So, Twitter has quietly changed the way it shows the search results now. With the update to its search result algorithm, instead of showing what was the latest post under a searched topic, Twitter now shows the most relevant posts first on the search result. This means in place of the earlier reverse chronological order, the search results will now appear in relevance order. Also Read - COVID-19 third wave: Twitter shuts offices as coronavirus cases riseAlso Read - Twitter Voice Tweets rolling out for iOS: What are they, how to send
Essentially, now when you look for something on Twitter‘s search bar, you’ll be shown tweets in the order of relevance from across the platform first on the results page. Apparently, the change occurred in September, and it is only now that it has been announced publicly. However, for people who feel that the older order of most recent posts was more relevant to their needs, Twitter has made the change flexible, by giving a filter which will let them set the tweets to the latest tweets format. Also Read - World Emoji Day 2021: Twitter reveals 10 most used emojis in 2021 in India
But, if you give it a thought, the entire idea of finding relevance in a search result for each user is so complex in itself. And in a blog post, Twitter admits this complexity. Lisa Huang, senior software engineer in Twitter’s search quality team, explains the difficulty of prioritizing the so-called relevant tweets in a search result. ALSO READ: Twitter to turn Vine into a six-second video recording camera app
She talks about how until now the tweets have been in a reverse chronological order, but that format could never ensure that users saw what they were actually searching for. 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. Therefore, in order to improve the results, Twitter search has been made relevance-centric.
But how does Twitter do that? So, 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 like, the candidates for the relevance-ordered results, the number of out-of-order Tweets to show, the layout of displaying a gallery of Tweets as against the individual Tweets, and other different ranking algorithms. The machine learning method helped the Twitter team to decide how tweets will be ordered. ALSO READ: After Twitter, Facebook, now Google refuses to help Donald Trump create a national Muslim registry
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). We can then use these models as scoring functions for ranking by treating the probability of engagement as a surrogate for the relevance of Tweets, the blog read.
Twitter on the blog reminds time and again about the complexity of the algorithm and the constant denoising it had to deal with while testing the format. (Noise, in the context of Twitter would be incidents in user behavior reading where a tweet is accidently or subconsciously liked).
However, in conclusion to all the complex machine learning algorithm, Twitter observed that people who experienced this new Search results page tend to not only engage more with the Search results but also Tweet more and spend more time on Twitter, which makes it a win-win situation for all.