Today I saw an advert at the bottom of a dailywtf.com post. It was for a backup service which charged $3 per GB.

I've had a thing about backup services for while now, I just don't think anyone has got it right. There was box.net - they got really close I think - I bought a year's subscription. The problem is, I've never actually used it! I've tried their Windows sync app, but it didn't work. I've tried to upload via the java widget but it's too much effort. It's also not checking for differences and I'm really not prepared to upload 2.8GB every night. It also lacks version saving. It's a shame because they've got some cool ideas and I like what they've been doing with what is, traditionally, a painful task.

Data Deposit Box are the closest yet. Their site isn't incredible, but it certainly does the job (nearly validates!), and the web app interface is functionally sound, but could do with a nicer design. They have almost everything you'd want:

• Synchronizing app that sits in the tray
• All backup services on background threads/separate processes - only working when machine is idle.
• Auto detects folders to back up, I only had to uncheck one folder - "My Downloads".
• Version saving with ability to adjust quantity of versions saved. Only charges for differences in files (as you'd expect).
• Continuous backup.
• Monthly bill rather than fixed tariff, so if you run out of space it won't stop backups.
• Spread single account across multiple machines.

What it's missing:

• Not suitable for servers - needs to be moved to a Windows service running under "system", the desktop application should be a control panel only.
• API
• Ability to share files. (They have "Shared Folder" in the web app.)
  • This is always a good marketing technique if nothing more.
  • Useful for blog posts etc - although this may have an impact on bandwidth.
• Drop-send like functionality (quick drag&drop upload and download of files/folders) url to share/email after upload etc.
• No free account, 30 day trials are OK, but a free GB would be much more appreciated I think.

A Possible Solution - Amazon S3 Powered Backup Service

• I don't see the harm in requesting payment info incase they exceed the free GB - warn them before billing of course.
• At the moment the service charges a lot for the first GB. Why? I'd say it's better to promote that first use, then have suitable tariffs something like these that include a certain amount of free space, eg:
  • Free Plan
    • $0pcm
    • 1GB included free per month
    • $10 per GB thereafter
  • Basic
    • $5pcm
    • 2GB included free per month
    • $5 per GB thereafter
  • Professional
    • $20pcm
    • 10GB included free per month
    • $4 per GB thereafter
  • Enterprise
    • $50pcm
    • 30GB included free per month
    • $3 per GB thereafter
• Now, who wouldn't upgrade to Basic? Double your space and halve the cost of additional storage for $5 per month! Anyone who got near the 1GB limit would surely give serious thought to upgrading?
• Remember that Amazon S3 (a potential storage method/host for any backup service) charge $0.15 per GB stored and $0.20 per GB transferred. So, using S3, your Basic plan user could upload their entire 1GB, then delete it all and upload it again next month, every month and it would cost you $0.70 per month. You'd be charging $5. So, slightly more realistic figures (1/20th bandwidth, 80% capacity (ie, your users, on average upload 5% of their total data each month and are currently using 80% of their storage space)):
• Free Plan
• $0.13 loss per month
• Basic
• $4.74 profit per month
• 1823% margin
• Professional
• $18.70 profit per month
• 1438% margin
• Enterprise
• $46.10 profit per month
• 1182% margin
• Now assume you have 3/100 of your users upgrade to the Basic plan, 0.5/100 upgrade to the Professional and 0.1/100 upgrade to the Enterprise. Pretty pessimistic, but Ryan Carson suggests the dropsend's upgrade rate is around 3%. Assuming you have 10,000 users:
• Free Plan
• 9,640 users on this plan
• $1,253 loss
• Basic
• 300 users
• $1,422 profit
• Professional
• 50 users
• $935 profit
• Enterprise
• 10 users
• $461 profit
• So 10,000 users attracts $1,565 before other expenses. It's looking a lot worse than I'd originally expected. Although this doesn't take into account additional usage above the quantity included with the plan. I think one of these variables would need adjusting to secure a profitable venture:
• The quantity of free space offered on the free account. (If it became 512MB the free users would cost only $626.50).
• The price of the plans.

Perhaps Amazon S3 isn't a suitable solution for a storage provider for a backup service. But I think it deserves further investigation.