Elon Musk’s efforts to turn Twitter around are uncoordinated, gut-driven, and self-sabotaging. But it doesn’t have to be this way. There are true and tried Growth principles that worked for many companies before.
In this article, I explain:
- The basic parts of a Growth model
- How applying these parts leads to 9 Growth ideas
- How Twitter could apply each of the ideas to turn the company around
My primary goal of this exercise is not to induce change at Twitter, only a few people can actually do this, but to inspire you. Some of these ideas might apply to your company as well.
The 3 quintessential parts of a Growth model
The following ideas are the product of the basic Growth model parts and underlying dynamics.
Every Growth model has three basic parts that apply to any tech company: acquisition, engagement and monetization. You need to bring users in, create meaningful engagement, and monetize the created value. You can’t build a scalable company without getting each of these steps right. The principles are universal.
Done well, each of these parts can lead to powerful Network Effects: unique moats that lead to competitive advantages. An example is TikTok’s algorithm, which creates a uniquely sticky experience that other platforms can't compete with at this moment. This Network Effect leads to more users coming on the platform, which in turn makes it stronger.
The basic parts of a Growth model can reinforce or compete with each other, which is why they need to be evaluated in isolation and holistically. Higher engagement can lead to more users, like a hot debate on Twitter encouraging more users to join. On the flip side, stronger monetization could decrease engagement (example: more ads in the feed) or increase it (more monetization opportunities for users that would also benefit Twitter).
Every model has local maxima and constraints. Local maxima occur when a Growth loop has reached its potential. An example is the fact that every market is limited in size by the number of companies or people on this planet (Total Address Market or TAM). Constraints are hurdles, or friction users need to overcome in each step of a Growth model. To sign up, for example, a user needs to provide their name and email address. If the value of overcoming constraints is high enough for users, they’re happy to do it. If not, you have a problem.
The following ideas reflect these basic parts and dynamics by addressing Twitter's three core growth problems and offering 3 solutions each.
Problem 1: Twitter has no user acquisition loop
Twitter grew by +26.3% (year-over-year) in 2020 but only +9.2% in 2022, meaning it's significantly slowing down. On top of that, Twitter was only profitable in 2018 and 2019 (lawsuit settlements didn't help).
Musk's acquisition resulted in an annual credit interest of $1B that comes on top of the current losses. Paying this incremental cost with only 237M users would mean Twitter had to add $4.20 (HA!) to the current average ARPU (average revenue per user) of $20 [source].
Making it profitable will take more than a clever monetization model. Twitter needs more users, but there is no user acquisition engine to bring more to its platform!
Idea 1: Build a user referral loop
One principle in user acquisition is to seek opportunities to strengthen existing growth levers before of creating new ones. Twitter’s strongest lever for user growth is the existing user base it already has.
Why is there no incentive for me as a Twitter user to invite my friends who are not on Twitter? This is Growth 101. Many marketplaces like Instacart, UBER or Airbnb already did this years ago (maybe decades). Heck, TikTok even asks for access to your contacts (though I don’t suggest Twitter go that far). Newsletters like The Skimm, The Hustle or Morning Brew reward community ambassadors that drive net-new subscribers with swag and other benefits.
Twitter has a real opportunity to reward users that invite new users to the platform in creative ways:
- A special badge that shows up next to the user name or profile image
- “Boost credits” that allow users to boost the visibility of their posts
- More algorithmic visibility for their Tweets
The constraints here are that people only invite their friends if the incentive is strong or the content good enough. Eventually, this growth loop will dry up as people have invited most of their close network. However, compared to other user acquisition loops, referrals have a longer shelf life.
Idea 2: Expand the partnership with Google
Another principle in user acquisition is to expand the surfaces you already have before creating new ones. Twitter’s second most powerful surface, right after its app, is Google. Twitter already has a partnership with Google (since 2009) to display the latest tweets next to user names.
Let’s think about what Twitter is for a second: a discovery platform that delivers access to the fastest update cycle on the internet. Where else do you get updates on recent development as quickly as on Twitter?
The next logical step would be to expand this partnership to cover generic terms with high-news intent as well. Searching for “crypto”, for example, shows a huge Top Stories SERP Feature covering the meltdown of FTX.
I argue there should be a SERP Feature for Twitter that shows the latest Tweets about the situation to drive new users to the platform. Twitter could flex its core muscles, meaning show the value that made the platform successful in the first place - and that’s relevant news and expert takes.
One big constraint I see is content relevance. Before Twitter expands its Google partnership, it needs to find a strong signal to highlight the most relevant tweets about a topic. You can see that this problem isn’t solved when searching for a topic on Twitter. One remedy could be to surface only Tweets from Blue Checkmarks or Official users or filter Tweets by engagement for a topic.
Idea 3: Stronger activation on the homepage
Building on the previous idea, Twitter’s core product is the content its users create. One of the most important surfaces for marketplaces is their homepage because that’s where a lot of users go after seeing ads, hearing about the product from friends, or clicking on backlinks from other sites.
Right now, Twitter’s homepage is just a feed of trending topics - not very engaging to sign up!
What if instead, users could select topics they cared about on the homepage and see a custom Twitter feed based on the latest and most engaging Tweets (“peak behind the curtain”)? That would give them a taste of the product’s core value and make the onboarding flow easier because Twitter would already know what the user is interested in and could suggest relevant accounts to follow.
Linkedin does something similar on its homepage.
The constraints are the same as in ideas #1 and #2: the content needs to become more relevant for this to work.
Problem 2: Twitter’s engagement suffers because of irrelevant content and low creator engagement
Once users are on the platform, they need to experience regular (daily on social networks) and meaningful engagement. Otherwise, users won’t retain. Twitter’s engagement has tons of potential but also cracks in the foundation.
Brand search volume suggests that Twitter gets more engagement than Linkedin, Reddit and Snapchat since users searching for social network brand names often want to log in.
The following data from Semrush shows brand search volume in the US and globally in November 2022 for 5 big platforms:
- “instagram”: 30.4M US, 330M global
- “twitter”: 20.4M US, 140M global
- “reddit”: 11M US, 24.4M global
- “linkedin”: 9.1M US, 49M global
- “snapchat”: 1.5M US, 6.5M global
Surprisingly, Twitter comes in second. Traffic supports the theory that Twitter’s user base is more engaged than often assumed (+71% comes directly to the site):
That engagement doesn't seem to correlate with user numbers, which goes to show that higher engagement doesn't always mean more users:
- Linkedin 875M global “members” [source]
- Instagram: +500M DAU [source]
- Snapchat: 363M DAU [source]
- Twitter: 237M “mDAU” [source]
- Reddit: 52M DAU [source]
Twitter could make so much more out of its potential for engagement. Many creators (including yours truly) complain about dropping engagement and users about irrelevant content (too much "sh*tposting"). Sometimes, this can be the result of algorithmic filtering due to rapid user growth, but we know that's not the case for Twitter. The combination of lower engagement for creators and less relevant content for consumers leads the platform in the wrong direction.
Idea 1: Add trust signals to the algorithm
To maximize the chances that every user sees the most relevant content possible, Tweets need to be filtered in a similar way as content is ranked on Google: by relevance and authority. Fake accounts and bots are a problem, but if you can create an environment in which they don’t have the same impact, the incentive for spammers drys up.
The way I think about content relevance for Twitter is in concentric circles: at the core are Tweets by users you follow. The priority is innate: you follow them because you want to see their Tweets. The next outer circle should be Tweets that the people you follow liked. After that: highly relevant Tweets on topics you care about with no connection. We see early signs of that on Twitter today but what seems missing are trust signals.
Examples could be:
- The Tweet was liked or retweeted by a user who has high trust
- The Tweet gets comments with positive sentiment
- Users have a verified profile
- Users have a profile with a certain age
- Users have a low amount of downvoted comments
Trust goes hand in hand with verification. Twitter should provide optional levels of verification like email (level 1), credit card (level 2), website through tag implementation (level 3) and photo ID (level 4) that give weight to a profile’s behavior. A higher verification level would allow a user to get broad reach on Twitter much faster, especially if their profile is new.
The idea of improving verification on Twitter was a good one but tying it to monetization was a bad one. The two should live independently from each other.
To be fair, I don’t know exactly how Twitter’s content algorithm works. I am also sure that smart engineers have worked on these problems before. I don’t want to pretend I’m smarter than them. Yet, I think they would agree that these problems are not yet.
Idea 2: User feedback for fine-tuning algorithmic recommendations
One mechanic that the best platforms that filter content algorithmically have in common is that they use direct feedback from users to improve recommendations. That often comes in the form of positive signals, but I think Twitter is at a point at which it needs negative ones as well.
Why is there no button for me to say that I want to “see fewer Tweets like this”? It doesn’t have to be a downvote or hard block but can be a soft factor to fine-tune algorithmic content filtering.
Twitter has started doing that with Topics users can follow, but topics are broad and have many facets. Feedback needs to happen at the Tweet level.
Spotify is a good example. They started providing “+” and “-” feedback for playlist enhancements (Spotify adds songs to a playlist they think fits well). The user feedback allows them to learn what users really want, and recommendations improve over time.
The constraint is that user feedback needs to have a noticeable impact to encourage users to keep engaging with it. If changes are too small, users think the interaction doesn’t have an effect. If they’re too strong, users might fear missing relevant Tweets. It’s a balancing act that Twitter needs to test.
Idea 3: Reward users for link sharing
Over the last couple of years, almost all social networks have become walled gardens by either removing outgoing links altogether (Instagram posts) or significantly decreasing visibility when a post contains links (Linkedin, Twitter). Twitter should strongly reconsider due to its nature as a content discovery platform.
The assumption behind limiting outgoing links is that users who click on outgoing links leave the platform with a negative impact on ad impressions. That might be true in the short term. But long term, users return to the source to discover more. Think about Google or Hackernews - they don't see less engagement because users leave the platform!
Rewarding users for posting links shouldn’t be binary. Twitter should measure how many clicks links get and boost the post’s visibility to encourage users to post more. They should also make the reach of posts with links dependent on who posts them (going back to trust signals).
The biggest constraint is finding the right balance between content and links. If users see too many outgoing links, the platform might lose stickiness. If they don’t see any outgoing links, they might look at other discovery platforms.
Problem 3: Twitter’s monetization model has no platform-fit
Most conversations about improving Twitter over the last few weeks revolved around how to monetize the platform to make Twitter profitable. We now established that user acquisition and engagement go hand in hand with improving monetization.
The key element to monetization is to charge for value. This sounds basic, but it’s not. Twitter has three value receivers on its platform: consumers and creators, who can be and often are the same people, and publishers, who receive traffic from Twitter.
Bill Gates's classic platform quote (“A platform is when the economic value of everybody that uses it, exceeds the value of the company that creates it.”) also applies to creators who build their business on a platform, not just to app ecosystems. The more successful creators and publishers can be on Twitter, the more Twitter will benefit itself.
Youtube is the most successful creator platform, with $50b in shared revenue over the last 3 years. Apart from its recent decline in growth numbers, this model has turned out to be wildly successful. So much, so that other platforms do the same, and so should Twitter.
We shouldn't discount publishers from playing a vital role as well. After losing a lot of leverage in 2008, they might be able to lean on Twitter much more for monetization and create a win-win-win situation.
Idea 1: Share revenue with creators
Sharing revenue with creators is not a new but powerful idea, as agreed by Elon Musk himself. Stephen King had a point when he said that he should be paid to be on Twitter. Popular creators like Marquess Brownlee agree [source].
The basic monetization loop looks like this:
- Creators share great/unique content
- Consumers see ads next to the content
- Creators and Twitter get a share of that ad revenue
- Creators are motivated to create more great/unique content
Creator monetization could go as far as Twitter taking a cut of any sold merchandise or digital goods on Twitter, similar to how Apple takes 30% of sales on iOS apps.
This idea is constrained by creator tools. Twitter needs to figure out what unique tools to build for creators, so they create unique content. Tweetdeck is a good product that hasn’t changed in a long time. Twitter should be able to serve inspiring prompts to creators and deliver statistics about what content has performed best in the past. Superfollows and subscriptions went in the right direction but didn’t go far enough. Creators of a certain audience size should receive unique ways to show ads next to their content, moderate content (as suggested by Noah Smith) and get better analytics data, so they understand what resonates.
Idea 2: Share revenue with publishers
In line with the idea of rewarding link sharing, Twitter should make deals with big publishers that allow them to monetize their content on Twitter, as suggested by Elon Musk before.
A common problem users complain about is finding paywalled content on Google. Twitter could solve this issue by allowing users to pay small sums for paywalled content on Twitter. Twitter Blue went in that direction, and there is real user value here. I already mentioned that Twitter is a discovery platform, so the idea of paying for gated content fits the product.
Let’s run through a hypothetical example. Say, Twitter users can buy access to a paywalled New York Times article for 50 cents, and Twitter would get a revenue share of 10 cents (20%). Let’s also assume 11.5M (5%) of Twitter's 230m users read it. The article would generate $5.75M in publisher revenue, and Twitter would get a share of $1.15M. The Times published a little over 50K articles in 2021 on its core site. You quickly figure out that the revenue potential for all publishers is massive.
Two assumptions need to be validated to understand the constraints. First, users are willing to pay micro sums for content. The topic of micropayments is controversial, and the verdict is by no means out. Second, micropayments don’t cannibalize publisher subscription models. Some publishers like the New York Times, The Wall Street Journal, or The Economist charge their users directly for content. Allowing other users to bypass this could have cannibalizing effects.
If these two concerns turn out to have no substance, though, Twitter could create a fluid ecosystem of monetized content.
Idea 3: Monetize user tools
Twitter has a low signal-to-noise ratio that can make it feel like the biggest firehose on the internet. It can be incredibly hard to sift through all the information sometimes. There is an opportunity to allow users to dial their experience up or down and charge for that.
Twitter Blue went in the right direction with Top Articles from people you follow and from people they follow.
But there could be a lot more:
- Limit your experience to the top 5 Tweets of the day
- Get an email summary with the top Tweets from people you follow by engagement or top Tweets on the platform
- Be able to turn Tweets people you follow liked on or off
- Follow hashtags and get a curated email at the end of the day
- Pay for an ad-free experience
- Pick which Tweets from people you follow to prioritize in your feed
- Get notified when your name is mentioned but not tagged in a Tweet
- Start group chats that can be turned into Tweets
The constraint with monetization is always whether customers see enough value in features to be willing to pay. On the consumer side, Twitter needs to figure out what core features users really want and whether they're willing to pay $5, $10, or even $20 a month.
Part of why Twitter Blue wasn’t a smash hit is because it was a bundle with largely unconnected features that didn’t solve strong pain points [source]. A paid consumer plan needs to go back to what made Twitter successful in the first place (discovery) and make that 10x easier.
The killer scenario: Bringing Vine back
There is one scenario that could catapult Twitter from shipwreck to spaceship: the resurrection of Vine in combination with a TikTok ban or blockade.
At the height of its success in 2013, Vine had 25M users and was the fastest-growing app worldwide. Twitter bought it in 2012 and shut it down in 2016. Even 8 years after its retirement, 90,000 people still search for the keyword “vine” on Google. [source]
The super short-form app was TikTok before TikTok was Tiktok. It offers substantial acquisition, engagement, and monetization potential. One doesn’t have to look further than TikTok itself, which has set sail to make an annual $10B in 2022.
However, this would likely only work if TikTok got banned in the US. Lately, the voices have become louder to ban TikTok due to the security risk of hosting all user data in China with seemingly unlimited access by the CCP. If not enforced by the government, Google and Apple could pull the trigger and decide that the app is a security risk for their customers and simply take TikTok off the app stores. The result would be equal to a regulatory ban.
But hope is not a strategy. Twitter can’t wait for outside factors to determine its fate. It needs to turn things around by building a stronger user acquisition loop, enabling meaningful engagement, and monetizing the platform where users experience the strongest value.