HackerFriction Feedback

The biggest usability problems on Hacker News
in the last 5 years

Try iPhone · Jira · Instagram · parking

What’s this?Who made it, how it’s ranked, and where to get the data
What am I looking at?

A ranked list of the specific usability problems people keep raising in Hacker News comments, such as “rejecting cookies takes more clicks than accepting”. Each one shows how many different people raised it and links to their comments. Search any product to see its own list. Inspired by Andrej Karpathy’s HackerTube, which ranks the videos Hacker News recommends.

Whodunnit?

Adnan Khan, founder of Justusability, which helps teams see where their customers get stuck. Claude read and grouped the comments; Adnan reviews the issues on these lists by hand. Not affiliated with Hacker News or Y Combinator.

Why?

Other projects built on Hacker News rank what it recommends: videos, books, tools. This one ranks what it keeps complaining about, because a problem that many people raise for years is one a design team can fix. Each reviewed issue names the usability principle behind it and suggests a fix.

Can I have the data?

Yes. Download every issue and the comments behind it as JSON or CSV. No cookies, no analytics, and search runs in your browser.

How is it ranked?

By how many different people raised each issue in the last 5 years. An issue is listed once at least 3 people have raised it, in at least 3 different months and 3 different threads. To rank in a period it also needs 3 people and 3 threads within that period. No thread counts for more than a third of an issue’s people, so one launch thread can’t carry it to the top. How we count

What does “Reviewed” mean?

Justusability checked that it’s a real usability problem, named the principle behind it and suggested a fix. Unreviewed issues, further down product pages, are grouped and counted by machine: you get the counts and comment links, but no quotes, principle or fix.

Is this what all users think?

No. Hacker News skews towards developers and technical people, and towards the US, so treat these as problems people report, not a measure of all users. The counts are a floor too: Claude has read only the part of Hacker News most likely to hold complaints. Method and limits

Something wrong?

Use the “Something wrong here?” link on the issue, or open a public correction report on GitHub.

How do I give feedback?

Send us feedback: email us, or leave a public note on GitHub. Ideas, products we should cover and things we got wrong are all welcome.

  1. 1

    Can't operate controls by feel, must look away from roadReviewed

    Car touchscreens · 621 people in the last 5 years · raised since 2010

    2009: 0 people2010: 1 person2011: 2 people2012: 13 people2013: 5 people2014: 30 people2015: 15 people2016: 15 people2017: 20 people2018: 51 people2019: 176 people2020: 102 people2021: 99 people2022: 176 people2023: 221 people2024: 108 people2025: 70 people2026: 71 people
    20092026
  2. 2

    Rejecting cookies takes more clicks than accepting themReviewed

    Cookie banners · 343 people in the last 5 years · raised since 2009

    2008: 0 people2009: 2 people2010: 0 people2011: 1 person2012: 4 people2013: 0 people2014: 0 people2015: 3 people2016: 4 people2017: 0 people2018: 39 people2019: 42 people2020: 72 people2021: 113 people2022: 110 people2023: 73 people2024: 67 people2025: 46 people2026: 56 people
    20082026
  3. 3

    Consent popups are artificially hard to refuseReviewed

    Cookie banners · 333 people in the last 5 years · raised since 2011

    2008: 0 people2009: 0 people2010: 0 people2011: 1 person2012: 0 people2013: 1 person2014: 1 person2015: 1 person2016: 2 people2017: 5 people2018: 82 people2019: 33 people2020: 98 people2021: 107 people2022: 99 people2023: 78 people2024: 69 people2025: 73 people2026: 34 people
    20082026
  4. 4

    Local account setup is deliberately painful, pushing Microsoft accountsReviewed

    Windows · 226 people in the last 5 years · raised since 2012

    2007: 0 people2008: 0 people2009: 0 people2010: 0 people2011: 0 people2012: 1 person2013: 8 people2014: 2 people2015: 5 people2016: 3 people2017: 7 people2018: 10 people2019: 28 people2020: 24 people2021: 50 people2022: 35 people2023: 63 people2024: 62 people2025: 29 people2026: 44 people
    20072026
  5. 5

    Touchscreen controls must be looked at, unlike findable physical buttonsReviewed

    Tesla cars · 199 people in the last 5 years · raised since 2011

    2009: 0 people2010: 0 people2011: 1 person2012: 5 people2013: 2 people2014: 13 people2015: 10 people2016: 15 people2017: 11 people2018: 29 people2019: 46 people2020: 46 people2021: 55 people2022: 63 people2023: 67 people2024: 34 people2025: 19 people2026: 15 people
    20092026
  6. 6

    Quoted terms no longer force results to contain exact wordsReviewed

    Google Search · 183 people in the last 5 years · raised since 2009

    2007: 0 people2008: 0 people2009: 6 people2010: 8 people2011: 22 people2012: 5 people2013: 10 people2014: 7 people2015: 5 people2016: 15 people2017: 11 people2018: 35 people2019: 50 people2020: 33 people2021: 43 people2022: 68 people2023: 64 people2024: 30 people2025: 18 people2026: 11 people
    20072026
  7. 7

    Mobile site constantly pushes app download, hard to useReviewed

    Reddit · 169 people in the last 5 years · raised since 2017

    2007: 0 people2008: 0 people2009: 0 people2010: 0 people2011: 0 people2012: 0 people2013: 0 people2014: 0 people2015: 0 people2016: 0 people2017: 9 people2018: 28 people2019: 39 people2020: 85 people2021: 100 people2022: 54 people2023: 75 people2024: 14 people2025: 6 people2026: 12 people
    20072026
  8. 8

    Banner covers much of the screen and hides contentReviewed

    Cookie banners · 165 people in the last 5 years · raised since 2012

    2008: 0 people2009: 0 people2010: 0 people2011: 0 people2012: 2 people2013: 5 people2014: 0 people2015: 0 people2016: 4 people2017: 4 people2018: 14 people2019: 21 people2020: 30 people2021: 45 people2022: 47 people2023: 41 people2024: 30 people2025: 33 people2026: 14 people
    20082026
  9. 9

    Slow to interact with, with little guidance on proper useReviewed

    Jira · 148 people in the last 5 years · raised since 2009

    2007: 0 people2008: 0 people2009: 1 person2010: 2 people2011: 4 people2012: 1 person2013: 1 person2014: 4 people2015: 10 people2016: 10 people2017: 14 people2018: 38 people2019: 21 people2020: 73 people2021: 44 people2022: 46 people2023: 37 people2024: 22 people2025: 24 people2026: 26 people
    20072026
  10. 10

    Takes ages to learn Nix with too little documentationReviewed

    Nix · 146 people in the last 5 years · raised since 2013

    2013: 1 person2014: 2 people2015: 2 people2016: 6 people2017: 3 people2018: 5 people2019: 3 people2020: 18 people2021: 29 people2022: 48 people2023: 31 people2024: 36 people2025: 27 people2026: 12 people
    20132026
  11. 11

    Ads play before videos and repeat every minuteReviewed

    YouTube · 145 people in the last 5 years · raised since 2010

    2008: 0 people2009: 0 people2010: 1 person2011: 2 people2012: 2 people2013: 4 people2014: 2 people2015: 5 people2016: 0 people2017: 5 people2018: 5 people2019: 12 people2020: 17 people2021: 12 people2022: 27 people2023: 60 people2024: 22 people2025: 32 people2026: 8 people
    20082026
  12. 12

    Results cluttered with spam pages, hiding useful contentReviewed

    Google Search · 142 people in the last 5 years · raised since 2008

    2007: 0 people2008: 1 person2009: 0 people2010: 5 people2011: 5 people2012: 3 people2013: 2 people2014: 2 people2015: 4 people2016: 5 people2017: 6 people2018: 5 people2019: 9 people2020: 11 people2021: 15 people2022: 40 people2023: 48 people2024: 23 people2025: 18 people2026: 11 people
    20072026
  13. 13

    Results look like ads, making organic results hard to tell apartReviewed

    Google Search · 137 people in the last 5 years · raised since 2008

    2007: 0 people2008: 1 person2009: 2 people2010: 1 person2011: 1 person2012: 5 people2013: 10 people2014: 10 people2015: 2 people2016: 9 people2017: 4 people2018: 10 people2019: 20 people2020: 47 people2021: 20 people2022: 36 people2023: 38 people2024: 30 people2025: 22 people2026: 10 people
    20072026
  14. 14

    Start menu shows adsReviewed

    Windows · 134 people in the last 5 years · raised since 2009

    2007: 0 people2008: 0 people2009: 1 person2010: 0 people2011: 0 people2012: 0 people2013: 0 people2014: 0 people2015: 1 person2016: 3 people2017: 7 people2018: 11 people2019: 10 people2020: 13 people2021: 6 people2022: 20 people2023: 38 people2024: 29 people2025: 31 people2026: 20 people
    20072026
  15. 15

    Settings are scattered across old Control Panel interfacesReviewed

    Windows · 133 people in the last 5 years · raised since 2008

    2007: 0 people2008: 1 person2009: 2 people2010: 1 person2011: 2 people2012: 3 people2013: 0 people2014: 1 person2015: 5 people2016: 13 people2017: 24 people2018: 19 people2019: 37 people2020: 33 people2021: 61 people2022: 32 people2023: 19 people2024: 35 people2025: 29 people2026: 14 people
    20072026
  16. 16

    Banners designed to push users into accepting trackingReviewed

    Cookie banners · 126 people in the last 5 years · raised since 2012

    2008: 0 people2009: 0 people2010: 0 people2011: 0 people2012: 1 person2013: 1 person2014: 0 people2015: 1 person2016: 0 people2017: 0 people2018: 10 people2019: 11 people2020: 16 people2021: 34 people2022: 31 people2023: 24 people2024: 37 people2025: 21 people2026: 17 people
    20082026
  17. 17

    Search returns unrelated videos after a few matches, hiding the wanted oneReviewed

    YouTube · 114 people in the last 5 years · raised since 2009

    2008: 0 people2009: 1 person2010: 1 person2011: 0 people2012: 1 person2013: 0 people2014: 1 person2015: 1 person2016: 2 people2017: 2 people2018: 3 people2019: 7 people2020: 5 people2021: 5 people2022: 32 people2023: 29 people2024: 18 people2025: 17 people2026: 18 people
    20082026
  18. 18

    Captchas are so hard that users fail them repeatedlyReviewed

    CAPTCHAs · 110 people in the last 5 years · raised since 2007

    2007: 2 people2008: 3 people2009: 4 people2010: 3 people2011: 8 people2012: 15 people2013: 6 people2014: 11 people2015: 4 people2016: 13 people2017: 5 people2018: 17 people2019: 28 people2020: 28 people2021: 29 people2022: 20 people2023: 28 people2024: 26 people2025: 24 people2026: 15 people
    20072026
  19. 19

    Recommendations repeat watched videos and push unwanted contentReviewed

    YouTube · 109 people in the last 5 years · raised since 2010

    2008: 0 people2009: 0 people2010: 1 person2011: 1 person2012: 0 people2013: 2 people2014: 1 person2015: 1 person2016: 4 people2017: 5 people2018: 11 people2019: 23 people2020: 25 people2021: 30 people2022: 33 people2023: 37 people2024: 10 people2025: 21 people2026: 7 people
    20082026
  20. 20
    2010: 0 people2011: 0 people2012: 1 person2013: 1 person2014: 4 people2015: 2 people2016: 3 people2017: 4 people2018: 6 people2019: 9 people2020: 7 people2021: 36 people2022: 15 people2023: 23 people2024: 19 people2025: 16 people2026: 14 people
    20102026
  21. 21

    Tweets fail to load, and viewing content forces a loginReviewed

    Twitter / X · 98 people in the last 5 years · raised since 2009

    2008: 0 people2009: 1 person2010: 1 person2011: 2 people2012: 4 people2013: 3 people2014: 4 people2015: 4 people2016: 0 people2017: 4 people2018: 2 people2019: 3 people2020: 1 person2021: 25 people2022: 40 people2023: 19 people2024: 15 people2025: 9 people2026: 10 people
    20082026
  22. 22

    File Explorer takes several seconds to openReviewed

    Windows · 97 people in the last 5 years · raised since 2010

    2007: 0 people2008: 0 people2009: 0 people2010: 1 person2011: 0 people2012: 1 person2013: 1 person2014: 4 people2015: 2 people2016: 4 people2017: 8 people2018: 10 people2019: 10 people2020: 16 people2021: 12 people2022: 11 people2023: 24 people2024: 7 people2025: 31 people2026: 32 people
    20072026
  23. 23

    Reject option hidden or less prominent than acceptReviewed

    Cookie banners · 96 people in the last 5 years · raised since 2016

    2008: 0 people2009: 0 people2010: 0 people2011: 0 people2012: 0 people2013: 0 people2014: 0 people2015: 0 people2016: 1 person2017: 0 people2018: 6 people2019: 7 people2020: 17 people2021: 22 people2022: 20 people2023: 19 people2024: 24 people2025: 18 people2026: 17 people
    20082026
  24. 24

    Cancelling Prime is deliberately hard, keeping people subscribedReviewed

    Amazon Prime · 89 people in the last 5 years · raised since 2014

    2014: 1 person2015: 0 people2016: 1 person2017: 1 person2018: 4 people2019: 6 people2020: 5 people2021: 9 people2022: 31 people2023: 43 people2024: 7 people2025: 9 people2026: 2 people
    20142026
  25. 25

    Autocorrect changes typed words to wrong ones, causing typosReviewed

    iPhone · 87 people in the last 5 years · raised since 2008

    2007: 0 people2008: 1 person2009: 0 people2010: 1 person2011: 2 people2012: 1 person2013: 2 people2014: 0 people2015: 0 people2016: 2 people2017: 1 person2018: 4 people2019: 0 people2020: 8 people2021: 4 people2022: 26 people2023: 19 people2024: 7 people2025: 22 people2026: 21 people
    20072026