The Algorithm - The magic behind dating apps

Every dating app needs a way to get two people together, it's the heart and soul of a dating app, the most important factor, but how each app goes about this is what makes them different. Some are aggressive and hyper specific while others are generalised, high volume and softer in their approach. It's a complex endeavour, but can make or break your chances at finding exactly what you are looking for.

But you should know what they do and why they do it, to really understand what works best for you.

Distinction Number 1 (Your Perfect Match vs Your Minimal Match):

The Perfect Match

In the first major way algorithms are different, you have two main schools of thought, one of which is to gather as much information you can about 2 people and try to find the perfect person for them by matching across all these data points. Normally this comes in the form of a questionnaire of sorts.

Pros: If you have strict ideas, you are more likely to only see people that fit within them.

Cons: Unfortunately, the world isn’t perfect, and the perfect match might not fit within your initial idea of the perfect person, which means you may miss out on that one chance.

ScatterGun Approach

The second approach is the type we’re probably more familiar with, the scatter-gun approach. This is where an app gathers the minimal information that you are likely to rank people on, age, gender & location (sometimes with religion, or culture) and then allows you to very quickly sift through a large number of people that meet your minimal criteria.

Pros: Many, many options

Cons: Depending on how the specifics of the algorithm operate, you may not get seen, or may not see, the people you want. This is important as this approach has many strengths, but can be mis-utilised / abused.

Distinction Number 2 (Gathering Likes - Swipe Farming)

As in today's day and age we seem to be focusing further towards the second category as are all the most popular apps, we’ll be focusing on that. The second distinct issue in a scatter-gun approach is how do you link 2 people together, again of which you can see in 2 distinct formats.

Sit & Wait, like when you are liked (Hinge)

Some apps, Hinge comes to mind here, have the ability to go like users, but then the ability to recall who has liked you and then you can make your decision, (note, many other apps have this feature behind a paywall e.g. Tinder & Bumble).

Pros: You know who likes you and when, leading to a quicker meetup period and interaction

Cons: A culture of waiting. When analysing the types of users, desirability and other factors, this can lead to a culture of those who don’t get likes, being the only ones to swipe and those receiving likes, to sit back and wait for the swipes. So unless your perfect match is someone proactive, yet popular, or you yourself and proactive in putting out swipes. You may potentially just end up waiting forever for the other person to put the first foot forward.

Spam Liking, go through and like everyone

Option 2 is that you are required to go out there and swipe on users to potentially get a like, it means your options go up, because each swipe gives you another chance to be matched, but it also puts you at the whims of the algorithm to ensure you are actually being shown to those that you have swiped on.

Pros: A larger pool of potential matches, as the culture is to swipe a lot, you’ve got a very big pool of people you could match with.

Cons: Really, the cons come down to the algorithm itself, each company has the ability to make it better or worse, for instance showing your account or not showing your account to users you’ve liked, spreading out the potential of a match too far, re-showing accounts that you’ve already disliked. All these factors can play a part.

Distinction Number 3 (Pseudo Data - Desirability, Lookalike, Timings)

Now we get into the nitty gritty. Here, there is no right or wrong answer and all outcomes are dependent on the details. Most algorithms beyond this point are a magic formula of a part of each of these outcomes, that can change your chances of winning.

The following are the factors most likely to contribute to your account being shown to another human being, and what they are:

Desirability: This is the number of swipes for your account vs the number against, how this is used can vary depending on the app.

Lookalike: Similar to marketing, dating apps can try to group you into groups that have perceived similarities, then based on the likes of users in that group, they can try to feed you people you are likely to enjoy too.

Activity: Some apps will take into consideration the time of day you are active, for how long, the day of the week, and more to try find people who are similar, or worse case, if you're not active enough or often, your account will be impacted and not shown at all or minimally to other users.

Distinction Number 4 - The final Distinction, Profitability

Now this final point really could have fit under point 3, however, I felt elaboration may be good here being a touchy subject. The fourth and final point is profitability. While a lot of apps have a good culture around what they want for their users, they are public or have directors and stakeholders that are, at the end of the day, tied to a financial number.

This will influence how an algorithm works when all is said and done, whether big or small. For instance, a user who has their potential matches spread out more (but not too much to the point that they leave you) will see more ads, more ads means more money, more money means better developers, marketing and more people in the app overall.

This distinction can help you to understand why sometimes when you were new to an app, or less active, in a new market, etc. You seem to be abundant in matches and everyone sees you, but as time goes on and you’ve maybe been more active, less picky, you seemed to be weaned off of being shown and instead had to work harder to try to organise that next date! It's simple maths, you’re still around, you're still happy (well, using the app actively at least). So they can afford to get you in front of more ads, less matches, right on that line of happiness before leaving. All for the benefit of the “app”.

What we do at Hookd!

Here, we want to reward our users for going out there and looking for their next big thing (fling). We take a mix of these steps and approaches, and without going into specifics, we can tell you this.

• We use a scattergun approach. We know love is complex, and sometimes fancy questionnaires can go overboard and potentially filter out the one person you are looking for.
• We take an exploratory approach, we want to increase the number of matches you could get by 10 fold, to do that we need users to be out there, putting out their feelers all the time! But… we let you know if a user has liked you and we feed you potential matches, incredibly often. We like to think of it as the best of both worlds.
• We use minimal pseudo data. We know love is a game of chance, and that really, you should have the say, we gather hashtags to help guide you to like minded individuals, but we do this as a gentle nudge as opposed to a sure thing. We believe letting you explore is a better idea.
• We don’t use any gambling strategies. A lot of dating apps, deliberate or not, use information like level of engagement and effectiveness to keep people around, we don’t, everyone’s account is treated indiscriminately, we don’t show people’s accounts more or less based on how often you are on, when or how much you swipe. Everyone has an even and fair shot.

With these points, and the fine-tuning behind the scenes in the speed and effectiveness of the algorithm, we believe we have one of the best, free dating app algorithms out there. Give Hookd a shot to see how we compare, or, give this blog post a share to let your friends on the inside scoop!