Hacker News Reader

Comment by davidw | Hacker News Reader

The Cynic: ask for picodollars and you shall receive picodollars.

Joking aside though, I think that I see something of myself in your marketing. Your web page, for instance, is very plain, which to someone like me is great. But people seem to not like that, for whatever reason. Maybe you should do some A/B testing with a "flashier" (but no flash, please!) version? I would be curious to hear about the results. This is one of those things that irritates me about the world at times: people ought to be looking at your product and its price, rather than fades, drop shadows and the like... but they don't.

Also: he's targeting people used to paying for stuff, whereas lots of Linux and BSD people are, well, kind of stingy. Maybe a bigger push towards Mac would be another idea?

View on HN

I agree.

The tagline is "Backups for the truly paranoid". Yet the site looks like it was built by one person as a hobby. Why would the truly paranoid trust this web site to store his files?

Tarsnap requires the user to be a unix-geek that can compile the program himself and use the command line. But if I'm a truly paranoid hacker, wouldn't I just write a quick python script myself to encrypt and backup my files to S3? Why would I trust you? One thing that may help is that you should let the user upload the files to their own S3 account, not to your S3 account. That would certainly make me more likely to use it.

Also, nowhere on the home page does it tell me what makes Tarsnap so special. To me, it's just another backup program. What's special about it that makes it good for the truly paranoid? Tell me that up front.

From what I can grasp from their page, the paranoid part comes from the encryption used.

And, ironically, I'm not sure which I'm more paranoid about: Having other people read my backups, or accidentally screwing up my encryption protocol and being unable to recover my backups.

My worst nightmare is securely encrypting everything with a single key and than somehow losing that key to a fire, a thief, a tragic laundry incident, or a filing error (oops, I changed keys and accidentally shredded the new one instead of the old one! Oops, I put the paper with the key in this bag but then I forgot about it for a week, and cleaned out the bag, and the piece of paper got picked up by the office cleaning staff!)

I wish I could afford to use nothing but the Linus Torvalds approach to backups: Upload everything to the cloud in plain text and try to convince people to mirror it.

IMHO, the point is not that the site isn't flashy, but that it doesn't explain why would you want off-site and crypted backups in the first time. The site I would expect for a backup tool is one which scares the visitors to the hell, full of well known horror stories, and showing how tarsnap would have avoided them.

For example, Jeff Attwood, a well known and "smart" guy isn't he? Well, he also thought his hosting provider was doing adequate backups...

Astalavista? They had backups, but the cracker logged into the backup servers using the credentials stored in backup scripts and deleted them. Did you know that tarsnap may use a key for writing, another for reading and another for deleting?

It's true that being an UNIX oriented tool makes it a "niche" product since most companies use a combination of UNIX and Windows servers. But there is still a big chance of success among web owners running small to medium webs for a living on cheap VPS or shared hosting accounts on foreign countries. They should get very nervous every time they hear an horror story detailing how some site dissapeared because of a bad backup strategy...

another suggestion is to start shipping binaries. i just looked at your download page and while it's not much of an issue for me to compile it, it may be just enough work to make other people not want to try it. especially when you start talking about header files and dependency versions.

make deb and rpm packages, get it into the freebsd/openbsd/pkgsrc/fink port trees. make your installation page as short as "pkg_add tarsnap" and only link to the compilation instructions for other, less commonly used architectures. (do you have the ability to see what os/arch most of your users are using?)

Binaries are on my to-do list. I could build a bunch of binaries and put them on the website right now, but I want to have a scriptable build process before I commit to providing binaries.

FWIW, Tarsnap is in the FreeBSD ports tree already, but not in others.

Explore dual-licensing options for the daemon, too. Tarsnap is the kind of thing that could become a ubiquitous sysadmin tool, if done right!

I would say this is a big one. I'm capable of compiling it myself, but honestly, for trying something out that's not very high on my to do list (since I have a workable solution of my own), the barrier to trial needs to be very low.

crypted off-site backups? wait, people pay for that? wtf, crashplan is free (forever). all you need is a buddy with some disk. a tb is about a hundred bucks these days.

good luck with that tarsnap thingy.

I think you are being downvoted for being so rude and straight but I think you make the point here. Tarsnap is shooting at the wrong target: everything about it seems to be designed to appeal geeks: the webpage style, the humor (the picodollars thing), the code to be compiled (!). Basically tarsnap is targeting the people that will never pay for this service. I am a geeky enough to recognize myself as possible target and therefore geeky enough to setup the same thing using 10 minutes of bash, truecrypt and S3 for half the prize.

Tarsnap does block-based differentials, in addition to its fancy crypto footwork. You can't set that up with bash, truecrypt, and S3.

Please don't get me wrong, it was not my intention to compare the features of tarsnap to a 10 minutes hack. I am sure there is a lot of good work behind and it looks like great code and a great project. I am also sure that a lot of businesses would be happy to pay for your service if they were directly targeted for it.

But I have no doubt the reason while your revenue doesn't make you happy yet it is because hackers are not the right target for it. I do differential backup using encFS and dropbox; it's not even remotely as cool as tarsnap but it's free for now. You see my point?

if your "backup" is on a single disk, it's not a backup

I think you're right on both counts. I have a friend who is trying to find a way to make the Tarsnap website look nicer without losing its essential geeky/minimalist character; if/when he comes up with something I'll definitely be doing some testing to see how it compares. That said, I don't think Tarsnap's conversion rate is particularly bad -- I'd say that most important issue is simply getting people to the Tarsnap website in the first place.

Promoting Tarsnap as a solution for OS X runs into the GUI problem: Most OS X users don't have a clue what to do with a command-line interface. Tarsnap will gain a GUI at some point in the future, but of course this is a rather non-trivial amount of code to write...

Have you tried Reddit advertising?

Not yet. I'm thinking about it.

EDIT: Ok, finished thinking. Let's see how much traffic $20 gets me.

I'd love to hear about how this turns out for you.

I'll probably post to my blog with the statistics and comments about the reddit approach to advertising. I don't think much of the "buy a fraction of a day's traffic, but we're not going to tell you what fraction you're getting" approach, for one thing...

Experiment day is Thursday, so I'll probably have useful numbers early next week.

Link that to HN. It will be very interesting if you put up a thorough analysis of how advertising on reddit worked out.

Assuming I write such a post, I'll submit it to both HN and Reddit. (And given how these things tend to work, I'll probably get more traffic from Reddit going to the blog post about advertising on Reddit than I get from the advertising itself...)

Good work. For cheap stuff like this, you need to act right away to get data. Thinking doesn't really do anything. I hardly know anything about A/B testing (except what I read here), but I don't think you need to let this test stop you from other tests (just create different landing pages, I think.) You mentioned elsewhere that you were going to wait 'till this experiment ended to start another. Don't delay, do it now.

I think there's a rather strong correlation between people that are used to the command line and people who don't pay for stuff online because they use opensource products. It doesn't make things better that tarsnap works only on linux of course. (If I remember correctly?)

That might be part of the problem as Davidw is hinting at.

It doesn't make things better that tarsnap works only on linux of course. (If I remember correctly?)

Linux, BSD, Solaris, OS X, and Cygwin. Probably other UNIXy OSes too, but I don't know anyone using AIX or Tru64.

Tarsnap doesn't support Windows (except via a POSIX layer like Cygwin) or GNU Hurd.

I've been following tarsnap from afar and think it looks like a great service. That said, an "unprofessional" look & feel is a deterrent for most people because it implies the service may be unprofessional (people tempted to click away before diving in to see whether you're qualified).

Think of it like the crackpots who submit their own math/physics theories. Sure, a professor can critically examine every submission, but it's not a good use of their time. It's easier to just read results in refereed journals.

If someone eccentric is doing something truly groundbreaking (Perelman solving the poincare conjecture), then most people expect somebody to tell them about it. Craigslist is like this -- if you saw the same layout on bobslist.com, you'd know it was some rinky-dink site. But craigslist built a reputation and that reputation is what made it spread.

Unsolicited advice, but long story short: people make snap judgements to save time, and often times they are correct (I suspect most unprofessional-looking sites are not run by professionals in that domain). This can be overcome with word of mouth, but it's ideal to not have to "overcome" their initial snap judgement and make a good first impression.

I think having a gui for OS X could really help it to spread.. I'd love to recommend tarsnap to friends but most of them don't want to use the commandline

You should consider partnering with (or contracting out) to someone already well familiar with OS X and interested in writing a GUI.

Tarsnap has one of the technically best back-end implementations available, but the barrier to entry for an interested customer is incredibly high -- not just on OS X, but on UNIX as well.

Even on UNIX, producing a working backup setup requires writing custom scripts to do so, and implementing anything like snapshot-based backups is complicated enough that even I spent a couple hours getting everything working/tested, and I'm very familiar with this area.

I'll be bold: You solved the backup problem, now the number one hinderance to your growth is solving the user experience problem.

The UNIX utility should have an exceptionally easy way to configure backup schedules. Possibly even an optional web UI.

A Mac release must have a native OS X GUI (no qt!), be easy to use/configure, and be well-integrated into system services.

After the first iteration on the above you should consider implementing support for network-based backup administration, so that we can control backups/restores across all of our desktop/server installations from a centralized administration console.

If the above is not your forte -- if you're only interested in the crypto and backup issues -- then you need to find a parter or contractor who can step in and solve them for you.

I'll be bold: You solved the backup problem, now the number one hinderance to your growth is solving the user experience problem.

I agree 100% -- and this was exactly my intention. I wanted to make sure that I had a solid backup system in place before I made it easy to use, because I hear far too many nightmare stories about backup software which is the opposite way around -- easy to use, but doesn't actually work.

A GUI will happen. Sample cron jobs for automating backups will happen. A web interface... well, that might happen, but really it needs to be built in to something like virtualmin.

I have a friend who is trying to find a way to make the Tarsnap website look nicer without losing its essential geeky/minimalist character;

Maybe you misspoke here, but based on this quote I think your approach could be improved.

1) You went to a friend. Unless this person is an experienced designer of commercial websites, that doesn't sound like the right approach.

2) Your primary concern is preserving the site's "geeky/minimalist character", although that seems to be your Achilles heel. The last time Tarsnap was on news.yc, I made the same incorrect assumption, that this was not really a business, but just a side project of an academic. You and I even talked about that in the comments.

3) You're framing the problem as aesthetics, but it's really about poor communication. Engineers often make the mistake of thinking that visual design = making something look nice. This is like saying a database designer makes data nice.

Just like a database designer sees important patterns to capture, a great producer or designer understands how to drive website visitors to the right understandings and behaviours.

It sounds to me that if you go down the route you descirbe, you'll have all the pain of redoing the site, without solving the more fundamental problems.

You went to a friend. Unless this person is an experienced designer of commercial websites, that doesn't sound like the right approach.

I went to a talented web designer who understands my vision. I think understanding my vision is important -- otherwise I'm likely to pay $$$ only to be presented with a website which I hate and don't want to use.

I think what people are trying to tell you is that--with all due respect--your opinion on your website is not worth a damn.

I mean, if you're interested in pageviews, and revenue--and your comments wrt Balsamiq seem to indicate you are--you need to realize that the aesthetic you're looking for is likely costing you customers and hence money.

Nobody's going to fault you if you prefer to make less money and keep your website any way you want, but if you're going to complain about sluggish growth you've gotta realize that your customers' habits and tastes are much more important than your own.

In related news, the demographic you seem to be targeting is quite possibly either unwilling or unable to pay for it; the former because they'd rather hack something up on their own and the latter because they may not be the decision maker for their organization.

Why not just take a gigantic leap of faith and build a 2nd front-end?

I mean, keep the existing website. And create a second one, using more traditional modern UX techniques. In my perfect fantasy here, this new one would be the .com and your current site would be the .net. That's likely not possible so you may need to find a variation of the domain or, hell, a completely new name for this "non geek" version. After all, "tar" doesn't have any meaning to the non-geek IT Biz Mgr anyway. Being sticky isn't a horrible visual for a backup service, but it's nothing special.

I'm dead serious, here. You have a vision. This is your vision. It's apparently not being successful in the marketplace. So why not just say, "hey, maybe I'm wrong" and go for an approach nearly 100% opposite of your current strategy?

I can't agree more strongly here. I elsewhere argue for a website change; this is even a better idea because it'll even more clearly show the difference. And hey--maybe all of us are wrong and the academic minimalist look is more profitable? If this is true, please come back and report it--maybe we're all wrong!

But unless your customer demographics are vastly different than those of many of the people here who earn good money with proven UI techniques, I'd bet you could do better with different marketing (website, etc.). But you don't really have to bet. What'll it cost you, 10 bucks for a domain name and some time? The first rounds don't really even need a designer, though you'll save some time if you procure one now instead of later.

Test, test, test. There are a good number of qualified people (not me) that are giving good advice, but your test results are king. Start running variations, and you and HN can look at the data and start making progress.

You can look here for ideas on minimalist design:

http://www.minimalsites.com/

Reply on news.ycombinator.com