Poll: How do you create the website for your startup or side project?

No CMS option?

End up wasting hours and hours choosing the right colour scheme for the website... in the end the side project ends up getting put on hold while I work out the best framework to use for the website.

edit: wait until I have to choose the 'right' font...

Good ol Twitter Bootstrap, Jekyll, and S3. Even an incompetent like me can put up website:

http://www.smalldatajournalism.com/

Er.. I've tried most of them :)

For the latest one I bought a template from ThemeForest. (so I voted that)

The one before that was a Bootstrap (the framework) effort, which ended up not very nice.

Before that, I tried Joomla.

And before that, I had a very basic HTML page :)

I've realized I'm really bad at the "design" part, I'm better off doing the backend stuff.

I think the consensus here is that I know absolutely nothing about web design; but I do it myself anyway.

Web design and web development are two very different things that overlap heavily. Judging from the options you have provided I suggest you pay someone to do it and focus on what you do well.

I've tried a few.

The problem with Themeforest is how muddled and complicated the HTML DOM in themes usually are - and even if the design is intuitive, it's not yours, so it's rather difficult to own completely. With that said, I've used it for some excellent WordPress projects before, and if you are willing to pay there is quality work there.

I prefer using a bootstrap framework to get stuff out there. Depending on the project, I'll spend some time customizing the defaults, adding plugins, etc. If the project is small enough, I won't even bother with that. Bringing in a design agency is only useful is you have VC cash to burn. It's usually far cheaper/effective to find a design freelancer to partner with. A decent designer can help you create a logo as well, which is an essential step, and something that's difficult for laymen to pull off well.

Remember the enemy of progress is the perfect. A side project/startup needs to iterate as fast as possible, so it's far better to throw something out there than to languish in design mode for months before giving up on your cool idea.

For my side projects, my rule of thumb is: a) the site isn't embarrassingly bad, and could be publicly associated with me. b) site design took less than 20% of my overall effort on the project.

I've used them all but settled on Themeforest. Yes you can make your own, but why would you when you can choose something that's already well designed for a couple of dollars. Save the design practice for the app itself.

As far as mechanics go, I've found a great compromise between the power and convenience of a CMS and the simplicity of a static page: using a static site generator. I personally prefer Hakyll[1], as you might have guessed :P.

This makes managing a structured site easy without having to configure a complicated server. Of course, it also means you can't have much dynamic content. For something simple--say comments--you could use an external service like Disqus, but you won't be able to do much beyond that without running your own server.

[1]: http://jaspervdj.be/hakyll/

I write open source libraries primarily. A nicely formatted readme on github tends to be best, since most people will be there anyway. Fancy websites tend to drive me crazy for code libraries, since they tend to distract people from the thing they are really looking for; a simple pitch and documentation.

That being said... I have been working on a javascript geoprocessing library lately and am considering a full blown interactive website. The reason being that many of the features are difficult for people to understand without being visualized. I still want to keep it fairly simple though.

The idea would basically be to have a page for each operation with 2 maps and a description of the algorithm. The first map would show the example input and the second map would show the output. I think this would allow developers who may not be familiar with more advanced geo statistical methodologies to be able to see what is possible with their data and spark some creativity. The existing readme tries its best to get the ideas across, but I doubt it would be that effective in its current form for someone who is not very familiar with traditional gis analysis already. https://github.com/morganherlocker/geo.js

For something that actually matters to me, I'll work with a designer on the core visual concept, meaning usually the logo, color palette, and overall feel. Once I have that, I feel competent staying within the given parameters to put together a pretty solid design.

It really helps to spend some effort learning the basic rules of typography and visual design. It won't make you a great designer, but it will make you a much smarter customer of designers, and make you able to execute someone else's design concepts gracefully.

Bootstrap/Foundation, strip out what I don't need (read: delete some of the LESS imports) and customise it. Both frameworks are actually fairly easy to customise these days, and it doesn't take a significant amount of effort to end up with a Bootstrap site that doesn't "look like a Bootstrap site". I'm nowhere near skilled enough (nor do I have the time) to write my own grid system, panels, basic buttons, etc.

The blog attached to my side-project will use the same CSS and just leverage jekyll/GitHub Pages/prose.io to keep things simple.

This is exactly what Drywall was designed for. It's a website and user system for Node.js

Feedback is welcomed.

http://jedireza.github.io/drywall/

Edit: more so for a functional side project, not a static website.

Use Strikingly of course. No brainer.

I've bounced around a bit.

Way back in the day, I had no budget, so I used OSS web themes with very light customization. Then when I actually started selling appreciable amounts of software, I had that website redone by freelancers (twice). It has not been my experience that cutting-edge web design has made huge differences to sales in the sort of markets I tend to operate in, though, so these days I mostly just get a themeforest/etc theme, have designers add in decent graphics (if required), and then do a bit of munging by myself.

Designers, cover your ears for a moment: There are many businesses which have sold many millions of dollars of product, including in our industry, with web design which is less impressive than things you can get on Themeforest/WooThemes/etc for whole tens of dollars. (This is not just applicable to marketing sites, by the way. There are plenty of applications which end up looking like they were designed by a software engineer which, for $15, could have had that same engineer just extract a Rails template from a Themeforest project and end up looking 100x better. Search for [admin] on Themeforest and feast your eyes.)

  > It has not been my experience that cutting-edge web design 
  > has made huge differences to sales in the sort of markets
  > I tend to operate in

  > patio11 made an understatement to get his point across -
  > that design doesn't matter as much as people think

Unless you really know who you're selling to, putting a lot of effort into a site design is probably a premature optimization. Just grab a reasonable free template or learn enough design [1] to do it yourself.

[1] I like the Non-Designer's Design Book by Robin Williams: http://www.peachpit.com/store/non-designers-design-book-9780...

Drupal + Bootstrap + a bit of custom css/html

neocities if it's small enough, tornado for internal demos of ideas. As a person whp develops deep learning back end stuff I often find that a decent UI to show case an idea is far more successful than circulating a paper (and I despise PPT)

I used to obsess over style and color and fonts and stuff. Now i'm likely to just to Slim/Laravel + Twig + Bootstrap and try to move on with my life.

Which is weird... I was the "art kid" for most of my life but I realized that when it comes to this I like the code a lot more than the design.

I know absolutely nothing about design. I use bootstrap and more or less copy pre-existing designs that have the desired aesthetic.

Problem is that I can tell when a UI is terrible, but coming up with the right one is not as simple as the inverse of terrible :(

Pharo Seaside, jQuery, Bootstrap or PureCSS, postgreSQL

I just launched Rate My App[0] yesterday evening on Hacker News, so this is great timing. :)

In the four days between idea and launch, how much time did I spend on the landing page, and would I do it again?

As an experienced engineer, I realized the first thing I had to worry about was the marketing content, so I spent the first day configuring my social accounts and creating and uploading a video.

Short version:

Yesterday, I implemented my own design because I wanted a simple site that had exactly the code I needed. I would do it again, especially since the next time I can re-use the build process I created.

Longer version:

Assuming you're familiar with CSS and HTML, you're better off finding a decent template or web site and re-creating the few parts about it you like yourself. Using a fancy template will be a time sink, both in understanding all the complicated things the author threw in, but also focusing your time on making your site perfect (since you just paid for it) rather than focusing on launching.

And in the two days in-between, I figured out the value I could bring to customers. I also spent a lot of time configuring payments, emails, form submissions, etc.

Establishing some process to collect feedback and getting your name and purpose out there is the main point of launching, so worry about that more than the landing page.

TL;DR: as far as systems, you're better off establishing some kind of a process (hopefully an off-the-shelf system that can function independent of your concerns) that allows you to iterate on your actual product quickly. As far as visuals, though, spending lots of time on making this tiny part of your product perfectly integrated with a particular template can make it emotionally difficult or time-consuming for you to replace or modify as necessary later on.

I'm writing a series of blog posts (@vla on twitter) about this, by the way.

[0] the app: www.ratemyapp.com

I have couple of websites. The first one[0] is done with basic Bootstrap. Second[1], I bought BootStrap theme from WrapBootStrap and reducing it to my needs.

Apart from this, the main issue is designing logos and graphics. Eventhough it takes time, I somehow do it myself(with GIMP) and enjoy it as well :)

[0] http://www.computeloops.com [1] http://www.remotewaker.com

I think most of us don't have budgets for hiring designers for side projects, because they are just side projects which may not merit spending more money on design, as a lot of hackers do know the very basics of design. If it does grow out of the side project phase and is actually showing potential to generate some serious cash flow, then more money is spent on proper design.

I know UI Design and a just a little bit of backend. My sideprojects are always more about design and less about technical part, however whatever little technical part it requires, it gives me enough to learn something new.

For http://typezebra.com/ I learnt a bit of Javascript while for https://bootmystrap.com/ I learnt to design websites properly using LESS CSS.

The way I keep myself motivated to finish sideprojects is that most of the part is right up my alley, design which gives me confidence I can finish it.