Meet the indie hackers cracking the X algorithm

For an indie hacker, posting on X is a lot like panning for gold. If you get good at it, you can get very wealthy very quickly (just look at Marc Lou and CodeFast!). 

The only problem with gold is the actual process of mining it. Think of the X algorithm like the gold mine. It changes location often. It runs dry after a while. And there are well-established ways to maximize your results.

In this article, we’ll dive into what the algorithm currently says and hear from some indie hackers on how they play the most important game in social media. 

How the algorithm works

When Elon Musk purchased X, he promised to open-source its recommendation algorithm. That promise was fulfilled in March 2023. 

Unfortunately, it hasn’t been updated since July 13, 2023, which makes it more of an archived version than an open-source model. Still, even though the algorithm frequently changes, there are valuable tidbits in the archived version that seem to stand the test of time.

At its core, all the algorithm does is mathematically guess what content you may like based on what you've liked in the past. It does this by ranking all of your interactions on the site. So, the algorithm notices whenever you like, reply to, report, or bookmark a post.

It will then take this data and show you 1,500 posts ranked in order of how likely you are to interact with them. The more likely you are to interact with it, the higher in the feed it appears. This is your For You timeline.

For creators, this ranking system makes the X algorithm somewhat game-able. According to Alex Finn, an X creator with 247,000 followers who has spent dozens of hours studying the algorithm, the current best practices are to:

• Prioritize replies and clicks on your profile, as these are worth up to 13.5x more than likes. In other words, you want your content to be thought-provoking enough so that people reply to you and check out your profile.

• Be verified (pay for X Premium). This is a must if you want any chance of growth. 

• Keep a 60% following to follower ratio.

• Never include a link to an external site in a post. If you do need to include a link, make sure to do it in a follow-up post that replies to your first post.

It's hard to overstate how big a difference the external linking rule can make:

Metas on metas

If these general principles stay constant, why are so many people so frustrated with the X algorithm?

Well, that's because it seems to have this weird habit of randomly prioritizing different forms of content. This gives rise to ever-changing "metas." We've already seen a bunch of them in just the last few years:

• In November, the algorithm was pushing content that had to do with the overall world, especially when it was political.

• October was ruled by interesting questions that could really pull in people's attention.

• September was dominated by videos.

• The summer was full of “honestposters”, which was basically just the anon account Kache yapping incessantly. 

• May was lifestyle porn and screenshot essays.

• 2023 was the time of LinkedIn style one liners, and 2021/2022 was when the Threadbois reigned.

As you can see, it basically takes a ninja to navigate these changes. 

Thankfully, I was able to talk to four such ninjas: John Rush, Edward Sturm, Dhravya Shah, and Channing Allen. 

John is an indie hacker with $2M in ARR, 55,000 followers on X, and tens of millions of impressions. Edward is a social media guru who has pulled in hundreds of millions of views and has worked with companies from Microsoft to Time. Dhravya has 24,000 followers on X and has already exited two startups...at just 19 years old. Last but not least, Channing is one of the founders of Indie Hackers, sports 19,000 followers on X, and has interviewed hundreds of founders.

Here’s our conversation. 

How to win on X

Q: How often do you feel like the X algorithm changes?

John: Every 1-2 months on Friday night. It also feels like they keep rotating the same algorithms, trying to play against those who “cracked the algo.” Like changing your wifi password every month by rotating ten passwords.

Edward: It, like Google’s algorithms, has around three major updates a year.

Dhravya: I think the algorithm is (definitely) always evolving, but human psychology always remains the same. I think the best way to grow is also the most obvious one - thinking from first principles, "what would my audience like?" Usually, people like others being real, or people that inspire, help others grow, etc.

Channing: There seems to be a "macro change" to the algorithm that happened around mid-2024 (if not earlier in the year) which hasn't changed that much, whereas "micro changes" to the algorithm are probably taking place constantly.

The macro change, leading up to the election, was a massive engagement boost for conservative political content — especially content that is dramatic and controversial.

Q: How can you tell it changed? 

John: My feed is totally new.

Edward: Because content from me or my companies consistently performs unexpectedly or I see new formats and tones in my FYP. For example, more negativity or more positivity. Or more or less threads.

Dhravya: Mostly from my own numbers. There are weeks when I get 10,000 views a day, and there are weeks when I get millions. It all depends and honestly feels very random.

Channing: My evidence is anecdotal. I'm seeing changes to my feed and to the engagements on my personal posts, and I'm hearing similar opinions from friends and entrepreneurs in my network, many of whom are leaving X for Bluesky.

Q: How do you find out what is the new “meta”? 

John: I open the feed from my passive account, the one where I don’t engage. So that I see the default feed.

Edward: I have a group of friends who are constantly sharing learnings with each other. I do the same thing for TikTok, Instagram, and YouTube. This is the fastest way to improve and learn.

Dhravya: Following interesting accounts. Every once in a while there's an account like that, the "pioneers" in a small social construct. They are hard to find, but most trends emerge from them. For example, aidenbai right now on react scan, itsandrewgao a little bit ago talking about new and upcoming startups, and levelsio with his real but insanely good content about life, health, startups, revenue, and tech in general.

Q: How much do you pivot your content, if at all?

John: I pivot it right away when the algo changes.

Edward: The only times I’ll pivot my content is for content types. For example, threads vs long posts. Videos vs images. It’s the same reason to post Reels on Instagram - they’re so so good for growth. Aside from this, I don’t believe people’s instincts in content change much. So the subject matter and tone stay.

Dhravya: I do not pivot. Sure, I do ride the wave every single day - gotta play the game. But, while doing so, I make sure to stay real and true to my audience. For me, retention is much more important than numbers.

Channing: I've been inactive for a bit, but I'm about to go heavy on quote tweets and on image- and video-based threads. If I share links, I'll make sure to bury them in threads that start out with compelling image-based posts with strong hooks.

Q: What’s the current meta? 

John: Double hooks with a media preview, followed by a thread.

Edward: Long threads with images/video in every other post, strong hooks, and hooks at the end of each post in the thread do very, very well. “They got obliterated. But you’ll never guess what happened next…” You write a thread and have hooks like that at the end of each post in the thread to get people to read the next post.

Dhravya: The current meta is different for everyone, but in general, for me, it is: generative UI, databases, the fight between payment apps, devil release + OpenAI releases and hype around it, people building something and sharing their journey, how good Claude is, and showing X analytics and payouts (lol).

Channing: Overall, this year, the algorithm has seemed to change in the following additional ways, from my point of view:

1. More emphasis on visual media content, i.e. posts with images, GIFs, or videos are prioritized over text-only posts.

2. The algorithm loves conflict and drama — even more than before.

3. Elon Musk's posts are massively boosted to all users, even those who don't follow him or engage content like his.

4. Text-heavy threads seem to be falling out of favor, whereas two years ago these were the best way to go viral. (That said, image- and video-based threads seem to work really well. So it's not an issue with "threads" per se, so much as text- or idea-based threads.)

5. The algorithm seems to adapt really quickly to trending topics. i.e. merely talking about popular topics and live events seems to give posts more engagement, even if you don't use outdated mechanisms like hashtags.

6. Link posts do terribly now, which Elon all but admitted in an exchange he had with Paul Graham.

Q: What kind of content succeeds regardless of the algorithm?

John: Cinematic stories in any format. Of course, having a mix of a story and the current thing for the algo is the optimal format. 

Edward: Regardless of the algorithm change, unique hot takes, juicy information, and viral stunts always do well. Always. People always want this type of content, and when they get it, they will consume all of it, share it, save it, comment on it, and use any metrics that are important to performance.

Dhravya: When you stay real to people, think from a psychology perspective and try to be deliberate about 'how you want to make your readers feel,’ anyone can grow their audience sustainably.

The player matters more than the game

If there’s one thing to take away here, it’s that the most important factor for X success is not the algorithm, but the poster.

There’s a reason why John, Edward, Dhravya, Channing, Levels, and Marc Lou get such great engagement, regardless of the algorithm. They know how to play the game, yes, but they are also genuinely interesting people who consistently post stuff people want to read.

And that’s the real game. You can growth hack the algorithm all you want, but unless you’re an interesting follow, any success you find will just be fleeting.

At the end of the day, it all comes down to posting bangers.