$ cd /projects/prtnr.me
prtnr.me — The Dating App That Deserved Better
GitHub: github.com/dabooze/smart-dating
Press: Business Insider Germany (March 2021)
PRE — Idea · Setup · Build
Goal: Build a dating app that matches people on psychological compatibility instead of photos and proximity. Use actual attachment theory — the science of how people bond (or fail to) in relationships — as the primary matching signal. Make Tinder look like the dopamine slot machine it is. I've been online dating since the early 2000s. When Parship launched with its scientific, data-driven approach, I was all ears. I'm data-oriented. I genuinely believe you need just enough of the right information to make good decisions — unless you let your emotions or your base instincts talk. Which is exactly what most dating apps are designed to exploit. I even freelanced at Parship at one point. Saw the machine from inside. Understood the market. And kept thinking: this can be better. The science exists. Nobody's using it properly. Then I read "Attached" by Amir Levine and Rachel Heller. The book explains attachment theory — originally developed by psychologist John Bowlby — in plain language. There are three primary attachment styles: Secure, Anxious, and Avoidant. And the combinations matter enormously. Here's what it looks like in practice. I researched forum posts on Parship's own community boards and saw the same pattern repeated hundreds of times: "I met this guy. We had this amazing weekend. Sex on Sunday. Breakfast in bed. Then he had to leave. And then I don't hear from him the entire week despite me sending him text messages. And on Friday he texts me as if nothing changed and we had another great weekend." Textbook Anxious-Avoidant combination. Pure poison. The Anxious partner craves closeness and reads silence as rejection. The Avoidant partner needs space and reads contact as suffocation. Neither is broken. They're just fundamentally incompatible in how they bond. And no amount of great sex or shared hobbies will fix that mismatch. Every dating app on the planet ignores this. They match on proximity, photos, shared interests — surface data. I wanted to match on the thing that actually determines whether two people can sustain a relationship. Stack: - Ruby on Rails 7.1 (API backend) - Swift / iOS (native App Store app) - Kotlin / Android (prototype) - PostgreSQL (German servers, encrypted) - Capistrano (deployment) - Eventbrite-style ticketing: none. Free app, €0. - Balsamiq (wireframes) - Fastlane (iOS build automation)
The Science — 36 Questions That Actually Matter
The core of prtnr.me is a 36-question psychological test that determines your attachment style. This isn't a BuzzFeed quiz. It's based on the ECR-R (Experience in Close Relationships - Revised), a validated clinical instrument developed at university level for measuring attachment anxiety and avoidance. I took the original research paper, studied the scoring model, and transformed it into a digital screening tool. Every new user takes the test before they can browse profiles. No test, no matches. Non-negotiable. The test places you on two axes: anxiety (fear of abandonment) and avoidance (discomfort with closeness). The combination yields your attachment type — Secure, Anxious, or Avoidant — plus a weight indicating how strongly you express that style. The matching algorithm then uses this as the dominant signal: Psychology (attachment compatibility): 70% Shared interests (media, music, sport): 10% Diet compatibility: 10% Kids compatibility: 10% 70% psychology. Not 70% "likes the same movies." Not 70% "lives within 20km." Seventy percent of your match score is determined by whether your attachment styles are compatible. The scoring matrix: Secure + Secure = high compatibility Secure + Anxious = moderate (Secure can stabilize Anxious) Secure + Avoidant = moderate-low (harder to bridge) Anxious + Anxious = moderate (mutual understanding, but...) Anxious + Avoidant = low (the toxic dance described above) Avoidant + Avoidant = variable (depends on weight) Each pairing is further modulated by how strongly each person expresses their type. A mildly Avoidant person paired with a Secure partner is very different from a strongly Avoidant person paired with an Anxious one.
The Stealth Features — Hidden Doors
prtnr.me had two features that never appeared in any marketing material. Hidden doors for invisible minorities. 1. MENSA MATCHING Verified members of Mensa (the high-IQ society) could activate a hidden flag on their profile. This let them find partners "on a similar cognitive level" without the app looking like an elitist nerds-only platform. The feature existed because gifted people genuinely struggle to find compatible partners — not out of arrogance, but because cognitive compatibility matters in long-term relationships just as much as emotional compatibility. 2. NEURODIVERGENT MATCHING Added in 2025: a similar stealth flag for neurodivergent users. Same idea — dating based on psychological compatibility, shared values, child preference, diet. The real stuff that makes or breaks relationships. No need to advertise it. The people who need it know they need it. The philosophy behind both: don't stigmatize, don't advertise, just quietly provide the door. Those who need it will find it.
The Product — What Made It Different
Beyond the psychology engine, prtnr.me had deliberate design
choices that went against every growth-hacking instinct in the
dating app playbook:
- No swiping. Users actively like, dislike, or skip profiles.
Swiping is a dopamine mechanic designed for engagement, not
for finding partners.
- 99-hour chat expiration. If nobody talks for 99 hours, the
conversation dies. Anti-ghosting by design. Either engage or
move on. No zombie matches cluttering your inbox for months.
- No tracking. German servers. Encrypted data. In 2025, I went
further and encrypted every personal field at the database
level — names, emails, phone numbers, cities, professions.
Because dating data is the most intimate data there is.
- Daily mottos. Small touches of personality.
- Complaint system. Because accountability matters.
A hybrid of Parship's depth, Bumble's respect, and OkCupid's
openness. Without their engagement-maximizing monetization.
Screenshots — App Store 2021
POST — Learnings · Afterthoughts · Timeline
What happened:
Three launches. Three deaths. Five years. Same lesson every time.
The first launch (January 2021) had momentum. Business Insider
Germany covered it. Mensa Germany promoted it through their
magazine. 350 users signed up. I saw conversations happening on
the platform. People were using it. It worked.
But 350 users spread across all of Germany is nothing. The
chicken-and-egg problem is merciless: nobody joins a dating app
with no users, and you can't get users without users. People
messaged me: "The app is nice but there are no people in my
area." The same feedback, over and over.
I tried to find partners. Sales people. Offered equity. Everyone
loved the idea. Everyone lost motivation within weeks. I don't
blame them — selling a dating app against Tinder's marketing
machine is a thankless job.
Shutdown in 2023. Silence.
Then 2025. Cyprus. I thought: it's a great algorithm, the world
should use it. Especially now, when everyone complains about
Tinder's enshittification. Full relaunch. Bilingual — I
translated every line of text into English myself. Database-level
encryption. The neurodivergent flag. Better tech across the
board. Back in the App Store.
Approached Mensa Germany again. This time they asked ME to write
the article. I did. Then I met with local business contacts who
made it crystal clear: "Your app is un-investable in its current
state. You need to find a proper CEO before anyone can invest."
They weren't wrong. I am not a CEO. I am a creator, an inventor.
Not a sales guy. I saw enough CEO types during my career to know: creator and
operator are different jobs. They serve their purpose. I serve mine.
Final shutdown. For real this time.
The thing that stings isn't the failure. It's that the science is
right. I'm 100% convinced that attachment style mismatch is the
#1 reason relationships fail. I've seen it in 20+ years of dating.
I've seen it in hundreds of Parship forum posts. I've seen it in
the research. The algorithm works. The product works. The
marketing doesn't. And in this industry, marketing is everything.
I learned this exact lesson 12 years earlier with bedster.de —
my hotel booking startup from 2009. Best product doesn't win.
Best marketing wins. And here I am, having learned it twice.
The commercial failure didn't sting because the product problem
was the interesting part. The science is real. Whether money
follows was always secondary to whether the idea was right.
Learnings:
- Attachment theory is real and predictive. The Anxious-Avoidant
trap destroys relationships with mechanical precision. A dating
app that screens for this would save millions of people years
of pain. Nobody funded it.
- The chicken-and-egg problem in marketplaces is a killer. You
cannot solve it with a better product. You solve it with
money, marketing, or a miracle. I had none of the three.
- Same lesson as bedster.de (2009): the best product doesn't
win. Best distribution wins. I've now learned this twice
across 12 years and two industries. Maybe it'll stick.
- Building for invisible minorities (gifted, neurodivergent)
requires stealth. Don't advertise it. Don't stigmatize it.
Just provide the door. The people who need it will find it.
- Creator ≠ CEO. Knowing which one you are is more
valuable than pretending to be the other.
- Teaching myself Swift and building a native iOS app from
scratch was genuinely satisfying. Hated the App Store
process, the localization, the marketing. Loved the building.
- With 10x the users, this would have been a great platform.
But "would have been" doesn't ship.
Timeline:
- 2020-06: NADA ("Not Another Dating App") — the first spark.
One SwiftUI commit. Two placeholder tabs. The idea existed
entirely in my head.
- 2020-09: First Rails migration. SmartDating is born. Three
months of intense building: Rails API, iOS app in Swift,
matching algorithm, the 36-question attachment test.
- 2021-01: Launch. App Store. Free. Let's go.
- 2021-03: Business Insider Germany covers prtnr.me. Mensa
promotes it. 350 users sign up. Conversations happen.
- 2021-2022: The slow bleed. Not enough users. "Nice app but
nobody in my area." Equity partners lose motivation. The
chicken starves. The egg never hatches.
- 2023: Shutdown. Silence.
- 2025: Relaunch from Cyprus. Bilingual. Encrypted. ND flag.
Better everything. Mensa article (this time I write it).
Business contacts: "un-investable without a CEO." Understood.
- 2025-07: Final migration. Final shutdown. 108 migrations over
5 years. The algorithm lives in a repo. The users don't.
Status: Complete. The science is sound. The product worked. The
market didn't care. The algorithm lives in a repo. The question
was answered.