Hacker News new | comments | show | ask | jobs | submit login
Bot wars - The arms race of restaurant reservations in SF (diogomonica.com)
166 points by RonileSille13 1309 days ago | hide | past | web | 159 comments | favorite

Soon - "Auction Sniper now supports urbanspoon! First 3 reservations are free!"

There's something frustrating about reading this. Perhaps its that a technological advancement that was supposed to make life easier is being capitalized on by a small group of technically competent people, shutting out the average user.

I was thinking of going to PAX next year a few hours ago and decided to do a bit of research. For this year, the 4 day tickets apparently sold out in 23 minutes, and the rest within 5 hours. I'm having a hard time imagining how its going to be possible for anyone to get some this coming year without a bot.

Yes, in todays technological society, people who understand technology have advantages. I don't see why that is surprising, given the website you're commenting on.

Ultimately this is not about technology, it's about economics. Surely the restaurant should raise their prices until not enough people want to book 2 months in advance to make bots and make this an issue. Or PAX should raise their prices so they don't sell out in 23 minutes but still sell all their tickets before the show. Or either business should increase their capacity. Or perhaps they are capitalising on the fact that they are oversubscribed as some kind of marketing device. Either way, any time there is a limited supply of something people are going to figure out ways to get some, and if the limited supply is available on the web, people who understand the web will have an advantage.

"Just raise the price" is an appropriate response to demand for a commodity, but it's an inappropriate response to these scenarios.

For example: PAX is trying to build a community event. The target demographic is not "wealthy people who love games", it's "people who love games" Changing the price results in a different convention.

Similarly, a popular $20 restaurant can't double their prices to keep demand manageable, because a great $20 dinner that costs $40 will reduce the perceived quality of the restaurant.

If PAX or anyone else wants to intentionally price something so that the demand is larger than available supply, then in any case the "effective price" will be raised so that supply matches the demand.

In effect, part of the "effective price" will be the monetary cost, and the other part of the price will be wasted time/emotions/connections/whatever that discrimiminates those who get the ticket from those who don't. This is basic, well researched economics, with plenty of real life examples in various countries, regimes and industries.

The common options include:

a) $x + significant time spent waiting in line - some wait in line and get nothing; some value their time a lot and don't attempt buying.

b) $x goes to the original supplier, $y goes to scalpers/touts/your employee bribes - the stuff gets sold at money market price, but someone else gets the difference;

c) lottery - either intentional lottery, where everybody has a chance of getting a ticket for $x despite that it's market worth is >$x; or an unintentional lottery, where random factors (showing up at the right time, getting a reservation that someone else cancelled) determine if you can or cannot get the ticket. In both cases, the "buyer" might be motivated to resell it if possible, since it's possible that he paid $20, he wouldn't go if it cost $100, but someone else wants to pay $100, so ...

d) (ab)use of advantages - political connections, bots, whatever; you buy at $x and sell at the market price.

In any case, (most) buyers don't get the intended effect of cheaper price; but the seller loses out. Only the middlemen benefit, without adding value. (In one sense, they add back the value that the seller reduced by mispricing)

I've seen what happens in USSR when the majority of economy functions this way - trust me, in most cases raising the price is actually the appropriate response; because otherwise there is so effing large amounts of effort&resources that is wasted - the seller gets $x, but the buyer anyway spends "market price" of $3x to compete with other buyers by giving $x of cash and $2x of wasted, unproductive effort.

It seems to me that the effective price would never fully bring it up to the money market level, and that difference is what makes it worthwhile to choose that strategy?

In the example of (a), there is no middleman, and the set of buyers becomes biased toward those who are dedicated over those with more money, because they intentionally distributed the effective price over multiple...payment channels (? I'm not much of an economist). Edit: actually there is a middleman; I recall the Wii launch, where I knew a guy who waited in line just to sell them on eBay. But still, there are "true fans" waiting who might be priced out at market value.

In the lottery style where tickets go on sale at x time and it's a free-for-all, some significant percentage of this round of sales will go to people that "deserve" them and get to enjoy a reasonable price with negligible non-monetary cost. Yes there will be some money lost to scalpers, but this can be reduced by forcing resellers into an official exchange channel where the original seller gets a cut.

Burning Man is an event where demand has exploded over the past few years and the supply cannot keep up. The organizers are also very concerned about getting the right people the tickets without jacking up the price. They have evolved a somewhat complex multi-pronged approach with lotteries, high-price presales, and a homegrown exchange system (STEP), where I believe they mandate reselling at face value.

edit: On second thought, I have neglected the cost of the people who expend time/effort and miss out completely. But, I still think there are benefits to pricing below market value with a sufficiently robust strategy.

There's this thing here in Copenhagen: http://www.goldentable.com/restaurants-copenhagen/

A number of top restaurants (otherwise booked, on a Saturday night, for 2 weeks) auction off a single table, where 50% of the proceeds go to the restaurant, and the rest to a foundation for gastronomical research.

Another approach is to require a credit card and charge almost the price of a full dinner if you don't show up. The 41 in Barcelona goes one step further and makes you pre-purchase the entire meal.

Some guests will spend another $300 on an expensive bottle of wine. I think it's fine to let some people pay extra in advance to be guaranteed to get a table.

Alinea and Next in Chicago both use a pre-purchase ticket approach. They also open up tickets on a monthly basis a couple weeks in advance, and all the reasonable times are gone basically immediately. And both are far from cheap places.

The result of pricing too low for demand is ticket touting, e.g.: http://www.ebay.com/bhp/pax-tickets

I don't understand why the solution, at least in PAX's case, is not to ban bots. That would open up purchasing to regular human users, and since PAX is a ridiculously popular event, tickets would still sell out.

How do you ban bots with out implementing an arcane, clunky, painful, expensive manual system?

PS: your answer isn't allowed to be captcha.

You don't. I'd much rather suffer through a captcha than have the tickets/reservations sell out in seconds.

Yeah, pretty much this. PAX is a huge, widely popular event. It has an entire community built around it. I suspect that the vast majority of people would not mind filling in a captcha if it means they will have a much better chance of hanging out with friends they made in previous years' PAXs.

Perhaps the reason your answer is not allowed to be "captcha" is that there are numerous services that provide manual captcha answers for bots.

Raising prices is the wrong way to address this behavior. Ultimately the restaurant wants to maximize yield, not just average meal price. Open tables are wasted inventory. So the priority should be to stop bots from grabbing reservations for people who aren't that committed to eating at the restaurant (and thereby stealing them from people who would actually go dine.)

A better solution would be to charge for making a reservation, and give that charge back as a credit against the meal. You don't have to raise prices, you increase the probability that someone making a reservation intends to show up, and when you have a no-show, you get compensated for the empty table (and can still give it to a walk-in.)

Restaurants like State Bird are a bit of a poor example, anyway. Regardless of their reservation book they're going to fill all their tables every night anyway right now, and they don't particularly want larger (5+) parties anyway because they're tiny. I think they want prices low enough that they can establish a loyal clientele that will keep them going once they're no longer the "hot spot" in SF. Or maybe (gasp) they like offering their food to a broader range of customers. Chefs tend to be a little more down-to-earth.

There are some places in D.C. that require a credit card to make a reservation and charge for no-shows and people are FURIOUS about it.

On the flip side, I made my first reservation through opentable a few weeks ago, and I was surprised when they DIDN'T ask for any cc info. Granted I was new to the experience, but the first thing that crossed my mind was I could not show up and it wouldn't be a big deal.

Feel like if they market this move a little better, they could convince their loyal customers that this will allow them a better chance at getting reservations.

Only very popular restaurants require a CC on OpenTable. One of the nice things about OT is that it's very easy to cancel (one click) which I suspect also helps restaurants maximize their yield, since they don't have to keep a table open for a reservation that never shows up. Taking a CC may discourage some people from making a reservation so if you're not already at capacity, it may not be a good play.

Sounds logical that only very popular restaurants would require a CC. Me having to take out my credit card is probably a friction most restaurants want to avoid.

Interesting 20 minute podcast on this topic:

Kid Rock Takes On The Scalpers




  1) More shows.
  2) Beat the scalpers at their own game.
  3) Go paperless.
  4) Don't sell the first two rows.

Funny how the rational, clear-headed solution to economic problems always turns out to be "fuck poor people."

Rational perhaps, but not clear-headed. It assumes that people's dining behavior is a relatively linear, simple relationship heavily based on price, in a perfectly efficient market filled with perfectly efficient agents who have perfect information.

This is sort of like estimating the weight of a cow by assuming it's perfect sphere with a radius.

But this isn't entirely surprising. It's a common sentiment on Hacker News, where highly complex markets are reduced to incredibly simple single-variable, Econ-101-stye supply/demand curves. The poster usually then goes on to believe they've solved some major problem and wonders how everyone else can be so blind as to see such a simple solution.

But far be it to leave at just snarkiness. There are a whole slew of reasons why pricing as a lever cannot be used to fine-tune demand. Simply off the top of my head:

- Desirability of restaurants is heavily determined by their pricing context. That hole in the wall Thai place with the IKEA chairs is a great place to get $5 Pad Thai, and is always crowded. If they charged $15 the expectations for service, food, decor, location, etc, would be entirely different. Cannot simply pull the price lever.

- Desirability of restaurants is heavily determined by perceived authenticity. That Chinese restaurant where all the Chinese immigrants eat is popular with everyone because it communicates authenticity. It loses that authenticity if it prices itself out of reach of the people who lend it authenticity. Funnily enough, authenticity is frequently derived from the working-class everyman. Cannot simply pull the price lever.

- Restaurants are divided into price ranges, with not a lot of intermediate prices. This limits the "resolution" to which you can tune your pricing before you jump into the next range (which comes with expectations, authenticity, and other issues as above). There are demand cliffs moving from one range to the next - in other words restaurant pricing is not a continuous function and cannot be modeled as such. It's not even a "sorta kinda linear" curve.

- As price increases demand does not drop off linearly, or close to linearly. If you've got a line out the door for $5 burritos, a $6 burrito may cause only a small dropoff in customers, while a $7 burrito may cause your business to largely dry up. The holy grail of "that price where everyone who wants a burrito can get one without waiting very long" may not be achievable by simply pulling the price lever.

Pricing food is complex. Until the day where all food becomes standardized, commoditized, and we can treat our restaurants like we treat our nuts-and-bolts supplier, it will remain like this.

It's a common sentiment on Hacker News, where highly complex markets are reduced to incredibly simple single-variable, Econ-101-stye supply/demand curves.

That's a strange phenomenon I've also noticed, since posters here usually have the opposite instinct in any other field. Most systems end up complex and difficult to model by simple rules, unless you restrict your model to a specific range of parameters or behavioral regime. Do you want to know how increasing or decreasing pressure affects fluid flow? Most knowledgeable people would immediately say that it depends: does the pressure change impact the flow regime? If it stays within a flow regime, there is one answer. But if it changes from laminar to turbulent flow or vice-versa, the answer may differ. What initially seemed like a simple relationship has a confounding factor of macro-scale regime-switching.

But in economics, HNers feel somehow confident just giving a blanket answer without analyzing the situation. Price and demand of restaurants are linearly related; raising minimum wages increases unemployment; etc. Blanket statements they'd never make in an unqualified form in fields they actually know about, somehow become fine in a field they haven't studied. Nobody seems to feel the need to inquire, well, is it a simple relationship, or can changes in a parameter change the operating regime? Very strange.

> Until the day where all food becomes standardized, commoditized, and we can treat our restaurants like we treat our nuts-and-bolts supplier, it will remain like this.

And after this, the minute someone opens a restaurant where food is not standardized, commoditized and that can't be treated like a nuts-and-bolts supplier... a market is (re)born.

I totally agree with you that it's an enormously complex system, but I wonder how risky a tiny bit of experimentation would be. Maybe it's worth upping everything by 1 dollar just to see? That would make a huge difference to profits. But maybe the hip customer base would notice the change and see it as a signal that their eatery has become bourgeois?

I have to disagree with your point 2: The blue-collar working man doesn't want to queue with a bunch of students for half an hour for his food, however authentic it may be.

In order to determine who wants X the most, you have to ask credible commitments. Getting up at 4AM to make a booking when it becomes available; or waiting in a three-hour queue; or spending time writing a bot is just as significant commitment as paying some extra dollars, so all of those will work to allocate the scarce thing.

But there is a major difference - if you spend three hours in a queue, you lost those hours and noone gained from that. If you spend three hours writing a sniping bot, noone really needs it. But instead if you spend those extra three hours even sweeping streets at minimum wage, and pay that extra money for more expensive tickets - then the seller has gained extra money, which (on scale) increases supply and the world has gained some useful work.

One of the neat things about dollars is that they are convertible into [someone else] getting up at 4AM, [someone else] waiting in a three-hour queue, and [someone else] writing a bot.

That's why you can hire a homeless to do it for you. At Disney World, rich families are actually hiring officially disabled people to get them line skip passes so that they can jump the lines.

In the olden days, you didn't need to be "officially" disabled, you could just have your brother sit in a wheelchair.

...or adapt the site so it's not bot friendly instead of raising prices just to punish people.

Raising prices isn't just punishing people - it also increases supply, since if PAX would gain double revenue then it would incentivize both PAX to obtain larger facilities, and others to make more similar conventions more often so that more of people who want to go can actually do so.

Except that PAX is a community event and you don't trash your community by pricing its members out. Not if you're human, anyway.

But the result of more people going is a side effect of raising prices; it shouldn't be the main reason. In fact if anything, prices are usually raised to decrease the number of people who pay out for it (ie exclusivity).

The issue here is bots, so the solution should be to make such reservation / booking sites unfriendly to bots. Then the event organisers / restaurants can choose to adjust their prices based on the actual demand and their long term vision instead of trying to play some silly price war with ticket touts.

The issue here isn't bots, the issue is that there are a lot of people who want to go, and they can't decide which ones will stay home (as there isn't enough space for all of them). You need to decide on some kind of system to decide who will get to come.

You can draw randomly, you can determine who you want to come and invite them, or you can determine who want it the most by asking them to pony up a lot of resources - money or waitingtime (live queues) or effort. It's your choice, unless you refuse to decide, and let scalpers/botwriters decide for you and take everyones money for that.

The issue here is exactly bots. The problem is people who have bots are gaining an unfair advantage over those who do not have bots. But raising prices, all you're doing is shifting that unfair advantage from people with IT skills to people with a higher disposable income. Thus you're not solving anything.

Where I work (PagerDuty), we follow up with everyone who cancels -- my favourite was "I only signed up to be alerted by a script as soon as PAX tickets went on sale" He got them.

There's probably a whole market of "notify me when (plane tickets get cheap, apartments look good, tickets go on sale, etc)"

If Amazon had a notification when product X dropped below price Y then not only would it allow people to track that something was now affordable to them, it would allow Amazon to feed back to manufacturers that they could sell N items if they dropped their price to Y.

Something like this? http://camelcamelcamel.com


I've used this service in the past: https://pricenoia.com/ it can notify you of price change on Amazon AND allows you to compare prices between different countries's Amazon local sites

Isn't there a risk that this type of notification would break the market segmentation that they're aiming for? If you make discounts too easy, maybe you'll start pulling in people who would have paid the full price.

I work for an online retailer, and every once in a while, we get an email from a customer asking if we could add a feature on the website to notify them when prices drop on items.

Maybe we should start trial runs of this feature..

You could even get some additional engagement out of them:

o Send me an email if the price drops below $_____

o Send me an email and put 1 item in my shopping cart for me if the price drops below $_____

Yeah, that kind of feature is getting built into a lot of price comparison based sites now.

I think this site is similar to what your describing: https://trackif.com/

One could always use a Captcha to stop bots if it became a problem.

But it's not a problem for the places taking the reservations. They're thrilled whenever they manage to book solid. It's only a problem for the non technically minded people looking for reservations.

Are all the bot reservations actually made with the intention of a human going to the restaurant? I wouldn't be surprised if some people used these bots just to prevent anyone else from making online reservations and screw with the restaurant/customers.

Based on the bot's code [1], it looks like all they ask for is a phone number and an email address. It might be a good idea to at least require a captcha, if not a credit card deposit of a few bucks.

[1] https://gist.github.com/diogomonica/6076911

A service that solves captcha for you can cost as low as $2/1000 captcha

Yes, but the services utilize human interaction to solve the captcha and that would almost certainly prevent all the openings from being taken in <1 minute.

Here's the thing though: you've got two people who are eagerly waiting for the purchase / booking of something to become available. One of them is going to do it all manually. The 2nd person has a bot to do all the form filling in leg work but will manually solve the captcha.

Who wins?

True, but there are multiple openings available. I guess they're probably already executing in parallel though. When I looked earlier the DBC average was 17 seconds... with browser form fillers you could beat that!

You folks don't actually think that opentable/urban spoon reservation availability is representative of a restaurant's entire table inventory on a given night do you? Pick up a phone (dirty little secret, restaurants segment customers too).

Restaurants typically hold back inventory from online reservation systems for two reasons. They either think they can fill seats and they rather not pay the per cover fees OT and UrbanSpoon charge or they want to hold seats for walk-ins. Well three if you count timing seating to keep the front of the house and kitchen from being overloaded.

State Bird Provisions however has so much demand that there is a line for walk-ins hoping for a cancellation an hour before the restaurant opens and they appear to not actually hold back much (any?) inventory specifically for walk-ins.

But generally speaking yes, you will often be able to call a restaurant and get a table or possibly walk in even when their online reservation system shows up as full.

Also, you can also put yourself on a waitlist at many restaurants. If someone calls to cancel, before they release the reservation back to the online system, they'll go through their waitlist and call and ask if you want it instead.

From their site: "We welcome walk-ins on a first-come, first-served basis and set aside about a third of our seats for those guests every night, including our chef’s counter and several tables in the dining room. "

I tried that once, very shortly after open on a Wednesday. The wait was 2-3 hours. It's a popular place.

trust me. this man knows his stuff when it comes to reservations. :-)

hi jordan.

Or walk in and slip them a $20. It's the old school hack.

also see: be a regular and a good tipper.

(only works in places where you are a regular and a good tipper)

(but it works amazingly, wonderfully well in those places where it does work).


You can also nook from the phone a couple of days in advance usually.

Ah, but then you need the correct phone number...

I'd suggest that this effect is a somewhat superficial expression of the deeply rooted problem of people in urban situations broadly losing their personal connections with their own expressions of creativity and art, primarily through doing work that they are not engaged with in a deep, meaningful sense.


- People in certain modern, urban situations crave certain types of creative expression that seem to be very hard to find in these environments.

- Social patterns place a high value on being seen to have had an experience, that provide a reward similar to being the artist.

- People capable of producing these artistic experiences are initially perhaps not so interested in managing the people who are the recipients of their creativity - they are more interested in creation.

- Once money and/or fame becomes involved, the artists are unlikely to widely share the skills that lead to the creativity arising, are motivated to protect methods of production, fetishize the product and create strong brands.

These factors together create this 'hipster' effect where small points of creativity are heavily focussed on and take a long time to replicate.

I'd hypothesize that this situation would go away if people were broadly connecting with meaningful, personal creative pursuits. Or, if the artists flipped their model once they smelled success and were motivated to share everything they were doing, allowing the experience to be rapidly replicated.

Wrong thread?

I love how everyone else who fights for reservations is a filthy hipster, while the author is merely a humble gourmand.

Great quote from a parallel discussion on Metafilter:

"A hipster is someone who enjoys the things I do, but in ways I disapprove of"

Given their popularity, why doesn't this restaurant just get rid of the reservation system altogether? A lot of the popular/trendy/hipster magnet places here do this since it saves hassle and creates a certain allure when lines are out the door for a table.

Given this place has a line outside of it well before it opens on the off chance someone cancels, removing reservations would mean everyone would have to wait for a table.

I can't think of a worse customer experience than taking someone on a date, waiting in a congested line outside at night for 45 minutes only to be told that there are no tables available.

Not getting a reservation on the other hand just means that I can just go somewhere else.

Seriously? I thought when most people saw a line at a restaurant, they just went somewhere else.

Perhaps the restaurant should raise its prices.

Indeed this I do not understand, nor do I understand why bands sell tickets so cheaply when they can be resold by touts outside the door for many times that price.

Kid Rock is tired of scalpers taking tickets away from his biggest fans.

One way to stop that: Raise ticket prices. If Kid Rock charged more for his tickets, scalpers wouldn't be able to sell them at such a big markup.

But Kid Rock doesn't want to raise prices.

"I don't want to break you by coming to see me, " he says. "I want to make as much money as I can, but I don't need to drive around in a tinted down Rolls-Royce or Maybach and hide from people because I felt like I ripped them off."


That interview was a great summary of what's going on in the industry. Kid Rock realizes that keeping his die-hard fans happy works out better for him on the tour circuit, since they'll spend more on t-shirts and other things while they're at the show. If they have to pay $400 for a ticket, there's no budget left for a high-margin souvenir.

It's the internet age as well and he realizes that. The moment he tries to charge a market price for every ticket, he starts eroding his fan base.

I'll leave inviting the Market Economics Fairy here to Eliezer; I just want to point out, that he won't "break you by coming to him", you just won't come. Instead, someone richer, who can afford the ticket without breaking, will come.


The real ideas of Kid are described not in the quote but in the rest of article; it's worth taking a look at.

Having to show your id and the credit card you used to buy the tickets makes the most sense to me.


Because people aren't generally motivated purely by money? Surely it's not hard to imagine why a band, once they're making enough money, might also want to consider factors like letting in the people who love them the most, regardless of whether they happen to be loaded?

Surely it wouldn't surprise you, for example, if someone accepted a lower salary for a job that they enjoyed more than a better paying job?

Do restaurant's scale?

Find a bigger venue. Open another store.

Maybe they should just deploy to Heroku or Appengine and have them scale to meet their table needs.

Unfortunately scaling good restaurants is hard. Its a service business not a product business. It can be done to some extent.

The demand doesn't scale, this is a fashion market where new restaurants are in incredible demand until ... the next hot new restaurant gets it's day in the sun.

I think a lot of the business is probably generated because they are sold out all the time. Raising pieces threatens that.

Or partner with Urbanspoon to charge a reservation fee based on market prices that won't affect menu prices. It's basically a cover charge.

Or open a second location.

Restaurants don't scale linearly.


Sure, that's one kind of restaurant. I've never heard of anyone trying to get a reservation there, though, and I'm pretty sure there's a direct relationship between the lack of demand for reservations, and the restaurant's ability to scale.

If they only had one restaurant, over 99 billion wouldn't have been served, and a good fraction of them would probably be on the waitlist for a reservation.

If they only had one restaurant because they were incapable of scaling, then perhaps people would get reservations there.... ...but that is not the sort of restaurant they are. They are a restaurant that can scale, and if they only had a single location despite that, they would not have a waitlist.

Or to put it more crudely: If your aunt had balls, she'd be your uncle.

I like Chipotle better than most restaurants.

Enough to bother getting a dinner reservation?

I doubt it; I doubt many would do that at all. In some situations corporate fast-food joints do take reservations (White Castle on Valentine's Day evening is the only example I can think of at the moment) but in those situations they are cashing in on people considering it a novelty. The novelty drives demand to levels that it otherwise never achieves.

If there were only a single Chipotle in NYC, there would almost certainly not be enough demand to require a reservation system (and if they used one anyway, their business would surely suffer). Perhaps you would be willing to call in a reservation for Chipotle, but one person calling in reservations for fast-food isn't going to sustain a business.

Have you had the McRib? It's in this limbo state of being popular enough to warrant introduction into the McDonald's menu periodically, but not popular enough to be a permanent staple.

Basically people will especially go out and get the McRib when it "comes back" and then get tired of it.

If McDonalds was limited to only one location, there would almost certainly be long lines, waitlists, and reservations. The ubiquity of McDonalds, and the fact that they can meet demand enough so that no one needs to wait more than a few minutes contributes to the luxury of not having to wait for a Big Mac.

Sorry. Restaurants serving food fit for human consumption don't scale.

You essentially need a completely new staff--including a completely new creative staff. This is important, and a large part of why most chefs who own multiple restaurants generally have fairly diverse properties that present different menus. But diversification still requires wheelbarrows of money; "the fastest way to make half a million dollars is to start a restaurant with a million" probably applies even more acutely to opening more than one.

Just because there's a large one doesn't mean they scale linearly.

This is remarkable.

Concert tickets are also being bought this way by bots for scalpers.

Ebay sniper bots are commonly used.

High Frequency trading is the high end this.

Perhaps it is unavoidable, perhaps other means for selling things online should be sought out, text messages with semantic replies required or something?

This is why we can't have nice things.

On the topic of scalpers, Louis CK apparently hired [0] scalpers to help build some anti-scalper checks (I assume it is similar to fraud detection rules: multiple purchases from the same IP, out of town billing addresses) that runs on his website where he now sells all of his tickets. If you get flagged by a rule, the ticket goes to will-call, so you have to show ID and the credit card at the event. That small step to prevent it significantly reduces the resale value for a scalper who can't attend the event.

Here's the interesting part: It worked! Roughly 25% of all tickets for major shows fall into the hands of scalpers. On Louis CK's last tour, of all 125k+ tickets he sold, than than 1% were scalped. [1]

What's really terrible about the whole scalping business is that it would be solved, but only by the people who care least: The venue has a perverse incentive allowing scalpers to buy every last ticket they can get their greedy little hands on. As long as a ticket gets sold, they don't care. Screw the fans.

Possibly NSFW language in these links. [0] http://www.wbur.org/npr/162514765/louis-c-k [1] http://boingboing.net/2012/07/06/louis-cks-direct-sales-conc...

> The venue has a perverse incentive allowing scalpers to buy every last ticket they can get their greedy little hands on. As long as a ticket gets sold, they don't care. Screw the fans.

Everything I've seen trying to explain the scalping phenomenon appeals to the idea that this is completely backwards from true. Specifically, the theory is that the venue's incentive is to fill seats, which is why tickets are so dramatically underpriced (price too low -> shortage of seats -> all seats get filled -> the act appears to be popular). From the venue's perspective, a seat sold to a scalper is a seat they lose money on and that might not be filled during the performance (since the scalper's incentive is to charge a realistic price), which is a double loss to the venue.

In sum, selling tickets to scalpers gets the venues an amount of money they don't want (they were guaranteed to sell out anyway), in order to generate an effect they don't want (some seats will be empty during the performance). Where's the perverse incentive?

Exactly, Scalpers are just buying "seat futures". If venues are selling these future contracts so low that they yield absurd profits, don't blame the scalpers...

I really don't get the outrage against scalpers. They're just buying commodity futures, taking a bet on the future value of the tickets. If the value goes up, they make a profit, if the band suddenly become unpopular or whatever they lose.

We're in a capitalist system, the price you pay the scalpers is the "real" price of the concert: it's the law of supply-and-demand.

Now, if scalpers win every time, this is an issue with the original ticket sellers, not the scalpers!

If Shell started selling petrol at half standard price and some guy comes round to fill up a tank and then sell it to people at 95% standard price, that's Shell's mistake, not evil behaviour on his part.

To the average fan, the comparison between a ticket to see Act X on a specific date in their town is nowhere near the same as pork bellies or orange juice. They're called 'commodities' because they're interchangeable. Tickets are extremely specific to a single date and location and act.

Good point, I guess tickets would be "infungible commodities"... But the fact that you can take market bets on them doesn't change

The outrage at scalpers relates to how people feel what something is 'worth' against what you pay for it.

The existence of scalpers is curious. Artists could get the money they do. There are reasons why they still exist, one of them being that artists use scalpers to get cash in hand.

There is a really good episode of Planet Money that looks at why there are scalpers, how they work and what can be done by artists if they don't want them to make money:


They are creating monopolies (cartels) and creating artificial demand by cornering the market. Things most capitalist systems regulate/limit.

They are speculators. People have instinctive dislike for speculators as they produce nothing and in scalpers case provide no benefit (such as liquidity) to society.

Scalpers do provide a benefit though: If you have more money than time, scalpers enable you to easily buy the best tickets, those which would probably have been sold out within a few hours if everyone could buy them at the artificially low prices set by the venue.

Speculation is just another word for betting, I don't think that's an inherently bad thing.

That's why I stopped buying things via auctions and stick only to "buy now" solutions. Sniper bots combined with friends bumping up prices on auctions means that a typical auction is a waste of money (price bumped up) and time (you can't win with 10 sniper bots aiming for the very millisecond the auction ends).

Yeah, playfield got leveled because now everybody and their dog can use snipers (hey, where's the reward for technical competence?). So the "game" looks similar to before automation, but is more complicated. A positive feedback loop of increasing complexity if you like. For me, "the only winning move is not to play". </rant>

>Sniper bots combined with friends bumping up prices on auctions means that a typical auction is a waste of money

This doesn't make sense, really. You enter the price you're willing to pay in an auction; what everyone else bids is irrelevant to you.

If you are getting outbid, it's because someone is willing to spend more money on the item, bot or not. Don't bid low trying to get a deal. Bid what you are willing to pay, and it works every time. It's a very efficient system.

If I wanted to I could just as well buy the item in a shop. The entire point of an auction is to get a good deal.

I hate auctions now, too. Instead I just shop around for the best deal.

Agreed. My bigger problem with auctions is that you can rarely get a deal that's worth the added risk.

The "solution" to sniping is only in eBay's hands; either sealed-bid auctions (which they effectively are today) or a policy of extending the auction one hour past the last bid.

Few years ago I wrote sth similar for registering to popular courses at my university. A lot of CS students gets an edge doing that.

Moreover, I have also participated in a few lotteries for various things (e.g. Half Dome permit), but did that completely by the book.

A few things that help make system fair (TM):

- price them right, maybe even charge a bit premium for reservation vs. walk-ins

- make cancelation non-free and charge at reservation time, even a few dollars prevents spurious reservations "just in case"

- have a waiting list, if there is a cancellation, you can automatically sell it to someone else

- do some load management, e.g. weekend prices, discounts...

I bet this bot race is doing a great deal to improve relations between San Francisco's tech population and...well, everyone else in SF.

If an establishment wants to give non-techies the opportunity to attend, yet remain "accessible" to those without piles of money, there are a few solutions.

1. Don't take reservations.

My favourite restaurant operates on this basis. On a Friday or Saturday night you will queue for up to an hour for a table. If you prefer to avoid peak hour, there is rarely any queue for lunch. I've eaten there over 100 times and I still don't begrudge the time spent in the queue.

2. Run a lottery

Popular theatrical productions are great at this. If you don't want to book months ahead (or can't afford to spend hundreds of dollars), Book of Mormon offers a ticket lottery before every show. Just show up at the theatre and enter for your chance to win one of 21 discounted, front-row tickets.

To clarify, I'm not suggesting that the establishment in the OP wants to let the poor and/or technological illiterate to attend. However other places have managed to do so without hurting their brand.

If I can't be sure to get a table/ticket I may come, but I am not bringing a date or a friend and risking ruining our evening.

But surely this is a good thing - people will have to take part in an auction for a table.

Want to eat at $POPULARRESTAURANT - you set a bid min and max and off you go.

Its allocating scarce resource at the most efficient level.

It's possible that your patrons are not interested in bidding, or otherwise participating in libertarian economic fetishism.

Have you seen people trying to get in for a top NY restaurant - they will participate in any fetishism :-)

You're assuming restaurants look at it as a one-time transaction. But most good owners put a lot of focus on repeat, neighborhood customers, because those are the folks that will come in and dine on a Tuesday or Wednesday evening and help keep you afloat (Friday / Saturday will book out anyway if you're doing ok.) Bourdain said he used to put the interesting specials on weekdays, and save the easy / no-brainer stuff for weekends, partially for this reason.

Seems like Urbanspoon should be using some sort of scraping counter measures to make things a little more fair. I sent them an e-mail through their contact form (of course I wouldn't be suprised if those are sent straight to the trash/spam folder).

Also, is scraping legal? I would imagine, at the very least, the OP and others are violating the TOS. I have written a few scripts of my own for fun, but I would be careful about bragging about it on a public blog.

Well, looks like they do actually respond to the contact form, got an e-mail from them a few moments ago saying they were "aware of the situation". Although, looks like this blog has been picked up by some mainstream news outlets, who have been getting comments from Urbanspoon:


Since you can't RESELL your reservation, I think the answer is for the restaurant to black list exactly one sniper, maybe just for one year, and publicize doing so.

Everyone else, who has been fighting to get a reservation for months the old fashion way, will not decry your loss.

And I doubt very much you would risk getting blacklisted for sniping reservations off their site, since you appear to love the restaurant so much.

What am I missing? Why would the restaurants care? Do they prefer non snipers eating there as opposed to snipers? Won't they just be happy to be booked solid regardless of who the bookings are with? I just don't see how this is a problem for the restaurants at all.

I bet you could cancel a fake reservation one second before making a reservation on behalf of someone else. The effect would be the same as reselling.

Honestly though, I don't understand why this needs to be available over the web in the first place. If your restaurant is that posh, I'm sure you're not going to lose business by making people call in their reservations. You'll still need to have someone answering the phone anyway, so it's not like you're saving money by cutting an employee. Whatever meaningful benefit having an online reservation system might have provided is clearly lost in the absurdity of this situation.

If you can't resell your reservation, then what's the point of the bots?

Some of the bot-herders probably just want to eat there themselves. After all, the author wrote his bot because he wanted a table.

The people saying "raise the prices" don't realize that the scarcity of the restaurant is an asset in itself. Restaurants become famous for being impossible to get a table at. The chef is more respected by his or her peers and the media. Newspapers write articles about the restaurant. The restaurant's cookbook starts selling more copies on Amazon. The chef is invited to an interview on Food Network. None of these things happen if they raise the prices until there's no excess demand.

Ferrari also has a long-standing policy of manufacturing "one fewer car than the market demands."

In London, a "no reservations" system seems to be happening more and more with popular trendy/hip establishments.

And it seems to work.

Places like Burger & Lobster or Meat Liquor usually have very long queues at prime times.

Does anybody but the establishment really prefer such a system though? Setting aside time in my evening itinerary for queueing is not my idea of ideal.

Probably only a matter of time until a side-business of people who are willing to wait in lines for you emerges.

As a customer it can definitely be annoying to experience at times. My tolerance seems to vary depending on the wait times and food expectations.

A perceived upshot is that the restaurant must put an emphasis on quality at reasonable prices, otherwise they won't draw those crowds and I'm definitely not going to wait for a crappy meal.

However, the "no reservation" method itself seems to help draw crowds too. Kind of like the lines outside a night club seem to draw people. Throw in a bit of edgy branding with some cool marketing and it pulls well - having an "image" works.

Queuing is a bit of a cultural thing/expectation over here. Having someone cut in line can be a huge no no. If someone bought their way into the queue, I could very well see it causing some highly vocal responses. But I'm sure there are people out there trying it out.


Reminds of the tickets to "walk" up Half Dome (Yosemite park). After some folks died a few years back, the number of folks is limited and you have to get a ticket.

All tickets for the year are gone on the first day they become available and then lo and behold you can suddenly buy the tickets (for a considerable markup) elsewhere. The folks scooping up the tickets like this are parasites. The solution is to limit the number of tickets per person or to disallow reselling.

(Haven't tried for a while out of frustration, might be different now)

I know, slightly tangential to the article.

How are the bot owners flipping these reservations? I imagine there must be some commercial value via secondary market (or direct to concierges, offline), as bot-writers probably aren't actually using all these reservations themselves (or their friends). The more sane solution is to require full name when booking and ask for ID at the door. A name change has to be manually handled by the restaurant or would require rebooking. That would stop transfers and any market for these bookings.

It could be that a few savvy people always want a reservation or two on deck. They queue up reservations, say one table per week for two months. When that week comes up, if they don't want the table they cancel it. If they do end up having a need for it, how cool does it look to be able to tell your date that you can get a table at their favorite hot restaurant tomorrow?

The website is unreadable on a mobile device. I can't zoom out.

It's a truth universally acknowledged, that a restaurant in possession of a good name, must be in want of a good web designer.

Having worked at Ticketmaster for a few years, the first thing that came to mind was the similarities in automated ticket-buying bots.

Thing is, is there really that much of an arbitrage opportunity with restaurant reservations? Is there a place I can go purchase restaurant reservations?

My experience with restaurants is that if one is too popular, just wait a little while. Eventually the mob's attention will turn to the next heretofore-undiscovered gem, and you'll find it's much easier to finally check out the previous one.

I don't understand why they don't do authentication to remove the bots if this really matters.

If it doesn't matter, they should raise prices.

This is an issue in NYC. Some $300/night restaurants have you hyperclicking the moment the reservations open up.

Priceless! When I studied history I learned how to recognize a significant historical document, and this definitely qualifies as one. I wonder what historians reading this two-hundred years from now would think of our culture.

Next step: high frequency trading techniques applied to reservation making

Ahh, that would be hilarious. Urbanspoon should open up a co-location server rack and charge people a monthly fee to set up their servers there.

just as information, I use the free service http://www.changedetection.com/ to send me email when webpages change (i.e. jobs opportunities). Unfortunately you can't automate the procedure of logging in and/or posting to searches, and I think the scheduling is once-a-day, so it wouldn't have been very useful for the OP cron case

Cool concept. Wonder if he could write a script to land a reservation at Dorsia. I think that might be asking a bit much, even for a bot.

A perfect place to implement a captcha.

Or the equivalent of a shadow ban. Once a bot is detected, always return "no reservations available".

Next step in escalation: make a bot that books with fake names, sell primo reservations on Craigslist.

Next week: Co-location

After reading that I want 10 mins of my life back.

tangential but has anyone been to state bird provisions? I keep meaning to go, but don't because of the reservation saga.


I'd try:

- detect obvious bots (e.g. clients requesting form without loading CSS or JS or filling it in instantly), suspicious IPs (EC2, VPS hosting) and release smaller batch of seats to them (don't deny completely in case of false positive).

- create waiting list and give reservations randomly to people on the list (while the list can be spammed as well, humans would still have better chance than reserving between 4:00:00 and 4:00:01). Also bot authors would have to be creative to avoid creating easy to spot patterns of names/times in the list.

Site renders horribly on an ipad

Do people actually use "foodie" unironically? I thought that was something people invented to stereotype suburbanites who'd caught on to "gourmet" meaning absolutely nothing; I didn't know until now that people actually called themselves "foodies". Wow, it's a weird world out there.

It's certainly pretty common in the UK. Seems much less linguistically dubious than "gastropub".

(Why do "suburbanites" seem to get so much stick nowadays?)

Well you have to have some way to differentiate them from the people that don't like food.

Yes, definitely, people use it unironically.

> This means that I'm bound to compete for reservations at good restaurants with the the hipsters

if this restaurant is so popular, isn't it too mainstream for hipsters? /philosoraptor

Please for the love of god keep the shitty memes on reddit.

sorry, forgot to apply the stick up my this morning, so I couldn't appreciate the seriousness of foodie reservations in sf.

Non-hipster technique would have included invoking Yogi Berra and his "Nobody goes there anymore. It's too crowded." line.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact