Spotify utilizes personalization as it crafts music recommendations for its users, drawing on the previous streaming history to curate the Discover Weekly, the Daily Mixes, and dynamic homepage pop-ups. Since the app accurately predicts what listeners want to hear, they have to make fewer decisions, which reduces friction for the user and overall improves their experience and perception of the app and its algorithm, driving up the listening time. Moreover, personalized autoplay keeps people engaged even while they are not actively browsing music, so the music listening sessions can go on for hours so long as they keep recommending music that the user likes. This ends up in a very high ROI; while personalization is expensive, the improvement in recommendation quality lengthens the listening time of most users, and increases the rates of user retention and of users who upgrade to Premium.
In LinkedIn, we see a different story as (most) users log into the website to invest in their career as opposed to for entertainment. Personalization here is centered around surfacing opportunities that feel timely and relevant like job recommendations, suggested connections, and a feed tailored to the user’s professional interests. Because LinkedIn’s core metric is session frequency, their algorithm is designed to give users small nudges to return more often: a post from someone in their network that performs well, a notification that their profile has been viewed, or a curated set of recommended jobs posted in the last 24 hours. These personalized touches make the platform feel active and responsive, which encourages users to check back multiple times per week. The ROI, while not as dramatic as Spotify’s, is still substantial, as increased session frequency directly translates into more ad impressions, more sponsored job clicks, and higher engagement with LinkedIn’s premium recruiter tools.
TikTok’s use of personalization is a lot more aggressive, as the platform’s entire experience is shaped by the For You Page. The algorithm quickly learns a user’s preferences, including but not limited to how long they watch a video, whether they replay it, and if they swipe away immediately. They then use this dense behavioral data to construct an engaging feed. This level of personalization boosts watch time, but more importantly, it strengthens TikTok’s ad targeting capabilities. Because the app can infer what users are likely to enjoy, purchase, or interact with, advertisers are willing to pay more money to put their ads on it, and the platform can deliver more relevant ads with better conversion rates. The ROI here is extremely high, because improvements in personalization both increase time spent (aka creating more ad inventory) and improve ad performance (aka making each ad slot more valuable).
