1. 15

  2. 8

    I’m a Tarsnap user and while I agree that Colin should setup some kind of recurring, automated billing (which he said was on his todo list), I have to disagree with a lot of the other points.

    For starters, the title of the article is “what I would do if I ran Tarsnap” but this guy isn’t running Tarsnap, didn’t write all that code, and isn’t standing behind the service which is what makes me want to use it. Colin is, and I’d rather he be more concerned with making things run properly and secure than how to squeeze every dollar out of the service. It’s why many local small businesses stay local and small, and don’t have the desire to sell out to larger companies or go national. Many customers are only patronizing that business because it’s running the way it is. They get good service, they have a rapport with the owner, they don’t like having to wait behind 100 other customers, &c.

    Even though the metered pricing seems silly with its picodollars, I like that I’m only billed for the space and transfer that I use. I just checked my billing records and it’s only costing me about $1/day to backup and store about 40 daily backups for all of my servers. Even if the account was “$30/month” for the same amount of space I’m using, I like that I’m in control of the cost. I can delete old backups to free up space and spend less, or I can splurge and throw a bunch of data at it for a one-off need.

    While the website isn’t the prettiest thing around, it’s usable, it gets the job done, and it’s unique. The new website design proposed looks generic. I don’t think anyone that is in the market for using Tarsnap is going to be won over by a shiny website, nor lost because the website is sparsely designed.

    I like that Tarsnap operates like a Unix tool. It is small, reliable, and limited in scope. It doesn’t have all the bells and whistles that this guy wants it to, because doing all of that means Colin has to write it (securely), run it, support it, and maintain it.

    While there are certainly possibilities for lowering the barrier to entry for Tarsnap, there’s something to be gained by not doing it. When you have more novice users, you have to do a lot more support (unless you’re Google) which takes a lot more time. Now you’re making the same amount of money but working more. You could raise the cost of the service to cope, but now you’re making your more experienced users suffer who probably never even need support, so they might leave. Now you’ve just replaced your good, quiet customers with needy, loud ones.

    1. 6

      Hypothetically, you’d still be able to use “Tarsnap Basic” and continue paying picodollars. It just wouldn’t be a headline feature on the website, but I think that’s fair. The G2G customers would all know about it. I don’t think he’s advocated for substantially changing the tool. It’s still going to be tar, which the people using it will know, but the web site doesn’t need to mention that fact to the boss who signs the invoices.

      FogCreek had some similar encounters, fwiw. The overhead of purchasing user licenses one by one made it impossible for some groups to buy the product. They couldn’t just buy 25 individual licenses in advance because the purchasing department doesn’t approve 25x of something for 17 people. We added a small group license which was good for up to 25 users, and suddenly that made everybody happy. 1x small group license gets approved instantly, and lets them grow from 17 to 18 to 19 people without hassle.

      1. 1

        I don’t think he’s advocated for substantially changing the tool. It’s still going to be tar,

        I was referencing this at the bottom of the post:

        What Else Could We Change About Tarsnap?



        Tarsnap doesn’t currently do the whole backup puzzle. It doesn’t have monitoring, it doesn’t have convenient ways to restore, etc. Tarsnap could easily create more value for users by filling those sub-needs within backups and could potentially even consider branching out some day.

      2. 1

        One of the main points in that article was (I think), “What if something happens to Colin?” That’s really the elephant in the corner.

        1. 3

          The same thing that happens if any one-person business has its owner die. The customers leave. If Colin isn’t running Tarsnap (because he died or because he sold it to a company specializing in selling Bingo Cards) then I probably don’t want to use the service anymore because it wouldn’t operate the same way.

          1. 1

            Agreed. But some other people (not you) might care if their backups become unavilable because something happens to Colin. These other people might not find it so easy to just leave, because their businesses depend on the backup service.

            And I think that’s the point of the article: The one-man-startup backup company may not work for everyone, and if that OP were running the company, he’d go after those other types of customers. As you point out, the OP isn’t running the company. But my other takeaway from the article is that there may be room for another service with equal quality (if that’s even possible), that addresses bigger/more-profitable markets.

            1. 2

              Sure, but then that is another service and no longer Tarsnap. I think it’s presumptuous to tell someone how they should run their (successful) business just because one thinks they could be squeezing more money out if it (although I know the author is on good terms with Colin and Colin probably isn’t offended). It’s like saying you could have raised someone’s child better than them because yours grew up to be a Congressman and theirs only a teacher.

              While I would certainly be disappointed if something happened to Colin and/or Tarsnap, I’m somewhat confident that the service would not instantly shut off the moment Colin got hit by a bus. Colin can also make plans to hand over the keys to someone in the event of his passing or other problem, which would not have to affect his day-to-day running of his business the way he wants to. He can put the keys/passwords in escrow, pass them to someone when he dies, that other person makes an announcement to all Tarsnap customers that they have 30 days or whatever to get their backups, and then shut the service down.

              I guess what I’m saying is that, as a customer, I would much rather use Tarsnap the way it is now and deal with the inconvenience of having to move my backups later, than have to have to use a different/worse Tarsnap now just for the comfort of knowing I won’t have to move my backups later.

              1. 1

                I don’t think it’s presumptuous if the business owner is the one asking you for ideas on how to run the business. And his advice doesn’t appear to be mainly about squeezing pennies, it’s about improving the experience for the customers. As the article mentions, pricing based on picodollars/byte means that you can’t tell how much you’re going to be spending on the service unless you either a) start using the service or b) do a bunch of complicated modeling of the storage you will be using. Having a tiered pricing scheme based on how much storage you reserve makes a lot more sense to people, even if you end up spending more. As long as the old pay-as-you-go system is still kept an option for everyone, adding this tiered pricing scheme can only be a good thing.

      3. 2

        Reference for those who want it.

        Ref: https://twitter.com/patio11/status/451262179858083840