’ve spoken to agree. Twitter is extremely emotionally taxing, especially for big accounts. Every tweet will ultimately attract disinformation bots, trolls, and people who insist on seeing everything as a political statement. There are few tools provided to users for countering toxicity; blocking accounts doesn’t hide their replies to your tweets from other users or even prevent them from continuing to reply.

The problem with Twitter’s previous attempt at algorithms and customization was entirely down to poor implementation, which seems to be the recurring theme. Twitter sought to replace the chronological feed by forcing users onto the algorithmic one. On platforms heavily based around chronological order, forcing a chance is never going to be popular. Ideally Twitter needs both a chronological feed and an algorithmic one. Once the algorithm proves itself, people will naturally migrate.

The other core problem was Twitter’s algorithm was it focused too much on accounts rather than content. You’d basically just end up seeing every single though from the accounts you most interact with, rather than the content you want to see. Twitter did later figure out recommending content rather than people with ‘Topics’, but by then the algorithm was already doomed.

Why TikTok is king

I’m a big fan of TikTok. Despite the controversy surrounding its owners, it is a genuinely great platform. What’s unique about TikTok is firstly, their algorithm is good, like really good. But more importantly, TikTok was the first platform to divorce itself from the follower/following model. Whilst you can follow accounts on TikTok, most users prefer to just let the algorithm find the content they want for them, and that it does.

Traditionally, social media reach is based on how many followers you already have, rather than the quality of each individual post. On a platform like Twitter, a single word tweet from a large account will easily gain more interactions that most people’s best content. TikTok creates a system where content is judged primarily on its merit. A brand news account with zero followers can get just as many views as the platform’s most popular users. Admittedly, algorithms are much easier to do with video content where you have metric like watch-time.

Not all algorithms are created equal

Twitter actually does the whole algorithm a bit backwards. With quote tweets and replies both increasing reach, I often find my feed filled with whichever god-awful takes are getting most ratioed that day, rather than content I enjoy. The platform implicitly boosts users who generate the most engagements, which rewards controversy and hate.

Another area where algorithms can go wrong is political polarization, which is already a massive problem in the US. Platforms like YouTube, TikTok and Facebook have all struggled with this is the past. It’s good to focus on what people want when it comes to their interests, but categorizing by political views is a recipe for disaster. Once a user has been isolated into a political echo chamber, away from all opposing views, it is natural that they will progress towards the extreme end of the spectrum. Without opposing views there is no norm, middle ground, or balance.

Ideally a good algorithm should neither isolating users from opposing political views nor blast them with extremism and hate.

Fixing the advertiser economy

Promoted tweets have been Twitter’s primarily source of revenue for eternity. They sort of work, but they’re not great. As is evident from their attempt at algorithms, Twitter has never been good at categorizing users by their interests. Instead, the platform encourages advertisers to force tweets into the feeds of random users, who likely aren’t interested in whatever’s being sold. Promoted tweets are also just normal tweets, so users can simply block brand accounts to avoid them. As a result, promoting tweets can lead to brands losing reach rather than gaining it.

Why the creator economy is superior

What other platforms get right is integrating advertisers and content creators. This generated revenue for creators, awhile making ads that people might actually watch.

TikTok is an interesting case study, because they have something similar to promoted tweets, but in an interesting way. Brands will often partner with popular creators to produce videos ads that are entertaining. People will often sit through ads for products they aren’t interested in because it’s entertaining or features their favourite creator.

The focus though, should be integration of ads into regular content. There are two ways to do this, one of which many platforms fail to capitalize on.

  1. In-content ads - this is only really viable with long form content (blog posts, video, podcasts), basically anything but tweets.
  2. Sponsored content - advertisers pay creators for shout-outs in a videos, a tweets about their product, or wearing their brand in an instagram pics.

Sponsored content has the benefit of being unaffected by adblock, non-invasive, and non-excessive. It’s easy for a platform to become unusable due to too many ads, but when creators over do it they’re punished by losing followers.

Most platforms fail to capitalize on sponsored content. Sponsorships are typically negotiate directly with creators, thus the platform doesn’t get a cut. TikTok really shined here with their “Creator Marketplace”. They built a platform where brands can find creators and creators can find brands. In doing so they provided huge value for everyone, and were able to take a commission in the process.

for smaller creators, it can be very difficult to find sponsors, especially when agents are only willing to work for large creators where they can make huge commissions. Most platform attempt to fix this by paying creators to enable the platform’s generic ads, but TikTok is the first to help with sponsorships. Studies show that consumers are highly likely to buy products recommended by small creators, as these creators build strong loyalty and trust through direct interaction with individual supporters. Smaller creators are also more likely to recommend product they like, and their fans will like, rather than the product of whoever will pay the most. As a result there is currently a huge amount of brands looking for small to mid-sized creators.

Conclusion

I think it’s clear that for a long time, Twitter has been on the right track but facing the wrong direction. It’s understandable that the new management may believe there is a need for a massive overhaul of their business model, but I really don’t think this is the case. The platform has been doing all the right things, just is the worst way.

Twitter has always been somewhat out of step with their users. They implement features without seeking feedback, do it in a way that’s less than ideal, then revoke it when everyone ultimately hates it. Musk and others have been very active in engaging with users, but still seem to be making the same mistakes of seeking user feedback only after announcing changes.

For certian Twitter need to shift focus away from trying to turn users into customers. There are features users will pay for, but they are few and far between. It’s unwise to assume users are getting value out of a platform where they create content for free. If Twitter instead enable users to make money, then there is need for subscriptions. Costs can simply be in the form of commission.