Ravelry - a very successful deployment of Ruby and Rails | Hacker News Reader

Ravelry - a very successful deployment of Ruby and Rails

tbray.org

96 pointscolinprince28 comments

I note that this has been mentioned in 5 comments before in the history of Hacker News. Which is approximately once for every terabyte of knitting patterns they have stored. Or once for every 100,000 registered users. Or once for every 700k monthly page views.

Excuse me. Daily page views.

But to look at how much attention they get among technically inclined people, you'd think they were less important than the average URL shortener.

There are a million and one underserved markets in the other half of the population, gentleman, and no one you know has any interest in competing with you if you slice off one of them.

Indeed. And the reason I look over my wife's shoulder now and then when she's on her computer.

She's a craft/fashion/music/wedding nut and I've had more than a couple ideas for that half of the population.

Agreed. I am currently working on a site to fill a need that my wife, an artist, has. I found that the best way to find great ideas is to ask friends and family.

The downside with friends and family is that they want to help you. Regular people want to help themselves. Find regular people, ask what their problems are, solve them, make money.

A couple of days ago, I read the following article: http://www.forbes.com/2009/08/27/million-dollar-businesses-e...

One of the businesses is Jimmy Beans Wool, started in 2002 with $30k and had $2.1 mil revenue in 2008.

now boasts an average of 20,000 customers per month, mostly through the Web site.

Definitely. My wife was a early member of ravelry, and watching their success made an impression on me. Before it existed, knitters were mostly members of some vbulletin-like board whose name escapes me. I suppose that a good way to identify these underserved markets would be to look for popular vbulletin/phpbb sites aimed at some particular hobby, and try to think of ways to make something more useful for those hobbyists.

Also, you don't have to limit yourself to the female half of the population. There are lots of industries with a significant male presence that use some crap software. I can't believe how bad some of the software is that I've used working for non-technical companies.

Working on my own: http://doleaf.com

Not that guys can't be into gardens as well. I love getting my hands dirty :)

I have a couple ideas for startups in this space. Would love to know what you all think. Products/Companies that cater to a largely female, largely suburban audience have a tremendous amount of potential.

http://news.ycombinator.com/item?id=804063

"If you listen to Paul Graham and whoever else, then you’ll be working on your startup while you have a day job."

Funny how he kind of got that backwards. pg says go fulltime. 37signals are the 'keep the dayjob' crowd.

heh. Oops.

My wife knew about this site, and she tells me that for a long time growth was somewhat hampered by their inability to scale. They had to manually allow users into the site in chunks (you email them your info, they email you the following Friday when you can finally post on the boards). He obviously had some hiccups, but he managed them well and learned as he went. That's an admirable quality.

I completely agree with patio11's comment about the other half of the population being under-served.

We had an automated invitation system - the usual, "put in your email, wait a while for the invite" thing. It was true that were only adding a couple hundred people a day during the early days. We weren't ready for more. We were so not prepared for the interest - it's funny to think back on it now. When we hit 5,000 users we were amazed.

I think that we had actual performance issues for about a month (maybe a little longer?) when we outgrew our single dedicated server and hobbled along until we built a more permanent home. Luckily, people were very patient and understanding.

Cool! Thanks for the info. It's nice reading about people's efforts to scale mid-to-large sites. All the press is on how the "giants" (Twitter, Facebook, etc) do it.

My wife knew about this site, and she tells me that for a long time growth was somewhat hampered by their inability to scale. They had to manually allow users into the site in chunks

It wouldn't surprise me if that "whoa, it's so busy that we have to queue to get in" factor actually contributed to its final success. I've seen it before with other apps. The "it's not really a big secret but it sorta is" beta of Stack Overflow was a little like that too. People wanna get into things they "can't" get into.

Good point, that hadn't occurred to me.

I recently introduced my wife (a crochet-addict) to Ravelry. The result? She's now a Ravelry-addict. I can hardly get on the laptop myself these days. She's taking photos, queuing up projects, checking out patterns and even going as far as entering her "stash" of wool so she can track it all.

The one problem is that there's no api so it's a walled garden... and my wife really wants a public portfolio site. If Ravelry could fulfil this need, I'd pay them money to get it off my todo list!

I often wondered about the business model but after seeing it in action, the advertising fits quite well.

http://www.ravelry.com/help/api isn't much but it should be enough to make a project portfolio. Post on the Ravelry board if you have any questions :)

hmmm... interesting.

It looks like you can only get info for in-progress projects though. Really she wants a public version of the projects page to show off to family/friends/non-ravelrers and ravelry is a much better interface for building this content than her current Wordpress solution.

Edit: ok, I've checked out the raw JSON data and looks quite possible. Thanks!

Excellent article about creating a niche community within an underserved market. I'm not sure the success here is as much about the platform as it is about the brilliant utility of the site to its members.

Agreed, although I don't know about niche (a lot of people knit...) or underserved. There are numerous resources and communities for knitters and crocheters out there. The only reason why I love Ravelry over them all is because it manages to wrap everything up into a nice user-friendly website that (if one cared for them) is packed full of all sorts of useful features that other sites couldn't even begin to offer. Try to find another knitting community online where you can look at a pattern done in another yarn without attempting to knit it (and possibly wasting a lot of time). It's a lot harder than you think. On Ravelry it's a single click away.

Ravelry doesn't even replace half of these resources out there - most of the patterns are still going to be in books and magazines or on other websites, all the knitter bloggers will keep on blogging on their own sites, people will still go on other forums for, say, a specific yarn brand to ask questions there. But it is the glue that holds it all together in a single place so you don't have to go searching in Google and half a dozen sites to find patterns and talk to other fiber arts enthusiasts.

900 new users a day. For a knitting and crochet site. Absolutely amazing.

700,000 members and still 'beta'?

Google Syndrome. "Beta" is meaningless now.

Pretty good setup, but Cogent? Eek! Those guys run the crappiest network I've ever seen. You certainly get what you pay for. And they're combining that with CloudFront? Seems like a weird combination that results in their not getting their money's worth (on Amazon's side).

I feel weird replying to all of these comments but here goes:

We were worried about Cogent at the beginning so we only used them for delivering some static files and for non-critical outgoing traffic. It was initially an experiment - to save on our in-datacenter bandwidth and to compare with Amazon (a very inexpensive experiment) and we've been very happy with it.

I think that we are doing about 20% of our traffic (+ offsite backups) over Cogent, 40% on our datacenter's fancy-pants bandwidth, and 40% on Amazon.

When you wanted customer input, what was the most effective way of getting it?

Input from users? Our forums - lots of time talking with our people in the forums. I wrote a little post about how we collect feedback here: http://codemonkey.ravelry.com/2008/06/29/beta-testing-and-be...

Check out BurdaStyle.com - also a very successfull DIY site that's built on Rails.

Reply on news.ycombinator.com