Hacker News new | past | comments | ask | show | jobs | submit login

Disagree. Craig Newmark has said that he's scared to make drastic changes to the UX specifically because he doesn't know what magic recipe keeps people coming back. Do I like the CL layout? Nope. Does that mean other people should steal the content and profit from it? Probably not.

The magic recipe that keeps people coming back is that Craigslist has a large network.

I would be willing to bet a large amount of money that adding a map to search results and letting me put an upper limit on the number of bedrooms I'm looking for would not make people less likely to use Craigslist.

Their magic recipe is that an amazing amount of non-technical people have figured out how to work with the site.

The Dutch version of Craigslist (Marktplaats, owned by eBay) is in a similar predicament.

They have a very sucky UI and have experimented with new and improved versions, but every improvement was quickly reverted after large drops in engagement.

It's almost surreal to look at their site (http://marktplaats.nl) and realize that there's tons of highly talented people behind it.

Yup, there's a catch-22 here: when your site has shitty usability but is the only source of information, people hit more pages and spend more time there. When you fix your UI mistakes, they're in and out quicker because they actually get what they need efficiently. Because you're optimizing for engagement by measuring time on site or something like that, you end up making sure never to make your site easy to use.

This is not a recipe for success. It's a symptom of measuring the wrong metric, especially when you have a near monopoly on a market due to network effects. If this is really why Craigslist is not improving their UI, they need to hire some analysts that actually know WTF they're doing.

Haha, I see what you did there. You confused UI with Usability. Craigslist is incredibly usable. Have you taken a look at Ebay lately? Does it look familiar? That's cos it hasn't significantly changed in 10 years. Is that because they don't care? Nope, it's because gradients and drop shadows make your product look better, but don't make it more usable.

Craigslist is a utility, not a website. They care about keeping it useful not only for users today, but for users in 10 years. That's not something startups care about at all. Startups are looking to amass users and then sell. Craigslist cares not about such goals.

Go ahead, get mad at their inflexibility. But then take a step back and think about what craigslist represents. And then wonder if you yourself would have the cajones to stick to your guns in todays startup market. I for one commend their ability to keep their product strong in the face of all the bells and whistles of web 2.0 or whatever it's called these days.

>You confused UI with Usability.

I certainly don't confuse the two.

I couldn't care less what Craigslist LOOKS like, but its usability is profoundly broken in some areas (Denver Metro, for example).

The problem is that there's a Denver, Boulder, Colorado Springs, and Fort Collins Craigslist. For those of you in the Bay Area, that would be roughly like having a distinct San Francisco, San Rafael, San Jose, and Santa Cruz Craigslist. If you're in between those places, which one do you look at? You have to look at more than one. Even if you're IN one of those places, which one do you look at? Sometimes three or four, if it's something you're willing to drive an hour to buy.

And if you're POSTING, where do you post? If you post a popular item to Boulder (because, say, you live in Boulder), then people in the next town over who only know about Denver won't see it, so you also have to post to Denver if you want everyone in a reasonable radius to see your item. Which means you have to violate Craigslist policies AND mess with your listing enough to get past its spam filter.

The sad thing is they've fixed it to the 80% level in the Bay Area, where the whole (extended!) Bay Area (even Santa Cruz, which isn't really Bay Area) is on a single Craigslist site with sub-area filters, but they just leave it broken here.

THIS is a usability problem.

And honestly so is the lack of "map these results," for times when you want to see which ones are closest, or which ones you could stop by on your way to work, or whatever.

It's not because it's not shiny or doesn't have gradients or drop shadows. It's because key features that I want from a classified site are sorely lacking.

I disagree, craigslist sucks. It's filled with spam and it's searching and sorting options could be way better.

The only reason they don't have many competitors is because they have an information/platform monopoly. Just like many other tech companies, ebay and facebook come to mind as obvious examples.

You can't view apartment search results on a fucking map.

Are you really going to argue that the site is more usable without that very basic feature?

Just because the site is simple in design, doesn't mean that it is built to be used efficiently. Craigslist, as is, is a usability nightmare. When I use the site, I am using it with various Greasemonkey scripts, Firefox extensions and 3rd party web apps to make their content easier to consume/see/understand/filter. These tools are the only reason that I can bare to use Craigslist.

If I am searching for a bike, a simple script to show image thumbnails by every listing will let me see which listings interest me. This saves me time, because now I only have to open a few interesting links. I don't have to scour through many of text links, just to see if they interest me.

The rentals section is constantly bombarded by spam and has very few filtering options. There is no way to exclude keywords. Boolean operators would greatly improve the usability of the site.

You can exclude keywords and there is a boolean OR operator: |

The problem is that you have to enter some search query before you can use the negation operator.

Also, protip to get past the spam in rentals section (Austin is holy crap bad about this): enter 'google' as your search term. This will only return ads that have an address specified and hence have the 'google maps' links. Plus, since you've entered a search term, you can filter out the locator who keeps spamming and screwing up your search by putting his phone number in the address section.

Here's a query I might have used during my last apartment search.

    google downtown | central -"512-555-5555" -"CALL NOW" -"austin apartments NOW"

Because you're optimizing for engagement by measuring time on site or something like that

Oh good, nobody actually told you how they measured but you already know! Bravo.

I can tell you from first-hand experience that they know what they're doing. They measure almost everything, nicely put together in dashboards with all the KPI's for the appropriate managers.

I think it's unlikely they used a suboptimal performance indicator to judge the site's new design, and would say the same thing about craigslist.

If only they would measure amount of spam in their for rent listing that currently nearly useless.

Remember, if the numbers don't support your opinion, it's because your measuring the wrong thing!

Keep in mind 95% of their site is not the product they are trying to sell. It is in craigslist's best interest to keep users glued to their site. If someone takes 30 minutes to find a used bike instead of five then great. Maybe there is a chance the user get's bored of cruising bike listings or on one of their many trips to see the latest postings will click on the job section. Craigslist uses their massive user base on the free portion of their site as a significant competitive advantage over all other job posting sites. It is actually in their best interest to make it harder for users to find what they want. Craigslist is not in the business of people selling stuff to each other, they are in the business of attracting enough eyeballs for enough time to their site to click on enough job listings to keep their revenues flowing. They have absolutely no incentive to change or improve. I am pretty sure when they look at their revenues (the ultimate metric), it is actually doing pretty well.

Despite the very basic UI, the Dutch version of Craigslist seems to be doing well based on the hundreds of apartment listings in, for example, Amsterdam:


But if I might inject an opposing data point, nobody seems to use the Brazilian version of Craigslist. Check out Sao Paulo:


Can you believe that there are just 5 apartment listings for a metropolitan area of 19.9 million!

It's incomprehensible to me why Craigslist doesn't work there, especially since all the alternatives involve paying some middleman for listing. The same goes for job sites and "things for sale" sites as well. Brazilians prefer to use some scammy-looking sites loaded with graphics and eye candy and often charging a fee, rather than the vanilla Craigslist.

I'm going to venture a guess that this has much more to do with trust (Brazilians want to feel that someone has vetted the data) than with good/bad UI or web design.

There are various examples of Craiglist just simply not working in various countries. In fact I'm going to go out on a limb and posit that Craigslist really only had a monopoly in the U.S. Brazil uses olx (I think?), Argentina and Mexico among other Latin American countries use mercadolibre where Craigslist is more for baiting foreigners with higher rent prices and international shared accomodation. Australia has gumtree, New Zealand has trademe (and they both use Seek for jobs) and I would use marktplaats in the Netherlands as mentioned above. I forget the one for Spain. The list goes on..

'alister' made some good points up top. Let me get further into the rental side of the subject.

As someone who, on many occasions, has rented in Brazil using classifieds sites, it's a mess. 'We' really, really need a CL-type site that everyone knows and uses. As a Californian, and former user of CL, it boggles my mind why CL-Brazil isn't used. It's there (here), it's usable, simple, and in Portuguese.

Instead, people use OLX (free) a little, Vivastreet (free) a little, EasyQuarto (paid) a little, and a few other sites that are either less popular or only work sometimes. Couchsurfing city-specific groups usually each have a sub-group for renting but a strange occurrence takes place there, almost no one puts the price-point. It's called doing business, paying money for a product or service, yet there's a real aversion to giving up the rental price in one's ad. Luckily, the more popular sites I mentioned make it so one has to list their rental price.

And my final two gripes are that very few people respond to inquiries into any particular ad (I think this is universal though) and a good 60% of ads here in Brazil only want women renters.

It's tough, I tell you...

I suspect CL doesn't fly in some countries mainly because other companies got the jump on them there.

Same here in France. It looks like the only ads on fr.craigslist.com are either posted by americans living in France (they're allin english) or spam. I have never heard of anyone buying/selling stuff on craigslist among my friends/relatives.

The (pretty recent) equivalent would be http://leboncoin.fr, I guess.

That makes sense. I think a lot of folks advocate maps and filter sliders because they over-estimate an average user's comfort level with such advanced features. I think a non-technical person is much more content with spending several times more time just sifting through listings manually instead of having to learn some whiz-bang (in their mind) gadget.

TBH that website isn't much of an improvement.

To be fair to Craigslist, since everyone and their grandma uses it, I'd imagine a lot of users have underpowered computers, outdated browsers, and low bandwidth connections.

I've seen padmapper slow wayyyy down on my i7 macbook with 8GB ram, I can't imagine what it'd be like on grandma's 5 year old budget PC.

Features can be added without alienating users with outdated tech. Just conditionally show the improved site only to users who can run it. It's a poor excuse, especially since Craigslist hasn't even tried to improve the layout of the site, which has no effect on load speed and compatibility.

The same is true of Facebook. That hasn't stopped them from evolving their design.

I don't see how the existence of padmapper could hurt the grandmas that use craigslist.org

Don't discount the speed. Every Craigslist page loads and renders very quickly. The simple, text-only UI helps keep it fast, and that matters.

While the pageload speed may be faster the total speed to find what I'm looking for is drastically slower due to the fact that the interface sucks.

Which paradoxically, seems to be an advantage, because it puts people in a feedback loop where they stay on the site, and every new search provides a tiny bit of reinforcement that builds the habit.

Failing to find what you're looking for because of a frustrating UI is not positive reinforcement. People stay on the site because it is the largest marketplace and they have no alternatives.

That's weird, I've never had any trouble finding what I was looking for on Craigslist. There definitely are ways that they could improve the search interface, but the current implementation works fine for me, and given their user base it works well enough for a lot of people.

It's actually much more variable reinforcement - of course, the most addictive kind.

Each individual search has a low probability of success, but occasionally you'll succeed.

Other good examples of these content-over-form sites are 2ch.net in japan, and 4chan for the western world. (4chan recently updated its HTML for the first time in something like eight years.)

I suppose network effects (and the posting model) make up a large part of their popularity, but there's something to be said for a reliable and functional interface that works well and doesn't change on a whim.

There's a lot to be said for it.

It's constant in hn discourse to put great UI up there and to give advice on it.

Yet if anyone here has sat on a Bloomberg terminal, used Craigslist, or even games like net hack or dwarf fort, it's pretty clear that interface can be damned.

Not in all cases perhaps, agreed. But in these cases, after a while of using it people get faster and develop the muscle memory to become efficient on the site, so the UI fades into the background anyway.

> after a while of using it people get faster and develop the muscle memory to become efficient on the site, so the UI fades into the background anyway.

Jared Spool has said that when you change the user interface - fairly dramatically - you frustrate your regular users because they are now unfamiliar with it and have to restart their learning/experience curve.

"At eBay, they learned the hard way that their users don't like dramatic change. One day, the folks at eBay decided they no longer liked the bright yellow background on many of their pages, so they just changed it to a white background.

Instantly, they started receiving emails from customers, bemoaning the change. So many people complained, that they felt forced to change it back."


"If it works, its good enough?"

Honestly thinking about other sites, digg died because of content, reddit survives because of content; even if someone could improve the interface, its not really going to matter.

It's interesting to consider digg... the canonical reason for its death (CEO replaced within a week, 40% of staff laid off two months later) was Digg-v4, a graphical overhaul that also removed features.

Google is definitely content over interface (most of the time). Facebook... not so much - millions of people join protest petitions against every minor interface change - but the value of having all your friends together is still so much greater than the alternative of changing service, whereas reddit was a short step away for digg users.

4chan definitely remains popular because of content - canv.as, a similar project by the same owner, has nowhere near the traction despite being more demographically targeted for mass-market appeal.

I think the lesson is, if you have a close-enough competitor and a fickle userbase, one screwup is all it takes to push critical mass over the edge. If i ran a large website i would study the Digg-v4 case very carefully. Going back to Craigslist, i don't think they have to worry too much about losing their top spot just yet.

> I would be willing to bet a large amount of money that adding a map to search results and letting me put an upper limit on the number of bedrooms I'm looking for would not make people less likely to use Craigslist.

So when can we expect to see your show HN: Craiglist done right?

For the housing vertical. It looked like this. http://www.padmapper.com/

Building the right interface is simple. The real problem of building any marketplace is solving the chicken/egg problem of having a lot of buyers and sellers.

Craigslist does not avoid disruption because it has the best product, it avoids disruption because it has a huge marketplace with a lot of buyers and sellers.

What utter bs - padmapper is not "stealing" anything - padmapper is the best method of viewing cl housing listings there is.

I berated Craig over on Quora about this a while back - to say that he doesn't want to break it is bullshit - he is scared and lazy to better the site through even the smallest of updates.

The fact that "this posting has been deleted by its author" and "this post has been flagged for removal" still show up in results is down right infuriating.

Providing simply no way to browse items by picture is also laughable.

The content on CL is what is valuable, and Craig has been rumored to make 25 million per year off the site, and he doesn't want to muck with it because he doesn't want to hurt that income - but while doing so, he is keeping the content locked up in an abhorrent UX.

You just answered your own question. If your project made 25 million per year, would you change it? I know I wouldn't... Additionally, he is keeping the data locked up because that is his prerogative. Why should he go through the time and expense of developing some kind of API if he does not want to do that? If he is happy with how the website looks and works right now, why in the blue blazes should he change it?

> [Craig] is scared and lazy to better the site through even the smallest of updates.

I'd be the same. Why mess with a goose that's already laying more gold eggs than you can use? For public service reasons maybe, but I'm confident CL is a (highly profitable) business foremost.

Bah. Just because improve a web site is hard doesn't mean that you don't have a responsibility to your users to TRY. Craigslist has not. There are ways to experiment in a low-impact way. Hell, they could offer a "plugin" system, so that users could add the padmapper plugin.

People keep building things using Craigslist data that users LOVE and Craigslist keeps shutting them down.

Craigslist is one of the great squandered opportunities for awesomeness on the internet.

Wow, if this is true I've lost a lot of respect for CL. Being scared to make forward changes because one doesn't know the source of success is such a terrible excuse to stop innovating.

You may have lost respect for them but you're still going to use CL. Which kind of proves the point.

People come back because craigslist has monopoly power. Craigslist knows this. They are afraid to release an API that would allow people to experiment with UI because they know that would train people to go to sites other than craigslist (I save hours using padmapper over Craigslist).

If they don't want to change their layout, why not let sites like PadMapper build on it the way that it has been doing? That way, old-fashioned users can keep using just CL and people in need of something slicker can use PM+CL.

I also don't understand why classifieds content is proprietary? Aren't they just facts? Please can someone explain?

Because that would endanger their monopoly.

Does that mean other people should steal the content and profit from it?

Craigslist is unwilling to let Padmapper buy a license to use their content. This isn't as black and white as "stealing content".

Content not available for licensing is a justification for infringement? That doesn't ring true at all for me - exclusive content is often much more valuable to a business than content that's licensed. I doubt anyone would think that goldman sachs coverage being distributed by morgan stanley would fit under some "no harm, no foul" ethos even if Morgan significantly improved on the layout of the pages.

I don't agree with the infringement, I just think that boiling it down to "stealing" is too reductionist. The reality here is a bit more complicated.

Seriously, I think I am the only person who doesn't think it is OK to torrent Game of Thrones. :)

Is there a rock bottom hackernews will reach for leaning on shockingly bad analogies to make a point?

Its their content, just because they won't let you use it doesn't mean you can break in and grab it.

They also don't have an API, and they also don't allow you to pay for a license or API for websites.

Do you have anything you're unwilling to sell? And if so is it cool if I just take it?

This is unfortunately an increasingly common viewpoint. See for example the recent "Take My Money, HBO" site where people justify pirating the content because HBO won't sell episodes/seasons directly to them.

While I agree with you that taking something somebody isn't selling is wrong, it is still a force any content creator must come to terms with if they want to be realistic about 21st century content creation.

The difference here is that CL still has the content. We always end up in the same fruitless discussion with different subjects: software patents, content, digital media. Should you be able to own something if copying it is practically free? The only right answer is "no", but nobody likes it.

What about if instead of taking it I take a picture with my "3D camera" and then have my "3D printer" create one.

Am I still stealing it?

If this is true, then he (Craig Newmark) really needs to read the innovator's dilemma (http://www.amazon.com/The-Innovators-Dilemma-Revolutionary-B...). TLDR: If you allow yourself to become a prisoner to your success, you're actually making your continued success unlikely. You can solve the dilemma by spinning out a smaller, irreverant version of yourself & giving them what they need to destroy the old company.

In craigslist's case, he could include a link on the old site that allows users to upgrade to a more modern version, and have the modern site basically consume an api/scrape the old CL till it gained momentum.

I get what you're saying... But it's hard to argue with a strategy that's been absolutely dominating its industry for over a decade.

The author of this book probabaly didn't create a Craigslist, so how would he know? </ad_hominem>

Padmapper doesn't make money from Craigslist listings.

It also sends traffic back to CL. It shows the summary of the listing, and then when the user is interested, they can click through the pin to the actual Craigslist listing.

Most landlords probably never even realize the traffic first got to CL through Padmapper at all. Whenever I'm asked, I automatically just say CL. Bringing up Padmapper is just likely to confuse.

CL is deliberately making its service less useful both to its paying customers and the public at large. And the sole justification for this is fear that Padmapper could grow to be a bigger, better competitor someday in the distant future.

Leveraging a monopoly to destroy technological progress for fear of future competition is irksome.

It's almost certainly not illegal, and maybe not even morally wrong. But it still sucks.

Yet for all that, I'm moving on July 15th, and I think there's about a 95% chance I'll find the place through CL. That's where the market is.

How does "Leveraging a monopoly to destroy technological progress" not fall under some antitrust laws? Is it just because Craigslist is free? I definitely feel the consumer is potentially losing out due to Craigslist's monopolistic practices.

CL isn't US Steel. I doubt Craigslist even has a straight 40% of rental market share. But for urban 20-somethings looking for cheap living situations and roommates? It feels like it has 95% market share. At least in Portland.

This is all theorycraft though. If anyone had hard numbers on CL's market share for rentals I'd love to see them!

You would think that they'd be able to measure these things. How hard can your flavor of a/b testing (or bandit testing for the bros) be to implement for a company like Craig's list.

It's just pure laziness and I'm sure the right talented team could make a product that beats Craig's list

I would still conjecture that Craigslist succeeded in spite of their design, not because of it. That's the problem with these things, it's really hard to reverse engineer a success like Craigslist and figure out which of the moving pieces were real contributors.

>he's scared to make drastic changes to the UX specifically

a/b test

Then open an API and let OTHERS do it at their expense...

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact