Key metrics

• Number of couples formed (self reported, as of 4 October, 2021): >= 15
• Number of users: 11623 (116.23% to 10k)

This what we optimize for (see explanation in the blogpost). Please reach out to me if you are a couple who met through the site.

General metrics

• Weddings (self reported, as of 4 October, 2021): 0
• Verified users: 11623
  • Female: 3784 (32.6%)
  • Male: 7784 (67.0%)
  • Non-binary: 55 (0.5%)
  • With photos: 6148 (52.9%)
    • 3275 always unblurred (28.2% of all users)
  • With "empty" profiles : 1721 (14.8%)
  • By relationship type:
    • Friendships only: 1042
    • Romantic only: 5242
    • Both: 4836
  • By FIRE type:
    • "lean": 1324
    • regular: 5857
    • "fat": 1405
  • By activity (when visited site last):
    • < 1 day: 124 (1.1%)
    • < 7 days: 369 (3.2%)
    • < 28 days: 841 (7.2%)
    • < 90 days: 1645 (14.2%)
  • By v2 design enrollment:
    • beta: 208
    • bleeding edge: 68
  • By registration:
    • < 1 day: 1
    • < 7 days: 14
    • < 28 days: 66
    • 
      
• Profile observations : 615661
  • < 1 day: 199
  • < 7 days: 1304
  • < 28 days: 6335
  • 
• Profile fullness
  • 
  • 
  • 
• Requests :
  • Sent: 97553
    • < 1 day: 28
    • < 7 days: 226
    • < 28 days: 1125
  • Accepted: 28427 (29.1%)
    • 
• Messages:
  • Sent: 307126
    • < 1 day: 65
    • < 7 days: 775
    • < 28 days: 4497
  • Read: 291018 (94.8%)
    • 
• Geography:
  • Top 10 countries:
    • United States - 7022
    • Canada - 900
    • United Kingdom - 497
    • Australia - 402
    • Germany - 360
    • India - 243
    • Netherlands - 203
    • Russian Federation - 168
    • Spain - 120
    • Switzerland - 109
  • Top 10 cities:
    • New York, NY, United States - 300
    • Seattle, WA, United States - 297
    • Los Angeles, CA, United States - 249
    • Chicago, IL, United States - 220
    • London, United Kingdom - 199
    • Toronto, Canada - 197
    • Denver, CO, United States - 171
    • Austin, TX, United States - 158
    • Atlanta, GA, United States - 153
    • Boston, MA, United States - 151
• Last hour activity log (max 100 items)
  • [an hour ago] Message sent
• Viktor's time spent (as of 1st of June 2021) : 605 hours
• Money spent (as of 1st of June 2021): 226.65 EUR (~276 USD)
• Volunteers: 6 (as of 23th of June 2021) = 3 designers + 3 React developers.
• Fun: infinity
• Learned: a lot (especially CSS, web design, public communication)
• Interesting people met: a lot, all over the world
• RAM usage
  • 
• CPU usage
  • 
• Hard drive usage
  • 
• Number of non-initialized accounts (these are not included in the counts above)
  • without validated email: 928
  • not attempted test: 557
• All known privacy incidents
  1. February 7, 2024 - Email addresses & usernames of 16 users were each exposed to a single, different user.
     I was in the process of rewriting message notification emails to include more details and make them more personalized, in the hope of receiving more responses. As part of this rewrite, I intended to include the sender's pseudonym in the title. Internally, user data is represented as a DatingProfile object. This object has a helper method "get_possibly_handcrafted_pseudonym" which returns the user pseudonym or age-gender. However, due to an oversight, I wrote "current_user.datingprofile.get_possibly_handcrafted_pseudonym" instead of "current_user.datingprofile.get_possibly_handcrafted_pseudonym()". Instead of calling the method, I referenced the method as it is. This wouldn't have been an issue, but in Python, this leads to a string "a bound method of DatingProfile...". For debugging, the string representation of a dating profile includes the username and email. As a result, both of these were unintentionally included in the email title. This bug was active for about an hour, and during that time, 16 such emails were sent out, each exposing the details of one user to another. I quickly noticed and fixed the issue by double-checking the impact of my changes. I deeply regret the unintended exposure this bug caused. I've informed all affected users, both the ones whose details were exposed and their recipients.
     I've mitigated this from happening in the future. I've removed all the user-generated information from the string representations. I will test such changes better in the future.