Twitter is bringing in more in-stream ads using Artificial Intelligence

Lately, people have noticed an increase in the number of ads in their Twitter feed (and even on users profile pages). Twitter has slowly been increasing its ad load for some time now, but it seems that in recent weeks, it had increased more than usual.

Twitter was asked about this, and they responded that their team regularly experiments and makes changes to their advertisement experience while also holding up their principles and standards for a high-quality experience for the users. Also adding that they are constantly testing and innovating, and will keep doing it and learn from it.

This means that as Twitter is trying new display methods, users are looking at more ads in their Twitter feed. One reason for this can be Twitter’s need to address slowing ad spend as a result of COVID-19. In its Q1 report earlier this year, Twitter reported a slight decline in ad spend. It noted that the advertising income reduced by 27% Year-over-Year from March 11th until March 31st. This is because of events around the world canceling and sheltering in place in the US.

Twitter is being affected by fewer businesses in operation because of the lockdowns, specifically from fewer sports and events taking place. In order to generate more revenue, it needs to increase the number of ads while simultaneously capitalizing on increased usage during the outbreak.

There is a 24% increase in usage reported in Q1, meaning that ad exposure opportunity for the remaining operating businesses is present, and with increased usage, Twitter can capitalize on this. So, the users can expect to see more Promoted tweets in-stream while the advertisers can have more opportunities to increase exposure to the Twitter audience which is more active than usual.

Twitter will likely maintain such increases in ads to maximize revenue generation if usage levels remain stable. And if they remain stable, you can bet that Twitter will look to maintain any such increases moving forward, as a means to maximize revenue potential. Additionally, Twitter has announced that it will show users more relevant content, based on their activity.

Twitter explains that in the app’s default configuration, the users’ home timeline will show the top Tweets first. The system uses a machine learning model to predict what tweets will interest them the most based on their activities. The model needs to learn this via training and must be constantly updated due to user interest changing all the time. Twitter also added that they have reduced the time required to refresh this model from 7 days to about a day by redesigning the data logging pipeline. This upgrade will make Twitter feed more up-to-date with rapidly changing interests and time.

Twitter says that their new model is much better as compared to their old model in terms of training and real-world prediction after the refresh of the model. Also adding that newer data and models are favored by users, who come to Twitter for the most up-to-date information.

Although there will be more ads in-stream, it will be more relevant in general, based on the new algorithms. Twitter does seem to give more relevant information (except a few tweets that pass through the filter) based on the topics selected displayed in the Explore tab. It will still interesting to see if it changes with the upcoming updates.


Photo: Omar Marques/SOPA Images/LightRocket via Getty Images

Read next: Twitter Is Testing An Instagram-Inspired Feature To Sift Unwanted Messages

No comments:

Post a Comment