It's remarkable what users will "put up with" when their core needs are being met. I don't like a lot of things about CL, but does it meet my core needs? Yes; I can reach an enormous number of people with a high rate of success.
It seems to me that CL is a great example of the 80/20 rule -- but unlike most other firms, they simply didn't bother to implement the other 80% of the features.
Still, why do those sites get so complacent? Seems to be the same with Twitter - maybe I am the only one using it via web, but they don't even have a tinyurl button or a search function for one's followers. What are they doing with their time?
I get the impression from a lot of the coverage lately that they've been working on an actual plan to make money. I'm speculating, but this could involve developers working on features or even an alternate platform conducive to that effort that wouldn't show up in the currently public-facing free system.
I saw something similar with Heroku. In fact, in the RubyConf talk, Adam Wiggins pretty much stated that they were asked by their investors to switch gears to the making money part of things. Thus, the Heroku service stayed more or less the same (and good, I might add) for a period of time up until this recent announcement that other stuff had been happening behind the scenes.
They've made pretty great improvements to the availablity and reliability. It still goes down, but not nearly as often as it did. That despite virtually doubling their traffic in the last 6 months. So my guess is they have just about everyone focusing on optimizing and scaling.
I agree though, lots more they could do including a solid oauth implementation for the api...perpetually coming soon...
I know PG posted that YC would like to fund a Craigslist competitor. The issue is that it's hard to build the number of people needed to reach a critical mass. Likewise, it's not that eBay is so great, it's just the only place with that many people.
The CL search is really simple - really just whole word matches. That makes it hard to search for things that haven't been adopted by the CL community as keywords for items. I'm guessing this might be related to the amount of hardware they run. While search isn't that taxing a thing to do, it isn't necessarily easy (especially on CL's budget which doesn't rake in dough from VC or ads).
Think about comparing it to something like Digg. CL has more traffic (http://siteanalytics.compete.com/craigslist.org+digg.com/?me... - go to visitors->visits monthly and engagement->pages per visit). About 3.5 times the number of visits and about 10 times the number of pages per visit. So, CL is getting 35 times the amount of traffic of a major site like Digg. It's hard to scale up to 17.5 billion page views per month (24M per hour, 6,800 per second) - especially when you're not monetizing most of that. Comparatively, Digg is getting a mere 200 page views per second.
I'd love to see improvements to CL, but I feel like some of them would definitely require more hardware. CL wants to be independent with as little commercialness as possible. In the end, that leaves CL what it is - a bastion of whatever goes where it's also hard to find things.
Craigs is simple enough that most of its users can feel like they understand everything about it. Despite its being huge in users and content, this makes it feel small.
If a user feels like they understands a site, they don't mind spending time on it finding things. It is the opaque, confusing sites that have people run away after a few searches. The challenge for any competitor is finding a way to expand functionality while allowing the user to still feel like they understand exactly how everything works.
I would love to create a craigslist competitor. An improvement on craigslist would be a site which allowed the user to search for item, price, location, and time together - "yoga classes on Wednesdays in Berkeley for under $100" or "garage sales Saturday with car parts" etc. What, where, when, and how-much could be specified separately and orthogonally.
Remember, getting calendaring, location, price and category to work together is a tough problem. Also, making sure postings remain fresh is tough if you do anything other than craigslist's system of making everyone repost their content every two weeks. And you would have to be willing not to monetize any more than craigslist, 'cause the users can tell.
CL makes something like 50 million a year. I just don't believe they are incapable of fixing search. Hell Plenty of Fish runs on something like 6 servers + a CDN and offers a very structured advanced search.
I think this is just the typical case of someone being surprised by their success, and being too scared to change anything, in fear of driving users away.
CL can spend 5-10 million on infrastructure out of profits...w/o selling out, not that it would cost that much...they just choose not to.
Hopefully Jeremy Zawodny will drop in here and give an answer, since he left Yahoo and is working on mysql stuff for Craigslist now (including search, I believe).
I've actually built a feature that addresses one of the concerns in the linked article... now I just need to do the last 5% to ship it, I guess.
No, I'm not kidding.
But, yeah, we have a TODO list a mile long at craiglist. A lot of time goes into keeping up with growth and those who want to do Bad Things. But we also hack on new features too. We all use craigslist in addition to working there.
The double edged sword of Craigslist is that they accept nearly everything. They're good enough for most people, which is why they'll continue to dominate.
However there are many vertical niches within classifieds that can still compete with Craigslist. Apartments, Houses, Cars, Crafts and Jobs have all split out into separate niches and seem to be doing well. There are still tons of smaller markets virtually untapped (Pets come to mind).
"If CraigsList had an API, it would be a game-changer."
Agreed. As a long time CL user, I have thought the exact same thing. Short of API functionality and extensibility, I know of numerous scraper/mashup applications that utilize CL data. Unfortunately, this does fall far short of API extensibility and likely leads to liability in concerns to "stealing" CL's data - I would be apprehensive about investing 'too much' on any project that relies on scraping as chances are that the project will be killed by CL before reaching the critical user mass of "success".
Craigslist is like a social network site - its users and brand name are its sole asset. The actual design of the site isn't ideal, but it's enough to keep the masses happy, and that's what matters. A single competent competitor could take it down (like facebook did to myspace), but it would require a compelling reason to use it, not just better design.
My main complaint about CL is the lack of user profiles. On ebay, I can (sort of) use the user profile and ratings to gauge risk. This is especially important on something like CL, where you actually meet face-to-face with a person, not just get a product shipped to your address. I have had some sketchy experiences on CL...
The newspaper industry needs to stop complaining about Craigslist and realize that there's a market need for a better product. Craigslist is a good site, but it's full of scams + the search results are a mess + it's too anonymous.
Yup. Short of an official supported API, we are pretty much tied on utilizing CL's data. While screen scraping CL is easy and can yield fully utilizable data for an outside application, the application will be doomed to failure (financially, at least) because it will always be arguable in court that your application is worthless and that it is their proprietary data that ultimately holds the value of your application.
Personally, I've had to start searching Craigslist posts with Google in order to find listings I'd visited recently. They should just outsource their internal search to Google, and if they feel like making many millions of dollars, run Adwords ads along the search results.
Is it that Craigslist is rotting, or is it that people's expectations of search have moved forwards? How much would it cost to license search from Google? Would Google provide it for free? Are enough people scared of Google now that it would be a PR lose?
Google Search Appliance? http://www.google.com/enterprise/gsa
We have a couple of those where I work - they seem ok. Way better than the search function that came with our stupidly expensive "Enterprise Portal".
It would be very easy to improve searching on CL without needing to license anything from Google, especially given that CL makes plenty of money. It's quite shocking just how bad the Craigslist user interface is, and has been for so long.
Here is my attempt to give Craigslist the features it badly needs (keeping track of which ads emails come from, keeping track of ads): http://listingninja.com
The problem with this approach is that it substitutes good somewhere else for badness here and now. I like craigslist. I like the honesty, the local-ness and the community. It is useful to me. I have never encountered scammers or slimers like on ebay. I've found real people I can related to. Running partners, indoor soccer teams, tennis rackets. They are doing good for the world, and they're doing it directly and sincerely. They can do it better--I like many of the features listed here that they lack--or a competitor will. But I hope they or the competitor retain the sincerity that is a unique gem in today's corpate world.
It seems to me that CL is a great example of the 80/20 rule -- but unlike most other firms, they simply didn't bother to implement the other 80% of the features.