Amazon: Primed for Matchmaking
Amazon enters the online dating business with Amazon Matched
Disclaimer: This article is my fictional imagination of Amazon’s future.
February 10, 2025
Amazon enters the dating business four days before Valentine’s Day. The company announced this morning the launch of Amazon Matched. While the company has been known for recommending products to people, this is the first time Amazon will be recommending people to people. Amazon Matched uses big data and AI models trained on the company’s massive accumulation of data across all of the business verticals in which the company operates.
The Data Acquisition Conquest
In early 2023 Amazon acquired Hinge for $7.2 billion, a dating app that was “designed to be deleted.” Compared to its competitors, Hinge focused on connecting users for long-term relationships much like Match.com and eHarmony. Unlike the aforementioned, Hinge geared more toward a younger demographic like Tinder and Bumble.
The acquisition bewildered many in the industry and shareholders causing Amazon’s stock to drop a hefty 7% on news of the purchase. Hinge was left to operate their business as usual. Behind the scene, teams of AI researchers and data scientists were tasked to integrate all of Hinge’s data into Amazon’s cross-domain dataset. This was the final piece of the puzzle that enabled the development of Amazon’s fully AI-driven dating service that launched today.
Like most tech giants, Amazon has plenty of data. What sets Amazon’s dataset apart is the spectrum of verticals it owns businesses and operates in. Let’s briefly examine a few of them and what information they allow Amazon’s AI model to predict.
Amazon.com E-commerce
Amazon.com is very much The Everything Store. You’ll find products across almost every category on the e-commerce marketplace. The platform has tracks what you’ve searched, viewed, and bought. Less obvious, the company also tracks the colour, size, price, quality, quantity, frequency, shipping option, brand, category, flavours, health consciousness, and more, of every purchase. Even information about your address location and credit card type are incorporated into Amazon’s dataset. An entire human profile progressively built on each customer’s activities over time. Combining all your usage data and account information, it’s possible to imagine that Amazon knows more about some people than they might explicitly know about themselves. Amazon’s AI model can predict what you might like—or who you might like—based on what the company can infer about your behavioural traits, habits, preferences, values, interests, and personality. To some, this sounds incredible, and others alarming.
Groceries and Supplies
Amazon acquired Whole Foods Market in 2017 for $13.4 billion. This opened the door to a more intimate dataset about their customers: what you eat and supplies you use at home regularly.
Books
Much like what TV and movies you watch can say something about you, the type of books you read can do that too. It makes sense why Amazon acquired Audible in 2008 and Goodreads in 2013. These acquisitions add another layer of information extracted from what you read, how you rate and review books, how often you read, and even the notes you take and save on the Amazon Kindle if you have one, all go toward enhancing your profile holistically.
Multimedia Streaming
In 2014, Amazon acquired Twitch for $970 million in cash. Hosting 91% of all video game streaming, Twitch was the undisputed champion in the space. Like books, Amazon collects data from your streaming preferences and habits on Twitch for games, Prime Video and Fire TV for movies and TV shows, and Amazon Music for music.
Photos
With Amazon Photos, Prime members have unlimited, full-resolution storage for photos, another rich source of metadata that Amazon’s AI models can utilize.
Linguistics
Joining the voice recognition business, Amazon bought Ivona Software in 2013. This kickstarted the groundwork for what became Amazon Alexa a year after the acquisition. Alexa’s prominence spread as it became built into not only Amazon products such as the Echo, but also 3rd party products such as Sonos. The voice recognition and virtual assistant software enabled the collection and analysis of data on users’ voices, speech, and language patterns.
Amazon Matched: No swiping. No setup.
In 2020, a team of creatives parodied the idea of Amazon getting into the dating business by building a website called Amazon Dating. The website entertained the idea of Amazon expanding its e-commerce website to listing and selling people for dates. What Amazon launched today, however, with Amazon Matched is a reimagining of how online dating can work.
The key differentiator of Amazon Matched is its elimination of the time and manual effort required in setting up a profile, swiping, and “filtering” matches by messaging and going on dates. By using Amazon’s breadth of data on users across its businesses, Amazon can auto-generate a profile of you with highly detailed and personal information about the user. This holistic profile of customers is not public and only used by the AI model. Next, Amazon required a dataset on what profiles match and convert to a relationship. That’s where the Amazon’s acquisition of Hinge comes into play. Hinge provided the dataset of what profile matches succeed. That’s the service that Amazon Matched provides. No swiping. No setup. Simply open the app, and start messaging with people already matched for you. All data-driven, and AI realized.
Using a hiring analogy, the user is the hiring manager and Amazon Matched is the trusted executive assistant. The executive assistant, having watched and recorded the manager’s activities across all areas of their life, knows and understands the hiring manager deeply at a fundamental level. More so than the hiring manager knows about themselves. Because of that, they create the perfect job description. The executive assistant has access to data-backed information on what candidate profiles are most likely to match and get the job based on the job description. Equipped with that knowledge, the executive assistant can accurately scan available candidates and pre-screen them for you. The result is a highly filtered shortlist of candidates most likely to succeed that’s provided to the hiring manager to interview.
Amazon Matched is a new mobile app that’s available to all for free, with one caveat. You need to have an Amazon account with “sufficient data” to enjoy the effortless activation with no setup and no swiping. For people who don’t have an Amazon account or haven’t used Amazon services enough for the company to have sufficient data for their AI to supply you with matches, you can still download and use Amazon Matched. You will be required to set up an Amazon account as well as a dating profile that is a copy of Hinge’s profile setup. Users with insufficient data will get a disclaimer that their matches may not be as high quality. To incentivize usage of Amazon services for users with insufficient data, the company is currently offering a free year of Amazon Prime loaded with $100 in account credit.
Once you’re all set in the app, you’ll see a shortlist of matches with whom you can begin messaging immediately. When you click into a match, the only information you’ll see about your match is their profile photo, age, and location. To help users begin interacting, three AI-generated icebreaker prompts are provided based on the profile of you and your match. Along with messaging, users can also communicate with their matches using the built-in video call feature. The feature is a stripped-down version of Amazon Chime, the company’s existing video calling service on AWS. Both the messaging and video calling features are end-to-end encrypted.
Amazon Matched has the same goals as Hinge, where the goal is to match people up for relationships. To progressively improve Amazon’s AI model for matching people that result in relationships, Amazon Matched surveys users in-app to find out if they’ve met their match in person or if they’re now in a relationship based on their app activity. Over time, data from Amazon Matched will also further strengthen the company’s AI model.
What’s Next
The success of Amazon Matched will rest upon its experience and relationship conversion rate compared to other dating apps. The company boasts a 99% match satisfaction rate from their beta testing period. Given the quality and quantity of Amazon’s dataset and their work on AI, I can see that statistic hold true. What remains to be proven though is how accurate the matches are in determining relationship conversions. One might say that’s out of the company’s control. Amazon Matched may do an excellent job in matching people, but whether it forms into a relationship is primarily dependent on the individuals.
If the app succeeds in the dating space, it’s conceivable that Amazon Matched might expand into matching people for friendship and professional networking. Imagine a world of meaningful connections made possible by Amazon. A successful Amazon Matched would also put the existence of Hinge into question as it will no longer serve a business purpose under Amazon’s ownership.
Lastly, the potential opportunities for other applications of Amazon’s people matching AI might be a question worth further research and exploration. How else might Amazon be able to add more value to the world?