logo

drewdevault.com

[mirror] blog and personal website of Drew DeVault git clone https://hacktivis.me/git/mirror/drewdevault.com.git

The-profitability-of-online-services.md (6003B)

    

  1. ---
  2. date: 2014-10-10
  3. # vim: tw=80
  4. layout: post
  5. title: On the profitability of image hosting websites
  6. tags: [money, linkrot]
  7. ---
  9. I've been doing a lot of thought about whether or not it's even possible to both
  10. run a simple website *and* turn a profit from it *and* maintain a high quality
  11. of service. In particular, I'm thinking about image hosts, considering that I
  12. run one (a rather unprofitable one, too), but I would
  13. think that my thoughts on this matter apply to more kinds of websites. That
  14. being said, I'll just talk about media hosting because that's where I have
  15. tangible expertise.
  17. I think that all image hosts suffer from the same sad pattern of eventual
  18. failure. That pattern is:
  20. 1. Create a great image hosting website (you should stop here)
  21. 2. Decide to monetize it
  22. 3. Add advertising
  23. 4. Stop allowing hotlinking
  24. 5. Add more advertising
  25. 6. Add social tools like comments, voting - attempt build a community to look at
  26.    your ads
  28. Monetization is a poison. You start realizing that you wrote a shitty website in
  29. PHP on shared hosting and it can't handle the traffic. You spend more money on
  30. it and realize you don't like spending your money on it, so you decide to
  31. monetize, and now the poison has got you. There's an extremely fine line to walk
  32. with monetization. You start wanting to make enough money to support your
  33. servers, but then you think to yourself "well, I worked hard for this, maybe I
  34. should make a living from it!" This introduces several problems.
  36. First of all, you made an image hosting website. It's already perfect. Almost
  37. anything you can think of adding will only make it worse. If you suddenly decide
  38. that you need to spend more time on it to justify taking money from it, then you
  39. have a lot of time to get things wrong. You eventually run out of the good
  40. features and start implementing the bad ones.
  42. More importantly, though, you realize that you should be making *more* money.
  43. Maybe you can turn this into a nice job working on your own website! And that
  44. means you should start a business and assign yourself a salary and start making
  45. a profit and hire new people. The money has to come from somewhere. So you make
  46. even more compromises. Eventually, people stop using your service. People start
  47. to *detest* your service. It can get so bad that people will refuse to click on
  48. any link that leads to your website. Your users will be harassed for continuing
  49. to use your site. **You fail, and everyone hates you.**
  51. This trend is observable with PhotoBucket, ImageShack, TinyPic, the list goes
  52. on. The conclusion I've drawn from this is that **it is impossible to run a
  53. profitable image hosting service without sacrificing what makes your service
  54. worthwhile**. We have arrived at a troubling place with the case of Imgur,
  55. however.  MrGrim (the creator of Imgur) also identified this trend and decided
  56. to put a stop to it by building a simple image hosting service for Reddit. It
  57. had great intentions, check out the old archive.org mirror of it[^1].  With
  58. these great intentions and a great service, Imgur rose to become the 46th most
  59. popular website globally[^2], and 18th in the United States alone, on the
  60. shoulders of Reddit, which now ranks 47th. I'm going to expand upon this here,
  61. particularly with respect to Reddit, but I included the ranks here to dissuade
  62. anyone who says "there's more than Reddit out there" in response to this post.
  63. Reddit is a *huge* deal.
  65. Other image hosts died down when people recognized their problems. Imgur has
  66. reached a critical mass where that will not happen. 20% of all new Reddit posts
  67. are Imgur, and most users just don't know better than to use anything else. That
  68. being said, Imgur shows the signs of the image hosting poison. They stopped
  69. being an image hosting website and became their own community. They added
  70. advertising, which is fine on its own, but then they started redirecting direct
  71. links[^3] to pages with ads. And still, their userbase is just as strong,
  72. despite better alternatives appearing.
  74. I'm not sure what to do about Imgur. I don't like that they've won the mindshare
  75. with a staggering margin. I do know that I've tried to make my own service
  76. immune to the image hosting poison. We run it incredibly lean - we handle over
  77. 10 million HTTP requests per day on a single server that also does transcoding
  78. and storage for $200 per month. We get about $20-$30 in monthly revenue from our
  79. Project Wonderful[^4] ads, and a handful of donations that usually amount to
  80. less than $20. Fortunately, $150ish isn't a hard number to pay out of our own
  81. pockets every month, and we've made a damn good website that's extremely
  82. scalable to keep our costs low. We haven't taken seed money, and we're not
  83. really the sort to fix problems by throwing more money at it. We also won't be
  84. hiring any paid staff any time soon, so our costs are pretty much constant. On
  85. top of that, if we do fall victim to the image hosting poison, 100% of our code
  86. is open source, so the next service can skip R&D and start being awesome
  87. immediately. Even with all of that, though, all I can think of doing is sticking
  88. around until people realize that Imgur really does suck.
  90. *2017-03-07 update*
  92. * mediacru.sh shut down (out of money)
  93. * pomf.se shut down (out of money)
  94. * minus.com shut down after going down the decline described in this post
  96. I have started a private service called [sr.ht](https://sr.ht), which I aim to
  97. use to fix the problem by only letting my friends and I use it. It has
  98. controlled growth and won't get too big and too expensive. It's on Github if you
  99. want to [use it](https://github.com/SirCmpwn/sr.ht).
  101. [^1]: [Original Imgur home page](https://web.archive.org/web/20090225014924/http://imgur.com/)
  102. [^2]: [Imgur on Alexa](http://www.alexa.com/siteinfo/imgur.com)
  103. [^3]: [Imgur redirects "direct" links based on referrals](https://dillpickle.github.io/imgur-please-dont-be-the-next-tinypic-or-imageshack.html)
  104. [^4]: [Project Wonderful, an advertising service that doesn't suck](https://www.projectwonderful.com/)