The image problem

[powerful_dusk_88623]

Image hosting, and especially hosting NSFW images, is by far the biggest cost and biggest concern for the long-term sustainability of this project.  Below are some thoughts on the image hosting problem, some of which I first posted about on FFA.

The main problem is not the size of each individual upload isn't a problem, although that will have to have some restriction for free users--what I'm looking at right now is a limit of 2MB per image for free users, with the option to pay to have the size cap for individual images increased. The big issue is total uploads per user. Even with something fairly conservative like a lifetime limit of 1GB per user, you're theoretically paying to store that content in perpetuity.

 

A very conservative estimate: say there's 10k users and they each upload on average 7 images a month that are 2MB each. At that rate, a user would fill 1GB of space in five years. For the site, that's about 15GB of image storage space a month, which costs about $1.50. That's not so bad, but then the next month you're storing the first 15GB, plus another new 15GB, so the second month you're paying $3. If you don't add any new users, by month 10 it costs almost $300 a month to host all the previously uploaded content.

 

If (optimistically) about 1% of users get a paid account, image hosting alone outpaces paid account revenue in 10 months if you don't add users. If you keep adding users (and paid accounts at the same rate), it's sustainable for longer, but even then image hosting costs outpace paid account revenue after about three years. The higher the average uploads per user, the sooner hosting costs outpace income.

 

I've put up some extremely sketchy numbers here https://docs.google.com/spreadsheets/d/1qCojrXN2Aa_P-x35F0hOMWoMtloiJrNyQQlg5WXC2pk/edit?usp=sharing if anyone wants to see the factors. This is really, really sketchy because it doesn't take into account anything like administrative costs, payment processing fees, legal costs, fluctuating numbers of paid accounts, additional revenue from paid community accounts or micropayments, or anything else.

NSFW or potentially NSFW images are the biggest price issue. The host I found that is both not-sketchy and friendly to fandom underage/noncon/etc charges 10 cents for a GB of storage, but Amazon charges .0023 cents for a GB of storage. So for $1, I could get 10GB of storage on the friendly host, or almost 450GB of storage on Amazon. With Amazon prices, image hosting would be sustainable pretty much indefinitely (or at least for about 20 years, which I think is basically the same thing in internet years). I'm trying to figure out the technical feasibility of sending NSFW images to the friendly host and SFW images to Amazon, but it's both a technical and moderation headache that would depend on people accurately flagging their NSFW images. If up to 30 percent of images are NSFW, that pricing would be sustainable for about ten years.

One option that might make both the technical and moderation issues simpler is charging a nominal fee (like $1USD) to allow users to upload NSFW images, in addition to allowing full paid accounts to post NSFW.  That would make it easier to offer free users more image hosting space, and make the more expensive NSFW hosting pay for itself.  It would probably mean more moderation of posts by free users that aren't flagged NSFW in an attempt to get around the charge, but it might be worth it for the cost sustainability.  I'm also not sure if this would be a turn off for users that are used to being able to post all their images for free, or if it would be a turn off for users concerned about connecting wallet names to their NSFW fandom images.

Another option might be to allow free users to exchange their allocation of free image space for a certain number of points that can be used for paid features that are less expensive to the site.  For example, if a free user gets 1GB of image hosting space, they could exchange their image hosting space for 10 points, and use those points to get a tag viewer, css customization, personal analytics, or increased number of stored messages in their mailbox, etc.  I haven't run the numbers on this yet, so I'm not sure if it would sufficiently offset the image hosting costs to just have some users opt out of posting images.

The other gigantic issue with images is bandwidth usage, which I don't yet know how to estimate well.