X (not) Gon’ Give It To You: How Threads becomes the De Facto Digital Town Square with better, Personalized Recommendations

What Tiktok did to personalize short-form video, Threads will do for the digital town square.

X (not) Gon’ Give It To You: How Threads becomes the De Facto Digital Town Square with better, Personalized Recommendations
August 10, 2023
 min read
Omair Khan

With the backing of Meta’s machine-learning infrastructure and real-time user personalization, Threads has become a formidable contender in the midst of X.com’s challenges. In this blog post, we will cover 4 reasons why:

  1. Taking a (For You) Page out of Tiktok
  2. Meta’s Ranking Infrastructure Advantage
  3. Leveraging Instagram to Solve the Cold-Start Problem
  4. Optimizing for Experience Vs. Ads Revenue
Threads is a text-based conversation app that fosters real-time discussions and connects individuals based on shared interests.

1. Taking a (For You) Page out of Tiktok

Before creating the billion-dollar companies it's known for today, ByteDance extensively tested a variety of app ideas listed on the left.

Many people don’t know this but TikTok started as a recommendation system and machine-learning infrastructure company. ByteDance’s (Tiktok's parent company) original mission was not to create a video-sharing platform; instead, it aimed to offer real-time personalized content recommendations.

Prior to TikTok's inception, ByteDance’s team explored around 100 ideas, ranging from multiple meme apps to a news aggregation platform, and even a video-sharing app. The engine that powered all of them was ByteDance’s recommendation system.

Different screenshots of the Chinese app Douyin that was launched prior to TikTok

Ultimately, the video-sharing format emerged as the most promising avenue. As they noticed a growing interest in short-form videos, ByteDance became laser-focused on developing a hyper-personalized video platform, leading to the birth of TikTok, as we know it today.

TikTok's personalized recommendations keeps users engaged for an average of 52 minutes a day. With a real-time recommendation system like Monolith, Tiktok is widely regarded as having a world-class, machine-learning, ranking infrastructure.

2. Meta’s Ranking Infrastructure Advantage

Instagram Explore’s state-of-the-art final-pass model infrastructure of how posts are ranked on feeds. The model predicts individual actions that people take on each post, whether it’s a positive action such as like and save, or negative action such as “See Fewer Posts Like This.”

Much like TikTok, Meta has also developed a robust machine-learning ranking infrastructure, allowing them to offer real-time personalized content recommendations. What is clear is that Meta's investment in personalization is paying off, as seen in its latest venture, Threads. Launching with powerful personalization has helped retain users on the platform and enabled content creators to achieve viral success from reliable, positive user feedback.

Without Meta’s machine-learning infrastructure, Thread’s rapid success wouldn’t be possible. Poor personalization would have immediately led to significant retention issues – a death sentence for social media products. Competitors like Minds and Bluesky Social have struggled with personalization on their platforms. This is due to their lack of machine-learning infrastructure and inability to solve the resulting cold-start problem.

3. Leveraging Instagram to Solve the Cold-Start Problem

The cold-start problem refers to the challenge of providing personalized recommendations and tailored experiences to new users when there is insufficient information to make accurate predictions or to understand user preferences. There are two types of cold start problems:

  • Cold-start users: Individuals who join a social network with little to no existing connections or activity history. These users lack the followers, or engagement that are often present for established users.
  • Cold-start items: Pieces of content, such as posts, articles, or videos that are added to a social network with little to no prior engagement or feedback. These items lack likes, comments, or views, which are crucial for algorithms to personalize recommendations.

There are a number of strategies that social networks use to solve the cold start problem, including:

  • Incentivizing users to connect with their friends through rewards and referrals.
  • Using algorithmic recommendations to surface relevant content to new users.
  • Providing a good onboarding experience by making it easy for new users to understand how the network works.
Threads accesses Instagram’s data to suggest who you should follow.

Threads has a significant advantage of accessing a huge amount of data from day 0 of launch, as it is backed by Instagram’s data. Upon sign-up on Threads, every user has the opportunity to connect with their followers that they had on Instagram.

Threads already knows who you follow, what you’re interested in, the millions of likes you’ve done since you created an account, and even which videos and posts you spend more time watching.

4. Optimizing for Experience Vs. Ads Revenue

Threads looks on….as Twitter/X is on fire

Since Meta has deep pockets and less short-term financial pressure, the platform has the luxury of not optimizing for advertising revenue. Twitter/X, on the other hand, continues to bleed money and remains unprofitable, forcing itself to place too many ads on users' feeds. As a result, Twitter/X has become even more crowded with advertising competition, while Threads remains clean from advertising for now.

In contrast, Threads has the perfect opportunity to play social media ‘superhero’ as it doesn’t need ads to sustain the business right now, which ends up being a huge user-experience win. Users see what they want to see. Threads leverages real-time signals such as Impressions, Views, Comments, Likes/Dislikes, Shares, and Follows to keep users engaged and informed constantly. Whether it's the latest news, trends, or interests, Threads ensures that users are engaging in lively discussions that have some relevance to them.

The image above shows a user’s personal feed and the type of content they might be interested in. Based on the accounts the user follows, and the posts that they have liked, Threads curates a feed that fits their interest.

In Conclusion

As Threads continues to grow, it will reshape the digital town square and provide users with a platform to host meaningful text-based conversations. Threads' origin story is similar to TikTok’s, underscoring the importance of real-time recommendations, infrastructure, and personalization in scaling a successful social media product. If you or your company is interested in real-time recommendations, infrastructure, and personalization, you can learn more on the Shaped blog.

Get up and running in just 1 sprint 🏃

For developers

Waitlist for public API keys
You're on the waitlist! We'll be in touch 🙌
Oops! Something went wrong while submitting the form.

For companies 🏢

Schedule a demo with your data️
Thanks for signing up! We'll be in touch 🙌
Oops! Something went wrong while submitting the form.