Netflix’s AI recommendation engine drives 80 percent of content discovered on the platform. Amazon attributes 35 percent of its revenue to personalised recommendations. Spotify’s Discover Weekly has generated over 5 billion streams since launch. AI personalisation is not a future state for these companies. It is the primary driver of their engagement and revenue.
AI marketing personalisation in 2026 operates across multiple layers simultaneously: what content a user sees, what products are recommended, when communications are sent, what messaging resonates, and which channel reaches each individual most effectively. The principle is the same at every scale: serve the right content to the right person at the right moment.
Netflix: Personalisation as the Core Product
Netflix’s recommendation system is, in a meaningful sense, the product. The AI does not just suggest what to watch. It personalises the thumbnail shown for each title based on each user’s viewing history. A thriller fan sees the tense action scene. A drama fan sees the emotional moment. The same content, shown differently to each user.
The recommendation model processes viewing history, search behaviour, time of day, device, content completion rates, and behaviour of users with similar patterns. The 80 percent of Netflix content discovered through recommendations versus user search is the outcome of this system. Without it, the Netflix library would be too large to navigate effectively.
The applicable principle: Every content platform decision should be asking: what does this specific user’s behaviour tell us about what they want to see next, and how do we show them the right version of the right content?
Amazon: Personalised Commerce at Scale
Amazon’s recommendation engine generates 35 percent of total platform revenue through ‘Customers who bought this also bought’, ‘Recommended for you’, and email campaigns featuring products correlated with recent browsing and purchase history. The system updates in near-real-time: purchase something, and the recommendations across the platform shift within minutes.
The email channel is particularly notable. Amazon’s personalised email campaigns achieve open rates significantly above industry averages because the content is individually relevant rather than batch-and-blast. Emails containing a user’s recently viewed items alongside complementary products they have not seen convert at multiples of standard promotional emails.
Spotify: Discover Weekly and the Taste Profile
Spotify’s Discover Weekly, released in 2015, delivers a personalised 30-song playlist to each user every Monday based on their listening history and the listening behaviour of users with similar taste profiles. In its first year, users listened to Discover Weekly playlists 5 billion times and saved 1.7 billion songs from it.
The underlying system identifies taste clusters from behavioural data, matches individual users to clusters, and selects songs from that cluster that the user has not heard. The playlist feels personal because it is built from individual listening behaviour. No human curation is involved in the individual playlist generation.
Sephora and Starbucks: AI in Retail
Sephora: Sephora’s Visual Artist uses AR to let users virtually try on thousands of makeup products. The AI analyses skin tone and facial features to recommend products personalised to the individual. The Beauty Insider loyalty programme uses AI to send personalised product recommendations based on purchase history, skin type profile, and browsing behaviour. Sephora’s email personalisation drives industry-leading open and conversion rates.
Starbucks: Starbucks’s AI-powered Deep Brew personalisation engine drives the Starbucks Rewards app, the company’s highest-margin sales channel. Deep Brew personalises offers, recommends new products based on order history, and times offers around individual customer visit patterns. The Starbucks app drives 57 percent of US company-operated revenue and the AI personalisation layer is credited as a primary driver of app engagement and average order value.
How Smaller Brands Apply AI Personalisation
Enterprise AI personalisation platforms are expensive and complex. But the underlying techniques are accessible at smaller scale through existing tools.
Email segmentation and dynamic content: Tools including Klaviyo, Brevo, and ActiveCampaign allow behavioural email segmentation (segment by purchase history, browse behaviour, engagement level) and dynamic content blocks that show different content to different segments within the same email send. This is personalisation at small-team scale for hundreds to thousands of monthly sends.
Website personalisation: Tools including Mutiny, Personyze, and Intelligems enable dynamic website content for different audience segments or individual visitors. A returning visitor who previously viewed enterprise pricing sees the enterprise case study on the homepage. A first-time visitor sees the introductory offer. This requires only tag installation and rule configuration.
Product recommendations: E-commerce platforms including Shopify have AI recommendation apps that surface ‘You might also like’ and ‘Frequently bought with’ based on store purchase data. LimeSpot, Rebuy, and Wiser all provide this without custom ML development.
Timing personalisation: Klaviyo and similar tools allow sending emails at the time each individual user is most likely to open, based on their historical open patterns. This single optimisation consistently improves open rates 10 to 20 percent over fixed send times.
What is AI marketing personalisation and why does it matter?
AI marketing personalisation uses machine learning to deliver individually relevant content, products, and communications to each user based on their behaviour, preferences, and context. Netflix, Amazon, and Spotify attribute a significant share of their engagement and revenue directly to AI personalisation. It increases conversion rates, engagement, and customer lifetime value across every channel where it is applied.
How does Netflix’s recommendation algorithm work?
Netflix’s AI analyses each user’s viewing history, search behaviour, content completion rates, time of day, device, and the behaviour of users with similar patterns. It selects recommended content and even personalises the thumbnail shown for each title based on what has driven engagement for that specific viewer profile. The system drives 80 percent of content discovery on the platform.
How much revenue does personalisation generate for Amazon?
Amazon attributes approximately 35 percent of its total platform revenue to its AI personalisation and recommendation engine. This includes ‘Customers who bought this also bought’, personalised search results, and triggered email campaigns featuring products correlated with individual browsing and purchase history.
Can small businesses use AI personalisation?
Yes. Email behavioural segmentation with dynamic content is available through Klaviyo, ActiveCampaign, and Brevo. E-commerce product recommendations are available through Shopify apps. Website personalisation is available through Mutiny and Personyze. Send-time optimisation improves open rates 10 to 20 percent without custom development. The enterprise approach uses the same principles at different scale and complexity.
What is dynamic content personalisation in marketing?
Dynamic content personalisation shows different content to different users within the same communication or webpage based on their segment, behaviour, or profile. An email with dynamic content shows product recommendations specific to each recipient’s purchase history rather than the same products to everyone. The email is one campaign but produces individually tailored experiences.
How does Spotify’s Discover Weekly work?
Discover Weekly analyses each user’s listening history and identifies their taste profile. It matches that profile against clusters of users with similar tastes and selects songs from those clusters that the user has not yet heard. The result is a weekly 30-song playlist that feels personally curated. The system generates all 400 million-plus individual playlists without human editorial input.
The Principle Is Consistent Across Scale
Netflix spends hundreds of millions on its recommendation infrastructure. A Shopify store using Klaviyo for behavioural email segmentation is implementing the same principle: understand what each user has done, infer what they want next, and serve it at the right moment. The scale differs. The competitive advantage from doing it well is consistent: better engagement, higher conversion, stronger retention.