Magic: The Story of an Accidentally Founded, Wildly Viral Startup
wired.com
wired.com
> I’ve never once actually gotten what I wanted from Magic.
So we should be celebrating a "company" that takes almost a hundred bucks from this guy but doesn't deliver him anything at all. That's not a company or a service it's a walking disaster.
The people that are having sour grapes (if they exist at all) presumably have them because they work on startups that actually do things or solve problems and have some small measure of success but don't get Wired articles because they don't happen to be connected to YC dinners.
What about all the "ideas are worth nothing" dogma that permeates all of the startup literature? What about it's the execution that matters? Doesn't "product-market fit" require an actual product?
How is texting a pizza order to Magic, who then text it to a pizza company (Introducing a middle man) making the world a better place?
It's so well suited to startup founders though - Hey I'm way too busy and important to do menial things, I need a personal assistant on my phone that can do things for me while I'm saving the world by making some crappy website, which will never make any money, but will get bought for billions by some bigger fool because we know people who know people.
Magic sounds like it's right out of "Silicon Valley" (The TV Show).
I'd be surprised if anyone remembers Magic in a years time. Either that or it'll get bought by Amazon for $20bn. Or maybe both...
Exactly. If they were't connected to YC or someone else that could grease the wheels they would have their merchant account and CC acceptance terminated in the first 72 hours for the obvious and flagrant violations outlined in the Wired article, and quietly fade into the ether.
It's the good ol' fashioned trade of money for time, and trading is what makes an economy work.
Call your pizza company and see how much effort it takes you to order a pizza to your house. They don't know your card number, your address, etc.
Compare with magic: "I want a large pepperoni pizza". Confirm the price, you're done. About 20 seconds of time involved.
Assuming your time is worth, I don't know, $20/hour or more, this seems to be worth the money directly--not even accounting for less cognitive load, loss of focus, daily happiness, whatever.
The pizza company sucks at taking your pizza order. Magic doesn't suck at it.
In any event, I usually order pizza online, and it takes a couple of clicks. Even on the phone it's less than a minute.
Do you really consider your time so important that you refuse to even spend one minute on the phone talking to another human being to order a pizza?
FWIW it's more like 5-10 minutes every time I've ever ordered pizza (admittedly not often, I'm sure my pizza skills could improve, if I worked on them. I don't want to.)
Magic doesn't magically know your address/payment, of course, but it remembers it in the future. For every service, including new ones. I'm sure you actually knew that already. Point is it's one-time, ever.
Granted, Magic has that for all transactions, but pizza can already be pretty convenient.
1) You're not at home. 2) You don't always order Dominos. 3) You don't even know what's available for delivery (I'm in a less urban area than most--Pizza Hut just started delivering in the last 12 months, and doesn't deliver to the whole city)
Essentially, when things get complex, Magic's cognitive load hardly changes at all. If it's a matter of routine (same pizza, same company, same house), your way is probably worth saving the money they charge.
Anyway, my pizza place has none of that. Including "online account," my pizza guy would have a good laugh at that.
But I could see a huge value in having someone make these decisions for you when it's something that's far outside of your normal routine. The last time I went skiing for example, we landed in the small ski town at night and needed to get dinner, we ended up calling around to 4 different places and making a bunch of arbitrary decisions and still had no idea if the food would even be good. Someone else (or eventually software) could have done that just as effectively and it would have saved us a lot of stress, in that scenario it really would feel like magic.
From the article: "I’ve never once actually gotten what I wanted from Magic."
"The pizza company's UI for ordering a pizza is far inferior to Magic's."
Whether they get the product delivered or not is most important, of course, but not really what I was talking about.
If they can scale, it won't be an issue, by definition.
One of the UI's is a piece of software. The other is an English speaking human required to be quick and accurate.
Last I checked the slogan wasn't "manually entered orders done by on-demand employees are eating the world"
My regular pizza company is completely painless to order from, both on the phone or online. They know my address, don't need my credit card number because I pay via credit card at the door, and they even have my order history. This is 2015. Maybe you need a new pizza company?
I have every expectation that Magic would handle that order for me, though. Probably slowly, if the article is any indication, but I'm assuming they'll get better.
I wasn't really talking about my local shop, though, so point taken. It'd be pretty enjoyable to get "hey mr. mod!" and I say "hey, give me what i got last time, please." and hang up.
I don't want to bother with paying at the door, but I'd tough it out, I guess.
I mean, you're going to have to tell someone the specifics of your order. It's all the rest of the details where you're saving time.
Local pizza delivery companies vary more widely.
There is some pain points (god I hate myself right now) that magic could solve, but ordering food isn't one of them.
Grubhub or seamless is just easier. It takes longer to tell someone your order correctly that to just to it yourself.
I also do wonder if a pay per use secretary will actually work. My secretary knows me so she can better make choices. Someone with just a chat log has a harder go of it.
In the old days we used to call this a personal assistant. Now an Uber for personal assistants is something that would be really useful because I sure as hell don't have the money or the need for a full time assistant.
Although I have considered hiring somebody on a retainer basis to just be available for a few hours of menial tasks a few times a week or month.
Some business models simply don't scale until some amorphous inflection point in tech/attitudes/etc hits, no matter how much we want them to.
There's a healthy argument to make for raising and burning a bunch of money to aggregate users together in anticipation of the market shortly hitting that scaling inflection point, but it's still a big gamble (albeit one I think Magic and a handful of competitors will all take).
And then ultimately if it can scale, I think Amex and Visa/MC buy up competitors and make it a loss leader/free service attached to their card services.
I predict a lot of new SMS services headed our way. e.g. If I'm EXP on American Airlines, why don't they have a concierge # for me to text with? Why doesn't Seamless have a # for food delivery? etc.
And if there were multiple SMS services for pizza ordering or flight reservations or whatever, then Magic would just subcontract out to those smaller services, and people would still just use Magic.
They're going to drown and the temporary hype will go away (needed to support the volume, which will fall out from under their service just about the time they're trying to raise serious capital), long before they can scale up and deliver on the premise.
Someone else will make a lot of money off of a service like this, but it is not going to be Magic.
edit: Next year, one of us is going to be laughing.
"Users love CD-ROM - it is user friendly." -- http://forge.fh-potsdam.de/~IFLA/INSPEL/94-3laan.pdf
In retrospect, that line sounds like Lucille Ball describing Vitameatavegemin: "It's so tasty, too! Just like candy." -- https://www.youtube.com/watch?v=gmOLpeKwWGI
Every minute that someone's spending on some random, menial task is a minute that someone isn't getting something else more productive done. Maybe the world isn't getting saved, but I find it hard to call a service that gives people more time in their day anything but "making the world a better place".
Also, it's a person spending their minutes on it, who might not have a better job available to them at the time.
Don't get mad because someone is willing to OFFER money for a service that YOU would find demeaning to provide. Just don't take that job. Problem solved. Sheesh.
It's not even a very good argument, in that you're just demonstrating the willful ignorance the previous poster was pointing out.
You dismiss this entire line of reasoning as "willful ignorance," rather than ACTUALLY RESPONDING TO IT AND TELLING ME WHY YOU DISAGREE.
Sorry, but that's just arrogant and rude.
I had a completely different problem with kefka, which was that they said "Great idea." As though this concept of turkification were something new. It's not something new. Therefore, objecting to THIS idea on the basis of it "being a new idea" is a bad argument. If you want to reject the entire category of turk jobs, that's ANOTHER DISCUSSION.
See? Two different arguments.
So don't conflate them, okay?
This points out a problem, which is that we are not acting with sufficient empathy toward the people who are doing the work:
https://news.ycombinator.com/item?id=9108579
Here you shallowly dismiss that the problem is a problem at all:
https://news.ycombinator.com/item?id=9108684
That's the willful ignorance I'm talking about.
It is not my job to solve your ignorance here, but suffice to say that I don't think the markets-solve-everything hypothesis stands up when you look at historical problems with labor issues over, say, the last century. Markets are fine tools when we set out to solve a particular problem, but they have been poor masters.
Markets (and capitalism) are great tools. They should not be seen as the end-all-be-all. Same with socialism, communism, as well as other socioeconomic -isms.
The failing of the markets is the lack of empathy for the fellow human. We know from the way the US looks how capitalism fails. Poor people are disempowered and marginalized from the system, because little money = little power over self. There is also a downward push to bring everyone into poverty.
The current gig economy and using mTurk are both ways to make previous corporate internalites to the public as externalities, whilst privatizing profit. Along with that, they also rely on having none of the protections of an employee, yet with all the requirements of one. It conveniently gets rid of the reforms we fought for in the '30s.
Are there solutions to the marginalization of the workforce, along with poverty working class? Yes.
1. Minimum income. Everybody gets it, regardless of a means test. This could be built up by the local level or state, where fees and fines and such could start the basis of a minimum income, even if it's $20/mo. Start small, look big.
2. Change laws to encourage worker cooperatives. Worker cooperatives, like Mondragon, have shown that they care for the fellow worker, because they all own the company as well as manage the company. It would also fit inside our current idea of ownership and contracts. The goal here is to enable political decisions to make company creation in this route easier. Thus the 3'rd...
3. Change unemployment law that allows you and 9 of your unemployed friends who want to start a cooperative to cash out of rest of unemployment to start a business: a worker cooperative. This has been a law in Italy since '85, and is greatly successful. And given how small businesses are celebrated, this should be a great law to pass regardless party.
When did I say that?
I actually said explicitly that I believe in minimum wage. That already puts me well out of free-market fundamentalism.
If you've read some of my comments on HN, you'll see that I'm also in favor of regulation.
I happen to also believe that a Basic Living Standard is necessary.
I wish we had single-payer health care.
And I'm all for free, socialized education. (With BLS [stipend] on top of it.)
So, no, I'd say I fall WELL outside of the free-market fundamentalism you're referring to.
So, when you call me willfully ignorant, it's pretty damned offensive.
I happen to think data entry is demeaning. Some people enjoy it. I think busing tables is demeaning. Some people enjoy it. I have a friend who loves moving boxes in a factory, because he has free time to think about his poetry. I would never want to sell tooth paste to a super-market chain. Some people like to deal cards at casinos. Some like to sell pull-tabs. Some are waiters and airline stewards. Some people clean bedpans. Some clean houses after grisly murders.
Some people LOVE being a concierge. Well, this is a digital concierge service. And just because you and bendoernberg think it's menial and degrading, doesn't mean IT SHOULD BE ILLEGAL TO OFFER THAT JOB. Or, that we as a culture should be embarrassed that job exists.
If a person hates their job, I want to give them every chance to get out of that job. Every chance. But you're angry the job exists... regardless of whether someone WANTS to do it, for the pay they receive?
> we are not acting with sufficient empathy toward the people who are doing the work
I can't imagine why you think that of me. My grandpa worked four jobs at a time, sometime - because he grew up during the Great Depression, and if you can get work, you work. He set bowling pins. He delivered milk. He drove a school bus. He laid brick. All kinds of crap. Meanwhile, I TYPE in an air-conditioned building, and they pay me VASTLY more than he ever earned. It's ridiculous. And my peers are upset because the product owner violated the spirit of Agile when he made them change priorities mid-sprint (because we were shipping defective product). WHAT? GET OVER YOURSELF!
I want affordable housing. I want upward mobility. I want meritocracy and elbow grease. I want a safety net, and government-supplied-bootstraps. I want people to be able to declare bankruptcy, unlike today. I want an immigration system that's less like the mythical Labyrinth, and has empathy for immigrants.
So, are you really "solving my ignorance," or are you reading WAY TOO MUCH into one line of thought - and then running with it, half-cocked, and totally arrogant?
I'm not commenting on your entire history as person, let alone going back generations. My comments are about what you wrote here, in the things I responded to. You might be Mother Teresa, but "Just don't take that job. Problem solved," demonstrates willful ignorance. And then "Sheesh" implies people are stupid for even asking, encouraging willful ignorance generally. If that's in conflict with what you believe, well, maybe you should try writing something different next time.
I'll note you haven't actually engaged me in conversation here, AT ALL.
All you've done is name-called, and then criticized me.
Can't you see how the combination of a secure net, PLUS a free market where you are free to not take a job, results in people having jobs that they want?
As in, "Problem solved."
The problem is not that there are jobs that you find degrading - the problem is there's no safety net, so people HAVE TO TAKE THOSE JOBS.
Or do you want ignore the actual topic, and just poke me with a stick, some more?
Engagement in conversation is not something you are entitled to. It is earned. You seem to think you're entitled to make facile comments and then not have those comments challenged. And when they are, you turn it into personal drama apparently so that you don't have to face dealing with what you've actually said. I have no idea why you think that makes you an appealing person to have a conversation with.
That was me explaining the background behind my reasoning. Including personal anecdotes. Personal anecdotes are emotional, and you call it drama instead.
You still haven't engaged the topic, other than to say that my proposed solution is "some imaginary world."
Yes, PROPOSED solutions are "imaginary worlds" until you act on them. That's how that works.
I get that you didn't like my first comment, and I agree I didn't explain my entire world-view well in 140 characters.
When I tried to go deeper into my world-view, I was met with nothing but insults and meta-discussion. I'm not interested in either.
However, many MMOs (Eve excluded) usually ban or restrict selling gold for USD.
mTurk and the up coming gig economy are legal ways to extract money right now, as a IT1099
You should watch "Sillicon Valley" if you haven't already. I'm pretty sure they took a lot of it straight from YCombinator. And it's hilarious. "Magic" would fit right in on there, as it's comedy gold.
Honestly, if you had to get those three things at 3 AM, how much time and effort would it take you? Add a screaming child to the equation.
Anything that makes it easier for a consumer to get what they need, when they need it, if they are willing to pay the price, makes the world a better place. Especially if there's competition for that service.
If I had to get those things at 3AM - I don't know where you live, but for me (Out in the country), it's a 10 minute journey to a 24 hour supermarket, then 10 minutes back. The child would likely nap in the car.
You're right, all problems anyone ever has are solved by the Walmart 10 minutes away.
I retract my assertion that sometimes consumers are willing to spend money for convenience.
Have you actually used Magic? I'm a reporter for a new startup Tech.pro and am writing a story about Magic. Looking for users. Please let me know! Thanks!
To pull off Magic, realistically, will require extraordinary resources and execution. Magic will flame out long before they can manage to acquire such.
A company like Amazon however, might just attempt a variation of it. For Amazon it can be another AWS service: ordering things as a web service. Amazon sources contractor labor, said labor gets reviewed / rated. It's a form of mechanical turk.
p.s. Your website - www.complexventures.com - appears to be down, just in case you didn't know.
It is/was available in various countries in Europe from several companies (call center companies, telekom companies, credit card companies). Several "magic" like concierge services were already on HN, never heard about them again. On Reddit it was mentioned a small startup company with seed money is behind 'magic'.
The next step is a personal software agent, a vision that Bill Gates proposed in his famous 1995 book "The Road Ahead". It's basically the next logical step of Siri/Cortana/Watson.
What the hell is magic going to sell me when I ask for roses from my wife? From the great florist down the block who wants $79 or from the chain florist that wants $49 and has 1 star ratings because he only delivers beat up looking flowers? When I dont like the goods, what happens then? I've already been charged. Who do I complain to? Magic directly? This florist? Even then, who cares?
I imagine magic is busy making connections to the low price leaders to be their exclusive dealers. Now you don't have some tastemaker worried about offending you, you have put yourself in a captive market situation where one organization decides what vendor to use, and that vendor may or may not have your interests in mind. In fact, it would be stupid if they did. They are there to please magic, not you. Making you happy with over-priced cut-rate flowers is magic's job.
He hoped Microsoft's "pen-based computing" (later Pocket-PC/WinCE) would get traction. Microsoft Bob (from Gates wife) and Office agents (http://en.wikipedia.org/wiki/Office_Assistant) flopped. The Microsoft Network 1.0 (short MSN 1.0) integrated in Windows 95 (was a serious Web competitor based on Win32 controls and Explorer as browser instead of HTML and Mosaic/Internet Explorer 1) flopped. The second edition of the book features the Internet/web instead of "information highway".
Bill Gates vision can be seen in the Comdex 1995 video: https://www.youtube.com/watch?v=-XxeY-OchwY&feature=youtu.be... (shows the "wallet PC")
General Magic was definitely a competitor. AT&T was on the General Magic bandwagon, see their TV ad: https://www.youtube.com/watch?v=TZb0avfQme8 (triva: Jenna Elfman of Dharma & Greg in her first role in 0:48)
This is the sort of thing HN should be celebrating as a community, not deriding. Sour grapes, much?
there is a middle ground between celebrating and deriding, which is where I think most people hope discussions of new startups wind up
The MVP culture for startups is not a one-size-fits-all idiom.
Launching a service now without thinking about scaling is an amateur move. It is the kind of thing that can kill a company, or at least wound it severely. At best, you are wasting a zillion first impressions in the hope that you will do well enough that they'll eventually come back for a second chance. There are plenty of well-tested approaches to prevent that from happening, and they're not particularly hard to implement.
It seems pretty reasonable to me that people have put the time in to learn how to do this right are critical when it is done wrong. It reminds me of the tension between a person who has worked their ass off to get into a good school and the one who got in because he won the ovarian lottery and his family's name is on a building. Sure, we should be happy every time somebody gets a good education, but I think it's asking a lot for skill never to resent luck.
I don't think they were shooting for the stars with Magic. It seemed a hack from day one. My impression is they didn't intend it to grow as it did.
But for the sake of argument let me grant that some people in YC somehow a) thought something was so important as to take time away from their startup, and b) never even considered that it might work. Once it started working, there were a number of things they could do other than the "make a complete hash of things" strategy that Wired's experience apparently suggests they are pursuing now.
A potential customer you had to turn away because you are to popular and you might contact again once you can safely deal with them is better than a customer you failed completely.
For example, humans try really hard to build "failure proof" nuclear reactors, and those still have problems. I claim that's evidence that 0% failure is just not a reasonable goal.
Strategic Vision + Execution > Nice Idea + Connections
It's not surprising that a concept which, thus far, is demonstrating the exact opposite hasn't been universally welcomed.
I'm sure things will change if it ingenious logistics and commercial partnerships turn it into an Uber-esque juggernaut. At least there's a clear business model.
And we already have "dynamically-allocated pools of secretaries!" Magic is effectively a recreation of the concierge service attached to (some) credit cards, which does exactly that.
This is not to poo-poo the business model—far from it, it's something people want—but I'm guessing the true outcome of this battle will just be A. the credit card companies opening up their concierge services to lower cardholder levels; and B. AWS building a concierge-pool service to go along with (and possibly run atop of) their Mechanical Turk labor-pool service.
The business model of such a service is they sell the labor of someone who solves problems for you, and charge you slightly more than it costs them to compensate the labor.
Or in the case of this specific virtual assistant service called Magic, they have people who don't in fact solve problems for you, and charge some unspecified opaque amount of money.
If you've ever wondered what service would be like at a big company help desk or call center without any rules, well, now you know. Some really good parts, some really bad parts.
This is slightly different, in that a concierge or administrative assistant does a lot of things they're not trained for, just because it's what the boss/client asks for.
UPDATED: Just saw that FH has released some real-time communication capability with their assistants. Very cool: https://play.google.com/store/apps/details?id=com.fancyhands...
Yo led to an innovation collapse in Silicon Valley as many tried to do a "Yo for X" without realizing that Yo became popular due to dumb luck. That time could have been used to spend time on apps that have a legitimate purpose and are not trying to chase stupid VC money.
At the least, Magic is not as easily cloneable. That just means the clones will be even worse.
Interesting you say that, as this article posted as a Show HN the other day was derided as having exposed how easy it would be to clone Magic.
http://blog.sendsonar.com/2015/02/23/add-magic-to-your-busin...
The hard part and main scalability concern is coordinating user requests with other third party services, and that can't be automated in any way. (For at least the next few years, anyways). That's not something you could implement and scale at a hackathon.
How, exactly, do they plan to handle chargebacks and allegations of non-delivery from third party service providers that they don't have any apparent reseller agreement with?
It might be valuable in that Magic might be accessing a market that basically didn't know such services existed. But I don't really see any fundamental process efficiency that Magic brings to the table for this market...
What happens when my pizza isn't delivered? Normally I call the pizza shop.
What happens when my Magic pizza isn't delivered? I text Magic and say so. The magician (may or may not be the same person dealing with my case the whole way through) calls the pizza shop, get's an update. They text me 'Oh the shop said they delivered it'. I text back, a different person gets on the phone to the shop again, etc.
How is that manageable?
It's clearly not easy to clone Magic. It's pretty easy to clone "software to SMS message with customers via a browser", though.
To really make it work and to scale rapidly, they might think about a franchise model, so that franchises can be started in numerous locations by people who actually know the local area.
Magic could make plenty of money providing a reliable and customizable back-end service for local operators.
"Zappos, how can we help you."
"I'd like a large pizza."
"Um, you relize we sell shoes. It's also 2 am."
"Yes."
"Ah, ok. Please hold."
"Ok sir, we've found these locations that are open and have the type of pizza you'd like. Here are the numbers."
--
Magic is cool. There are some big scaling problems. They will probably not capture all the buzz the best they could.
Creating an on demand pool of virtual assistants + some machine learning to optimize the solution paths those VA's take so as to provide suggested solution routes for then to follow, ie for Pizza in SF at 2am, here are some options..
At some point most tasks become routine.
And yes, there already exist lots of va companies. Magic isn't surprising if you know about VAs, but many people apparnetly haven't read The 4 Hour Work Week.
A Universal API for all resources, labor, goods and services seems like a damn good product to build, perhaps magic can be that.
Otherwise, how can we begin to create a Universal Resource API?
"Hsieh likes to end his presentation with what he calls his pizza story. One night, he and some vendors returned to a hotel room late. Someone in the group was craving pizza and was told room service had ended. As a joke, Hsieh suggested calling Zappos. You can probably guess the end of the story—even though Zappos doesn't sell pizza, the customer service rep found a list of local pizza places that would deliver to the hotel. It's a fun story that seriously reinforces Hsieh's theme of customer service."
I've found that the tech behind this is actually quite scalable, thanks to Twilio. As other commenters have pointed out though, Magic's concept has already existed (free) for a long time, through concierge services like Visa signature and Amex Platinum. What's different and innovative here are the sms-interface and the hidden, Instacart-like surcharge.
Which is funny because part of the value proposition of Magic is not trying user interaction to an account.
- I asked "what kinds of things can you help me with?" and got a response back: "We're magicians! We can do anything." But then when I actually asked for some help, no reply.
- The next day I asked if they could get me delivery from a local pizza place which doesn't usually deliver to me. They say sure, great. After almost 4 hours of back-and-forth, we successfully communicate about what I'm hoping for. Apparently the price had already been sent to me, but I never go that message. And the price? $50 for a pizza that's $12 on the restaurant's menu. Even generously figuring $10 delivery, $5 delivery tip, $5 Magic fee, I can't explain how $50 is reasonable. Maybe some minimum delivery charge? Okay, so tell me that.
Average response time is 20-30 minutes depending on how you count a message they never responded to. I actually like this idea a lot, but I hope the implementation improves.
The app can also expand by striking some deals with smaller merchants to handle the customer service part for them.
The benefit I see from this is that you asynchronously ask for an answer and get texted back whenever, instead of waiting on the phone line.
You can write good dispatching software, but at the end of the day your scaling bottleneck is adding humans.
I hope they can overcome these teething hurdles and find a place.
Personally, I don't have a use for this service. I have a smart phone that can pretty easily search Google for anything that I would need.
Concierge services are popular at the high-end and there's always been an opportunity to bring these down to a different financial strata of users.
That doesn't sound similar at all. The substantive difference is that when you asked for a car, a car showed up. Unlike Magic, where you order things and pay for them and nothing happens.
Behold! It's gone!!