The Magic of Magic

You know you’re at the crest of the on-demand wave when your girlfriend uses a concierge service to “mariachi-bomb” you and your buddy at a bar on a Tuesday. Once I got over the fact that mariachi-bombing was now a thing, and once my buddy and I got comfortable with the fact that all of Croxley’s Ale House in Williamsburg had ceased watching the Rangers game and March Madness basketball and instead were eying us with the sort of bewildered curiosity one generally reserves for those who choose to eat hot, messy, open-faced sandwiches on the subway, I turned my thoughts to how this group of fine, charro-suited musicians ended up at our bar booth.

The answer, of course, was “Magic.”

Magic is the latest in—and I think the most literal interpretation of—the “Uber for X” startup gold rush. The on-demand economy is giving birth to so many iterations of this concept, it’s hard to keep up. But Magic caught my eye, and not only because they dropped a Mariachi band on top of hockey night. No, it was after I got a text from the girlfriend with a simple “So? What did you think?” that I sat back and marveled at the ease with which she had dreamed up, organized, and executed such a bizarrely hilarious experience on a complete and utter whim. I had to know more.

It turns out that Magic was born out of a weekend side project that a small team of YC founders hacked together while taking a break from their main startup idea, an app to monitor and improve your blood pressure called Bettir. It was something co-founder Mike Chen said he had always wanted for himself, a service you could shoot a text to and get *anything* you wanted.

The fact that you had to SMS your “order” was clearly a function of the constraints they were operating under, but to me (and probably them and many other users as well) it’s a great feature. Texting “someone” gives the service a feeling of humanity. In a world where things happen instantly at the push of a button, this felt different, more personal (yet still immediate). Because there was a real live human at the other end using natural, casual language and actively “working” on finding me a solution to my “problem”—whether it be mariachi-related or something else equally dire and pressing—I felt like this was a true concierge, and my subconscious was already accepting a premium price because there was nothing transactional feeling about the experience. Justin Kan had some interesting thoughts on this in a recent episode of Ryan Hoover’s Product Hunt Radio (starts at 3:38). Even when the price is brought up before confirmation, it’s part of a sentence written by a human, and it, at least for me, reduces the friction between acknowledging that price and replying “Yes.”

Of course, this service also crystallizes a lot of what’s crazy about startups at the moment (the fact that there were rumors they were raising $12M at a $40M valuation from Sequoia ratchets up the crazy a bit more too). The reactions have been mixed, pointing out that this feels like just another niche product catering to the valley’s elite and nothing more; that it’s solving a problem a million other companies already solve; and that it illustrates how lazy and entitled we as a society have become that we all need personal butlers at our fingertips 24/7. Some of those reactions are below:

The MVP approach of the product mixed with the massive virality it achieved (and the explosive demand it received) also raises concerns of:

1) Scaling - obviously there needs to be a good amount of automation introduced so they can move away from relying on humans for every request. As one HN commenter noted, if they could get the automation such that 90% of the requests were handled by computers and 10% were handled by humans (allowing for a true hands-on, above-and-beyond experience), perhaps then will their costs be manageable enough to scale to a real business. This would also help manage churn, because this is definitely one of those products where one bad experience could lose a customer immediately. Low response times and consistency are key in that case, as with any on-demand service, and the more technology is driving that, the better.

2) Pricing - While bundling their premium into the quoted price does reduce friction in the moment, their markups seem arbitrary and not clearly defined; better transparency and communication about that—a la Uber and its surge pricing issues-—would be important for gaining and maintaining customer trust.

3) Competition - I had forgotten that a few months ago I’d signed up to be notified when Operator, at the time a stealth startup coming out of Garret Camp’s Expa “startup lab,” launched. In the midst of the Magic hoopla I got an email alerting me that Operator’s beta was now live. I downloaded the app, flipped through a few How To Get Started slides, and realized Operator was a slicker, better-funded version of Magic. Except it was really pushing fashion as it’s primary use-case, at least to me. It was interesting to explore, because I could see they were trying to “inspire” me to ask for specific things, and they had a whole Discovery section to reinforce that nudge to get started. I actually didn’t even put a request through, because I didn’t need any clothes and I got kind of confused what else I should ask for. Nevertheless, they had a splashy article announcing their launch and amassed an impressive 80,000+ waitlist. I’m curious to see where this and other competitors take things.

4) Use-case - What Operator was trying to do got me thinking about intent when it comes to these concierge layers built on top of the on-demand ecosystem. When faced with an empty text box, what do you ask for? Is this a utility, a luxury, or just a novelty? I think what all these on-demand products are selling, when it comes down to it, is time. And how you position your product as it relates to time is important.

Most of these on-demand things follow a similar path across the spectrum of importance to daily life. They start out being viewed as a novelty (as Travis Kalanick says a lot, with Uber him and his buddies just wanted to be “baller in SF,” and Uber did that at the push of a button). Then it shifts to a mere time-saving luxury (it’s Saturday night and we’re late to a party, I’m willing to pay a little extra to get the four of us there on time in an SUV because hailing a cab right now will be a nightmare). Then it shifts to a basic utility, where Uber is the defacto choice for getting around a city (Hailing a cab? Calling a car service? What’s that?). The pricing tends to reflect this shift: as volume goes up and efficiencies increase due to better technology and operations, margins improve and pricing can decrease without killing the business, further increasing volume, etc. Obviously the current availability of growth capital to companies on this trajectory accelerates this process by allowing market leaders to decrease pricing substantially in order to capture more market share and squeeze competition.

(Chris Dixon wrote a great post about this phenomenon of important new innovations often looking like “toys” early on: “The next big thing will start out looking like a toy”)

But jumping the gap between novelty and utility (utility’s not the right word but something along those lines where a service goes from serving an occasional splurge to serving a routine need) is a difficult maneuver. There’s also an inherent catch 22 in the spread of these things: the novelty of the product or service is often what gets people talking, but redefining the value you deliver from novelty to “oh, I actually need this” is where the magic happens (sorry, I got this far without doing that, it was inevitable). I’ll be interested to see if Magic (and Operator, and any other startups that jump into this layer of the on-demand ecosystem) can find their use-case that demonstrates a more utilitarian value than simply “get anything on-demand.” As Ian Hunter, one of the founders of Zaarly, mentioned on Magic’s ProductHunt discussion, “the problem is quite obviously fulfilling blank slate requests.”