[SMK] Social Media Knowledge


Twitter algorithm tweak: relevance trumps recency

With precisely zero fanfare, Twitter changed the way it picks and shows search results.

Following Facebook

Emulating how Facebook handles its feed, Twitter search now shows the most 'relevant' posts first rather than the latest ones. 

Aware that not everyone will like the change, Twitter remains flexible. Users can still manually filter their search results according to ‘relevant tweets’ or ‘latest tweets’. The default setting, however, is now all about the relevance.

Complex relevance

Understanding what makes something relevant is pretty hard for a computer to do. It's something the engineers at Twitter have grappled with.

In blog post from late 2016, Lisa Huang, senior software engineer in Twitter’s search quality team, explains the difficulty of prioritising relevant tweets in a search result.

“They could be searching for popular Tweets to engage with or to better understand context around the search query, and the most recent Tweets are not necessarily best suited for that,” writes Huang.

Making it relevant

To ensure greater relevance, Twitter retrieved tweet candidates from various sources within a larger time range and ranked them with a machine-learned model and performed several experiments with variables.

The machine learning method helped the Twitter team to decide how tweets will be ordered.

“A person’s behaviour on Twitter provides an invaluable source of relevance information. This data contains both attributes about the Tweets shown and the consumer’s reactions to them. Using this information, we can train machine learning models that predict how likely a Tweet is to be engaged with (Retweets, likes and replies),” writes Huang.

Relevance engages 

Apparently, the Twitter team actually implemented the algorithm update in September of 2016, but opted to ‘test the impact’ for a couple of months before publicly announcing the change.

Since going live, Twitter has found that users who experienced this new Search results page engage more. Also, users presented with relevant results Tweet more and interact more with other users when compared to users presented with recency-based results.

Copy Transmission is a Melbourne-based agency :: Better Brands. Loud & Clear.

Learn with SMK through May

Start your SMK: Digital Excellence 7-day free trial today and unlock unlimited, on-demand access to hundreds of hours of digital masterclasses, training courses and hands-on tutorials.

Your risk-free trial offers the perfect opportunity to expand and upgrade your digital intellectual property.

Each month SMK releases 30 hours of new and updated social media and digital marketing educational course content, ensuring you never get left behind – be it on digital strategy, tactics or implementation.

Alongside this, SMK offers live help and support weekly within the SMK Working Group. It might be a quick fix or the root of a bigger problem; either way, a problem shared is a problem halved. On a day-to-day basis, SMK’s team gives you hands-on support and fresh ideas.

Not to mention a shoulder to cry on, occasionally.

May's Courses, Live Help & Support Options

Live Member Clinics every Monday & Tuesday from 1 pm – 2 pm AEST

  • Live help and support from SMK’s team of analysts.
  • Book in to request a personalised discussion for 15 or 30m via Zoom within the Facebook Working Group.

Weekly Technical Labs: Meta Business Suite on Wednesdays from 1 pm – 2 pm AEST

  • Technical Labs explore the technological process and workflows related to key digital marketing activities.

Google Analytics 4, Data Analysis & Evaluation Masterclass on Thursdays from 10 am – 12 pm AEST

  • Module 1: GA4 Optimisation, Key Features, Tools & Reports
  • Module 2: Setting up GA4 Conversions & Understanding Analytics Events
  • Module Three: Analytics campaign tracking and report analysis
  • Module Four: GA4 report round-up, conversion attribution & visualisation

Influencer Marketing Masterclass: Organic, Paid & Commerce on Fridays from 10 am – 12 pm AEST

  • Module 1: 2024 Influencer Marketing Trends, Forecasts & Opportunities
  • Module 2: Organic Influencer Marketing Best Practices
  • Module 3: Optimising Influencer Campaigns With Social Ads
  • Module 4: Evaluation, Reporting and Influencer Commerce

Leave a Comment