Ask HN: Scared of freemium test being all free and no 'mium'...

[beforeistart]

One of my many unfinished projects has now languished for so long that it's actually in danger of being finished sometime soon. I don't want to give away the specifics yet, but it's an ajax tool that allows you to tinker with some controls, and each time you do that it generates a dynamic output image of approx 100k. It's something you might use anywhere from just as a last-resort lookup, to an essential tool for a job/hobby that takes up every second of your spare time. At 100k per click, that'll add up lightning fast, and if it hits the right audience I can see it getting swamped. There's a few national magazines on the subject that sing the praises of much less impressive tools - one mention there and it could be adios! :/

Consequently, I wouldn't know where to start in estimating bandwidth use, and that's really all the cost I'll have until I'm ready to market it. The cost-free limit on my current host is 15G/month, but I haven't got the time/resources to set up anything better just yet. I have no money to throw at this, at all, no matter what. I'm in a situation where it costs me nothing or I put it off for a year :(

So, I'm planning to do this as a freemium with a fairly hefty restriction on the free and usage-bracketed costs on the premium so as to be fair, but I'm still scared the numbers might not add up and that the freemium could potentially cost me an arm and a leg before I even notice.

Does anyone have any nuggets of advice that might help? Even if it's just "It's a risk, get over it" :) I'd rather not abandon the freemium idea (pictures of it in action are OK, but using it is a hundred times better), but maybe the slower, safer approach of only using bandwidth that's already paid for is warranted for a skint coward?

[Joakal]

Startups are generally told that popularity is a nice problem to have. Further it sounds like you're over-scaling before you release it. I have no idea what you're releasing though.

If you wanted to look at monetisation opportunities, even if it'll cost a lot of money, you need to say what you offer. Then people can say what you can sell. eg Allow people to do 10 images. Want to do more than 10 a day? Premium account! But how do you define a price?

Recommended material to read about: Marketing, monetisation and something about releasing often (self-educational?)

[beforeistart]

> I have no idea what you're releasing though

No - I remember the knockoff debacle with the phone-me-to-tell-me-I'm-awesome idea, and while I agree with the "get over yourself, no-one cares" advice, I'm finding it a bit hard to swallow. It's a very obvious gap that's been sitting around unfilled for ages, and I'm not exactly quick to get things done :)

Thanks for the advice - over-scaling is right on the mark, I think.

[revorad]

A friend's site, where users upload lots of pictures every hour, gets millions of hits everyday. It costs him $20-$30 on S3 per month.

You're probably overthinking this. Just run the test. Chances are no one will care about your app.

But if you get lucky and get a bazillion hits, you'll find out very quickly. Stick some Google ads on. If you capture emails, you should also be able to convert some to premium.

[beforeistart]

Offloading the image part to a cheaper service is a good idea - the rest is fairly lightweight, and once they're generated, the images are themselves are fixed/cacheable. Thanks!

[rubinelli]

Can you generate low-resolution intermediate images and only create the full-resolution image in the end? What if you changed your UI to display some kind of thumbnail timeline? Another operation, another generated thumbnail (taking, let's say, 5k) is added to the list. This adds some interesting benefits, like letting your users compare and revert changes easily. When the user is done, you can upload the finished product to S3, and maybe ask unregistered users to create a free account before you return the link.

[myhnhandle]

Get a virtual linux box in Germany, most are sold with unlimited data plans (in german: 'Traffic-Flatrate'). They might disconnect your machine when you hit the terabyte ceiling but the financial risk would be minimized.

[beforeistart]

Thanks, I'll look into that for offloading the image storage.

[custominstall]

Dreamhost and Bluehost seem to provide unlimited bandwith on their shared plans for around $9/mo. I was thinking of moving to them myself. But not sure if the shared hosting will meet your needs,

[beforeistart]

I've run into both those names in a negative context without even being a customer of either. Thanks for the suggestion, but those two aren't worth the risk :)

Aside: Wasn't Dreamhost the one who put an answerphone message from a distraught customer on their blog with an entry taking the piss? I tried to check, but I can't find any reference to it amongst the myriad other complaints...

[Joakal]

Look up their reputation. You'll find them worse than another provider with the same price.

[realschool]

shared hosting is very limiting, if you want to do any real tinkering you will probably need a VPS.

[petervandijck]

The best way to "estimate" your bandwidth is to launch :) You can always turn it off and re-think.

What's much more likely to happen is that almost nobody will use your service.

[beforeistart]

Rightly or wrongly, I dread turning it off once it's launched as much as I dread being out of pocket :)

I'm thinking that's the fear that I'll be getting over first, though, as it seems fairly unanimous that I should should just shut up and get on with it :)