What an algorithm does in a dating app
Every dating app has more profiles than you will ever see. So something has to decide who shows up on your screen first, who comes later and who never appears. Everyone calls that something the algorithm. At its core it is an order: a stack, a list or a daily selection.
The order matters more than it looks. Hardly anyone scrolls to the end in one session; most people put the phone down after a while. If you sit far back in the stack, you practically don’t exist for them. That’s why so many people wonder whether they are being ‘hidden’ as soon as the matches slow down.
This guide explains what can be observed from the outside and what the providers themselves describe in broad terms. Nobody outside these companies knows the internal formulas, and neither do we. Where we describe an observation, we call it that, not insider knowledge.
What providers disclose and what they don’t
The big providers usually name the same factors in their help pages: how active you are, which settings you’ve chosen, where you are and how you interact with other profiles. None of them publish weights, thresholds or formulas. That makes sense, because a published formula would be gamed immediately. For you it means one thing: any guide that promises to ‘crack the algorithm’ is guessing.
One stubborn myth is the secret attractiveness score. For years the story went that Tinder ranked profiles using the Elo system from chess, where a like from a sought-after profile counts more than one from a less popular profile. Tinder itself has said it no longer uses that approach. What it calculates instead, the company hasn’t disclosed in detail.
That’s why our methodology doesn’t rate the algorithm itself, but the effects you feel. Whether an app is built for relationships shows in profile depth and matching logic, and that feeds into seriousness. What an app delivers without payment belongs to fair pricing. Both can be checked without looking at the code.
Activity: the factor almost every app rewards
The factor practically every provider mentions is activity. An app wants to show you people who actually reply, because a match with an abandoned profile wastes everyone’s time. Profiles nobody has opened in weeks therefore slide towards the back. That isn’t a punishment, it’s the obvious way to sort.
This gives you a simple rule. Short, regular sessions work better than one long session on a Sunday evening. If you leave the app alone for three weeks and then come back, the stack often feels noticeably quieter. In most cases that’s not because your profile got worse, it’s because of the break.
Activity also means replying. A match nobody ever writes to tells the app that nothing is happening there. Collecting lots of matches while having hardly any conversations mainly costs you time, and it probably doesn’t help your ranking either.
Picky or indiscriminate: what your behaviour changes
Many apps learn from your reactions who to show you next. It’s the simplest form of personalisation and works much like a music app: what you skip comes up less often. If you like every profile, you give the system no information at all. All it learns is that you’ll take anyone.
Liking everyone has a second effect. You collect matches with people you never looked at properly, and then you have to sort them out one by one. Several big apps limit likes on the free tier, partly to slow down exactly this behaviour. A limit is annoying, but it forces a real decision on each profile.
Our advice from testing: only like profiles where you can picture a first conversation. You’ll get fewer matches, but the ones left are more likely to be people you actually want to write to. For most people that’s the better trade.
The launch effect: why new profiles get seen more at first
One pattern shows up again and again in the big apps: in the first days after signing up, matches come in far faster than later. From the outside it looks as if new profiles are first shown to lots of people so the app can learn how they land. Providers don’t confirm this in detail. It’s plausible all the same, because a new profile has no history yet that it could be ranked by.
That leads to a practical rule: finish your profile before you make it visible for the first time. Photos, text and settings should be in place in the first week, not the third. Starting with a half-built profile wastes exactly the days when the most people are looking.
The popular trick of deleting your profile and starting over to get that effect again is a bad idea. Some providers spot these re-registrations and restrict them. More importantly, it doesn’t fix the real problem: if a profile gets little response after launch, the cause is rarely the ranking and usually the photos or the text.
Paid visibility: Boost, Spotlight and the rest
Many big apps sell visibility for a limited time. Tinder calls it Boost, Bumble calls it Spotlight, and the principle is the same: for a while you are placed further forward in other people’s stacks. It’s the only part of the ranking you can buy directly. Which is exactly why it deserves a sober look.
A boost changes how many people see you, not how they react to you. A profile that gets few likes without a boost mostly gets more rejections with one. It makes the most sense at times when lots of people near you are active, and only with a profile that already works.
In our matrix, what an app delivers without payment belongs to fair pricing. Tinder and Badoo each score 2 out of 5 there, Bumble, Hinge and Dua each 3 out of 5. An app that sells reach has a conflict of interest: the less you’re seen without paying, the more likely you are to pay. That doesn’t have to be exploited, but you should be aware of it.
Filters: the part you control
Before an algorithm ranks anything, it filters. Distance, age and the gender you’re looking for decide which profiles are even in play. These settings are the one part of the system that is entirely in your hands. They have a bigger effect than any ranking.
A tight filter is therefore a double-edged tool. Set age to a three-year window and distance to five kilometres, and you get a small stack that runs out fast. Some apps then ask whether they may show you profiles just outside your limits. That isn’t random, it’s an attempt to show you anyone at all.
Some filters cost money. In many apps the basic filters are free, while finer criteria such as height, religion or intentions sit behind a subscription. On embla, radius, age and language are always free; origin, intentions, religion, wish for children and minimum height belong to the finer search that comes with embla+.
embla: a table instead of a stack
embla is an Albanian dating app for iPhone and hasn’t launched in the App Store yet. So we assess the concept and the model, not everyday use. There is no swiping. Every day there is a daily question, and whoever answers sees other people’s answers at a table with eight seats.
When it comes to algorithms, that’s a different approach. Instead of a stack that refills forever, eight people sit at the table, and contact only happens through a specific sentence or a specific photo. With embla+ the table keeps refilling, up to a hundred people a day. The limit is visible rather than hidden inside a formula.
embla has no Boost or Spotlight. The only form of priority is the Wildfire: priority for one day, only within the search radius the other person has set themselves, and with no promise about a number of views. ‘Who likes you’ is free for everyone. In the matrix, embla provisionally scores 5 out of 5 for seriousness and fair pricing and 2 out of 5 for reach, because its pool is still growing.
Verdict: what to take away from all this
Nobody outside the providers knows the formulas, and anyone who claims otherwise usually wants to sell you something. But what can be observed is enough for a clear strategy. Activity counts, liking everyone hurts you more than the app, and the first days with a finished profile are valuable.
Paid visibility only pays off for a profile that already works without it, and only at times when many people are active. Filters are your strongest lever, so set them as tight as necessary and as wide as possible. And if an app keeps feeding you matches that never turn into conversations, that’s more a reason to switch apps than to rebuild your profile.