Ask HN: Are you working on a desktop software?

- I used e-junkie.com for sales and distribution, it works really well, plugs in with paypal and automates everything

- I offered a 4 day demo trial, 1 in 300 people that downloaded the demo went on to purchase, although my product wasnt very good the numbers for others ive read are still very low.. tons of people just download apps just to try with no real intention of buying, this makes adwords advertising hard to make economical unless you convert REALLY well or price really high

- download sites never sent me very much traffic, but I was in a very crowded niche ( typing teacher programs )

- if you are developing for msft, make sure you test your program an various versions of windows.

- Really polish your software, I made the mistake of throwing something that wasn't very good out there. Your software has to be VERY GOOD to get people to pull out their credit cards. You really need to be the best in your niche. You cant think that 'oh photoshop makes 100 million a year, I just need to make a photoshop clone that has 1% of the features and is 1% as good, then I'll be a millionaire' It doesn't work like that.

About 20% of those who downloaded were not able to install it due to the following reasons:

* Could not find "where did it go?" after downloading (not joking)

* Were running some kind of "security suite" which was crippling our installer or didn't let our product connect to the internet and users couldn't figure out which program performed the function of a firewall, let alone configure it.

* Their computers were heavily modified by other installed Windows software (non-standard DLLs in /windows/system), missing/modified registry entries, rootkits+spyware+malware+adware. Windows allows this kind of thing.

Technically the product was nothing fancy, an Internet Explorer plugin, but it surely knows how it feels to be a domesticated cat dropped from a comfy helicopter into amazonian jungle. [dev. machine -> an average user's machine]

I spent most of my career developing Windows libraries for other developers (ActiveX controls, signal processing, industrial automation, that kind of thing) and never faced a "6-pack Joe" as a customer. Once I've tried, I got rid of all my Windows computers, forced myself to open a whole new UNIX world and yes, switched to mostly Web development, which I find extremely trivial and boring (HTMLization of DB schemas), but I'm glad I don't have to deal with "Where did it go?" anymore.

* For distribution, I just have a website. I also put it up on the Apple downloads listing, which is where most of my traffic comes from.

* A bunch of other sites pull data from Apple's listings, so I get traffic from there too.

* More updates = more traffic = more sales, because each update gets you near the top of the download sites

* Sales are done through PayPal, which is really easy to setup

* Support email isn't bad at all, maybe 1 email every other day

* The free trial limits the amount of data you can enter, rather than a time trial. That way users can try out future updates too.

* When a major site did a post about it, sales AND feature request emails spiked. But after a week, it goes back to normal.

The biggest difference I've noted since moving to the web is that on the desktop (Windows at least), you need beta testers.

Your users are more distant and therefore disengaged (they got your app from download.com and have no idea how to contact you), older versions of your app will float around forever, and user's systems are varied enough that you can never do adequate testing inhouse.

Pay attention to those users that do contact you, because for every one of them there's several that had the same issue with your trial version and walked away. At the same time, learn to recognise unique problems and fucked up Windows machines and walk away - they're just time sinks.

Entice people with screenshots and a sales pitch, then make download and installation as painless as possible. Try and keep the installer size down. Definitely avoid making people download and install any sort of runtime environment separately.

Tying registration codes to identity (regcode = hmac_sha256(secret + emailaddress)) will cut down on regcode sharing a bit during your early days, but keygens and cracked versions will eventually appear. There are a few turnkey anti-piracy / activation / etc solutions about these days, I never evaluated them.

A retail version with a CD, box and a little printed manual will appeal to lots of people who wouldn't go for a download. Some may not have internet access at home, and most won't upgrade. Make your retail version once that's proven itself as a download for as long as is feasible :)

Most of the discussion on here is about web-based apps. We create both web and desktop based software. One of our products (SourceGuardian.com) provides encryption for PHP source code. Our experience was to create something that we wanted to use ourselves, as the competition at the time was charging $6000 a license, and then we turned that into a product. We have created several other applications and are working on some backup software which links with Amazon EC2/S3 and is due for launch in Jan 2009. For this, and for the other software under this brand we'll be looking primarily at a reseller model. Finally we are also working on hybrid apps, that take the power of the internet/servers/cloud and use that for processing, bringing the data back to a desktop app/widget as well as iphone. The latest one is competitive intelligence online and my own feeling is that desktop apps still have the edge for usability right now. Why not tell us what you're working on and I'm sure some people on here will provide you with help in terms of what you are looking for.

(edit) actually that whole discussion is relevant - http://news.ycombinator.com/item?id=322630

Business: http://www.bingocardcreator.com Blog: http://www.kalzumeus.com

Just hit $20k in revenue this year does a little happy dance.

Distribution: Website. I get people to the website by mostly SEO and AdWords, with an increasing word of mouth contribution (more than a thousand paying customers means that, increasingly, you hear "Oh, I was looking for JUST THIS and then Myrtle at church told me...").

Sales: Web site? See blog for copious tips on the use of payment processors (I use http://www.e-junkie.com as a wrapper around Paypal/Google Checkout) and how to write sales copy.

Download sites: Nearly worthless for PC software.

You can sometimes live on eyeballs with a web app but, in desktop land, you need to make something people are willing to pay cash for. If you get to that point, you're golden.

I like distributing with Java because I want schools to use the app, and when I give it to teachers to try they don't have to have special permission to install it. As long as java has been cleared for their computer, installing my app is as easy as using a .pdf file.

The sales just aren't there, though. (Many have downloaded the demo version, which (a) is good enough or, more likely (b) convinces them the app is not for them. Dang.

I will launch a complementary product in Jan 2009, which will hopefully provide a synergistic lift to the whole effort!

We're using WPF and this is the operations tool which will eventually be complemented by a browser app. It's our hope that we'll be able to force enough of the WPF code into Silverlight to make it the front-end for the browser app. We're also going to try using WPF in the browser but I think the installation headaches will be too much for us.

I should probably make my own thread about this, but if anybody wants to check it out it's available here: http://netifera.com/downloads/

It a cross-platform desktop-based time-tracking application.

We used JRuby + Swing plus Monkeybars. It needs a license key to run; we have a Web service handling that to generate 30-day trial keys or fully-registered non-expiring keys. That's handled by Ramaze on Glassfish.

Sales are done through Shopify and Google Checkout.