3 min.
Is it possible to optimize your Tinder matches?
1L’art de la gestion de projet2Un projet à succès commence par une bonne gouvernance3Cascade, agilité, demandes de changement?

Is it possible to optimize your Tinder matches?

Social Media Paid Media & SEM

Making a first impression with the future love of your life or just your next date (let's be real!) happens much faster and more frequently today thanks to dating apps such as Tinder. We are far from the dates arranged by our best friend where we anticipated meeting a person who interests us and for whom we made sure to be at the top of our game to make the best possible first impression.


With Tinder, a first impression often comes down to 1 or 2 seconds, and this, between two other first impressions. Singles therefore compete to convince the user to swipe to the right with as much confidence as possible.

As a digital media analyst, I can't help but draw connections between Tinder and the most popular digital advertising platforms .

After all, it's the same principle: we must make sure to present our product (ourselves) as well as possible by managing an online campaign where we target an audience by geolocation (km), by demographics (age, sex) and where you reveal your preferences and specifications, such as your size, through a profile.

The first photo of his profile then becomes the ad copy or the creative that will generate future conversions and with such a prospecting campaign, it is essential to get the user's attention from the first moment. The primary profile photo is then the first point of contact with future suitors and this moment is decisive in the success of his love life on the fiery application.


Last January, during an internal Hack Othon Mike Hall, Tinder's Machine Learning Lead, developed a concept to help singles on the platform optimize their matches. Nine months later, Tinder is launching Smart Photos , or Smart Photos, an option you can activate under your profile pictures that makes sure to show the photo most likely to get a “right swipe” first. 

How it works?  Just like in a real online marketing campaign, it is important to constantly test your ads using the proven technique of A/B testing.. Some ads are more evocative than others according to the user and it is not always the ad that we prefer that turns out to be the most efficient. The same situation applies with your photos on Tinder. While showing off your favorite photo of yourself seems like the most logical thing to do, Dr. Jess Carbino, Tinder's resident sociologist, says we're often in a terrible position to figure out which photo sells us best. Singles on the platform choose their first photo for all sorts of personal reasons, but they don't consider the impact of that choice on other users. Carbino claims that a profile gets rejected more frequently if in the first photo:

  • You are not smiling;
  • A portion of your face is hidden (even a very small portion);
  • You are in a group photo;
  • You wear a hat;
  • You wear glasses (all styles combined).

In order to remedy this problem, Hall and his team developed an algorithm that would show your photos in rotation in the first position each time your profile obtains an impression, and thus determine which photo causes the most “right swipes”. By calculating your "Swipe Right Rate" (SRR) on each photo, i.e. the number of "right swipes" out of the total number of "swipes", the algorithm is able to determine which photo to display first to allow you to optimize your matches as much as possible . Or your conversion rate. According to the company, some users have seen their matches increase by up to 12% after enabling the Smart Photos option.   

The SRR is therefore the equivalent of a "click through rate" (CTR), the most decisive metric in order to obtain a good Quality Score during a campaign on Adwords.  The higher the Quality Score, the more likely you are to appear first on the search engine, which allows you to have better visibility and therefore, consequently, more sales or conversions depending on your objectives.

It goes even further, because unlike a search engine, Tinder is a two-way platform, in the sense that, to get a match, a "right swipe" must be made on both sides. Which means that the most optimal photo varies according to the impression, depending on whether your suitors are looking for the corpo side in you with your photo in a suit and tie, or the adventurer that you are on a tropics beach.


The “Swipe Right Rate” is an important metric in your personal score on Tinder which is called the Elo score , a term borrowed from chess which is used to rank the best players in the world. Each app user has their Elo score which is used to rank users and introduce them to people with the same degree of 'desirability'. This score is not shared with users – probably to avoid certain deep depressions -, but it is however possible to improve your Elo score thanks to certain good practices  . An algorithm does not pass judgment, it simply reports the results of your actions through complex mathematical calculations that take into consideration many factors including your photos and your SRR, explains Sean Rad, CEO of the company.

Jonathan Badeen, vice president of product for Tinder makes a comparison with the popular game Warcraft. He explained  during a meeting with Fast Compagny that when he played with a player with a very high score, he ended up earning more points than if he played with a player with a lower score than his. He believes this system is a good way to quickly match and rank users based on who they have matches with. In other words, the more matches you have with people with an Elo score higher than yours, the better your own score will be, and the reverse is also true. Sure, Tinder's algorithm is more nuanced than that, but how others view your profile definitely determines your score. Data analyst Chris Dumler sees the platform as "a vast voting system" that indicates which people are more desirable than others at different levels of the desirability scale.

Thus, to ensure its success on Tinder, as to ensure the success of its online campaigns , it is important to adopt the right behaviors  . Even if user behavior differs according to gender, certain practices such as right swipe everyone, very popular among boys, certainly penalizes the user in terms of the quality of the profiles presented to him, in his scope, and therefore in terms of number of matches as well.


So how do you always present the best photo first when it depends on a multitude of factors at the same time? Mike Hall approached this optimization problem with the Multi-armed Bandit . It was originally a probability distribution problem where players of lottery machines had to decide which machines to play, how many times to play each machine, and in what order to play them in order to maximize the sums. won in a sequence of play. 

In order to address the Multi-armed Bandit problem , Hall and his team implemented Epsilon Greedy's algorithm to find out which of your photos generates the most "right swipes" without wasting impressions on your worst performing photos. . Your photos enter an exploration phase where each photo is tested and then fall into exploitation mode where your best photos are shown first. The team also thought of making some changes to the algorithm and so when you add or remove a photo, it is taken into consideration. 


Whatever one thinks of this application, the success of Tinder is definitely linked to its ease of use with its simplistic design of a ranking of cards. However, behind this intuitive experience lie significant technological challenges. By introducing Smart Photos, Tinder demonstrates that it has taken another step towards a more efficient platform . 

In the end, there is no perfect way to go that will ensure you get more matches. However, if you present honest and recent photos, take the time to "like" only the people you are interested in and use the application frequently, you ensure the best chance of finding the person you are looking for.