Hacker News new | past | comments | ask | show | jobs | submit login
Twilio's Seed Pitch (avc.com)
319 points by mathattack on June 27, 2016 | hide | past | web | favorite | 74 comments

Twilio also had a very powerful stock demo, which is (IMHO) the best demo since the Mother of All Demos. You've seen it if you've ever seen a Twilio demo.

If you haven't: the presenter "live codes" a quick Twilio application in front of the audience, showing a server-side program which serves TwiML (Twilio markup) that does some trivia action, like "Say hiya to the caller." They then have one person call a number which connects them to the application. It says hiya to them.

People clap a little.

Then they ask the entire room to call the same app. It says hiya to everyone.

People clap a lot.

Then they say "We can get the list of numbers which just called our number from the API, in one line of code. Here, let me print some on the screen with the last four digits covered." Nervous laughter happens.

"And then I'm going to have it call all of you, and bridge you into a teleconference."

Presenter hits enter. Every phone in the room rings at once. Crowd goes wild.

I've seen this demo 15 times and it never ceases to be absolutely magical.

I did that demo 2-3 times per week for over two years. It was fun.

The best time ever was when I presented to Bob Metcalfe's class and he heckled me: https://twitter.com/CaseySoftware/status/170301958727536640

I got him back in my closing though: https://twitter.com/CaseySoftware/status/170304358905425920

The biggest crowd this was ever done with was NY Tech Meetup and it was bonkers: http://johndbritton.com/2010/08/04/live-coding-demo-at-new-y... (too bad the video doesn't work...looking for another).

Fred Wilson also wrote about that one: http://www.avc.com/a_vc/2010/08/how-to-pitch-a-product.html

"...my birth cry will be the sound of every phone on this planet ringing in unison."

Lawnmower Man?

Yes sir.

Circa 1992 and seemed strangely appropriate.

I saw this done in person at the NY Tech Meetup, probably around 2010 or so, and it was the most effective startup presentation I have ever seen.

I'm sorry, but while impressive, it's nowhere near the mother of all demos. There's only one video out there that impressed me as much as the mother of all demos.

care to mention which one it is?

Bret Victor's inventing on principle

Does anyone have a video of this?


I remember Twilio giving this exact demo in 2009, though I can't find the video. There were audible gasps from the audience back then, the moment people realized their web development skills suddenly applied to telephony.

I went to an event once where a Twilio developer evangelist presented and he did the live coding exercise. Using flask he wrote a quick server, and started receiving text messages from the audience in about 5 minutes. Then he applied a NLP library to the text messages, all in all quite impressive.

Here's a copy of his presentation, it was pretty awesome to watch. https://github.com/RobSpectre/Talks/tree/master/SMS%20For%20...

That was probably Rob Spectre: http://twitter.com/dn0t

Rob also built http://trumpfilter.com/, built http://www.jeterfilter.com/, and sings in Adventure Capitalists: https://adcap.biz

Yup! He's a pretty great speaker. Definitely sold me on Twilio.

Rob is a beast. Love that guy!

Rob was a regular at PennApps, the hackathon where I cut my teeth. He was among the most helpful & friendly of any sponsors, helping debug code that didn't involve Twilio at all. He also provided valuable "hackathon strategy" advice like "don't build that feature" to young coders who needed it.

I'm sure Twilio's success is due in no small part to dev evangelists like Rob, I hope he's well rewarded for it.

It is based on "Show, don't tell" but it is not that easy to do. I think the Twilio presentation was impressive because automating SMS and getting a phone number was difficult and required slow back-end setup. It was a simple concept to understand but difficult to implement. Stripe example also simplify the work with payments.

The problem is: how do you follow these principles for less obvious applications? For example, how do you impress your audience if you sell a database engine? It is a good startup exercise to think about it. I would be impressed with a tool to quickly design a web application in the same way Visual Basic did 25 years ago instead of talking about Bootstrap, Foundation, AngularJS, React, etc.

> I would be impressed with a tool to quickly design a web application in the same way Visual Basic did 25 years ago

Me too! So we built one - https://anvil.works

The "Hello World" app takes 45 seconds (there's a video on the homepage).

The "Display a slideshow of images from Google Drive" app takes a whole five minutes, because I talk through it slowly (video at https://anvil.works/learn - scroll down).

If you want to demo (eg) a DB engine, you'll want the "uplink" that lets you call code on your workstation from your app. (There's a three-minute demo of us controlling a Raspberry Pi that way: https://anvil.works/blog/uplink).

Great! What features are you planning to add in the near/long term? I think it would be great to have ready to use access to web APIs, separate the front-end from the back-end, and target other popular programming languages like Node.js.

I recommend to do a "Show HN" to get invaluable feedback from HN users.

I know it can be painful but listing your competitors will be useful.

More features...where do I start? Certainly more service integrations (there is already ready-to-use integration with several APIs, including Stripe, Google auth/drive/sheets, Active Directory for corporate users, etc, but there's always more to do). There's also a nifty DB upgrade on its way, as well as a beta of push/pull Git access to each app's source (that one's live today - there's a Git Access button in the IDE). Ultimately, though, it's all being driven by user feedback now.

We deliberately chose Python because Javascript (language and ecosystem) is a dumpster fire whose weaknesses keep getting papered over with a new framework-of-the-week. (Fun fact: The median modern web app is written with five programming languages and three frameworks. Take a moment to soak in how nuts that is).

Python lets us implement simple, idiomatic APIs for things that would require multiple frameworks, nested lambdas and async hell in JS. (We translate into async hell in the compiler :D)

Expect an HN post or two before long...

Overall it looks like a good tool, but a major part of the magic of VB6 is intellisense. Proper, full blown, everything you type, intellisense. And the modern version is phenomenally good these days, compared to C# and typescript it's painful programming in javascript, Ruby and Python as you have to remember so much more minutae and context switch a lot more.

You don't even seem to have basic auto-completion at the mo. Even hinting the names of your declared controls would be a start.

I worked in the Microsoft c# & asp.net world for many years (since 1.1). The company I worked for got acquired and our product essentially killed. Their product is in Java (among other things) and using eclipse. It is horrible.

I love every time I go back to c# to write a tool or test harness. Autocomplete and intellisense allow me to barely type and let the code fly out from my fingers getting my production closer to the speed of my head figuring the stuff out. The difference is amazing.

We know, and that is most definitely on its way. While we considered the half-way houses you suggest, it's sufficiently important that we don't want to half-ass it. Sublime Text-style autocomplete from other words in the document is nice for power users, but can be actively misleading for newbies, so we didn't want to offer anything except the real thing.

Interesting point. I think that's the first half of the presentation:

"we have taken the entire messy and complex world of telephony..."

Twilio _chose_ and solved a problem that sucked. If your product doesn't solve a tough problem, it will be almost impossible to get the same "wow" factor that Twilio does.

Maybe the answer is that less obvious applications will simply be at a disadvantage right out of the gate.

I think it's important to point out that not only did Twilio choose a great problem to address, it was (critically) also one that a large audience already understands is very difficult. There are plenty of equivalently "valuable" problems to solve, which aren't as easy to communicate effectively, because most people lack enough domain-specific knowledge to understand the difficulty of the problem.

This, I think, is the core truth behind the oft-repeated observation that you should "pay attention to the startups with a few customers that absolutely LOVE their product". Not all valuable problems are ones that a general audience is likely to comprehend, but when you narrow it down just to the audience that does, how do they feel about what you're doing?

Put differently: the Twilio demo is so effective not just because they have a very good problem, but because it's an exceptionally common one as well. They have an impressive solution to a tough problem, sure, but they laos did a very good job aligning "the people who have this problem" with "the people listening to this demo", and that's what makes it great.

Regarding impressing the audience with a database engine, I think it was rethink who did the demo of a cluster of a dozen Mac minis and they'd start unplugging them and whatever steam of data never ceased.

I found that quite impressive. Was it Rethink or MemSQL or some such?

Anyone else remember that video?

That's a direct clone of the Tandem demo of the mid 80's.

Today's generation wasn't even born in the '80s. :)

Alan Kay was posting here... probably he comes from the future.

I heard about a demo like this from a DataStax employee not too long ago, maybe it was Cassandra then?

Edit: or it's just a common demo. :o)

I'd be interested in a link, too.

> I would be impressed with a tool to quickly design a web application in the same way Visual Basic did 25 years ago instead of talking about Bootstrap, Foundation, AngularJS, React, etc.

It's called ASP.NET Web Forms. Version 1 was released in 2002. Demos of it back then were quite impressive.

I tried it circa 2007 with Sharepoint and it was a pain in the * to use the web components/AJAX. It was the first time I used Sharepoint for development so take it with a grain of salt.

Could you please point me to some modern demo?

What this boils down to is that they weren't really pitching anything in the traditional sense. There was no salesmanship going on, no grandiose statements or numbers tossed around about market potential or "even if we can manage to just get 1% of the market..."

Instead, they had something to show, and touch, and feel. They actually had a product. That is the key to a good pitch.

Ummm ... I'd call that salesmanship.

I also heard the CEO of SAP (I think) talk about a sales call he made while he was a JR sales rep at Xerox. It was also pure salesmanship but the end of the story was that he closed the deal without the woman who owned the company even turning on the copier or typewriters.

So, I think the thread that ties these two pitches together (one that was all demo and the other that was zero demo) is probably that the salesman got the decision-maker to trust him.


I have given similar demos. You first explain what you are going to do. Before the demo, they go all go into "no, it is impossible. I have been in this business for 20 years. Blah blah." You demo it, and once in a while, they get it, their jaws drop and they appreciate it. But mostly, it is a suspicious disbelief, followed by adversarial challenges where they insist you disclose everything under the covers right then and there.

I have even heard this: If this was possible, why hasn't IBM done it?

Sometimes, it is pretty depressing.

Reminds me of the whole "Just run this cURL request and a payment is made, it's that easy." messaging that Stripe used to have right at the top of their landing page in the beginning, showcasing not only how easy their payments API was but also that they were a company by developers for developers. Brilliantly simple :)

I just realized what stripe has done is now I always tell clients "payments are the least of your worries" thanks to stripe. Hard to understate that sea change.

Which is true, unless your business falls in their "Prohibited" list [1]. It's quite a long and wide ranging list, which would appear to restrict many businesses from accepting Stripe payments. For example, 51. Personal computer technical support and many things to do with travel and tourism: 13. Airlines, 23. Currency exchanges or dealers, 27. Cruise lines, 49. Timeshares, 50. Centralized travel reservation services or travel clubs

[1] https://stripe.com/at/prohibited-businesses

That list looks to me mostly like a list of things that are either scams, or that Stripe and the CC companies have perfectly Bayesian-rational reasons to believe are scams. Travel and tourism is the only one on there that strikes me as coming from CC agreements rather than that.

"Personal computer technical support", for instance, is almost certainly referencing the scam where someone cold calls you and tries to explain that your computer is infected. There aren't a lot of legitimate businesses that could be called "personal computer technical support", and I suspect of the ones you may even be thinking of that Stripe may classify them differently. I imagine the mom & pop shops are classified as hardware vendors who happen to also offer services and the more serious versions of that business are "consulting". (And at that point, probably not using Stripe anyhow.)

I'm not so sure. I think it could be argued that many open source service models fall under Personal Computer Technical Support. Let's say I set myself up as a Linux guru (I'm not, but for the point of the argument, let's pretend). If someone then phones or emails me after borking their installation and I fix it for them (either in person or perhaps remotely), it could be argued that I'm offering personal computer technical support. However, the above scenario could be a legitimate open source software business model in many cases - develop consumer open source software and provide paid for service to support it.

I didn't say that there is no possible overlap with legitimate industries. I'm only pointing out that all the instances that I know about (not necessarily a representative set) of little computer support businesses also have a storefront where they are selling a usually-modest array of hardware for upgrades and such, and Stripe may classify their business as something other than "personal computer technical support". For all I know, that's why they all seem to sell a modest, almost perfunctory, array of hardware upgrades.

"Age-restricted businesses" - wondering what kind of scam this would be...

My guess is things like alcohol and porn, things that are not intrinsically illegal but are still legal hot potatoes of compliance. An online alcohol vendor that had an "age check" that was easy to game ("yes, I'm 21") would be easy to set up and make a lot of money, but a lot of laws would leave Stripe legally in bad shape when the hammer came down, but the PR could be an even bigger disaster.

In addition to the obvious -- bootleg cigarettes from reservations and booze.

Are they scams or regulated industries? Mix of both?

> That list looks to me mostly like a list of things that are either scams

That's easy to say right? Oh it's a scam.

I can't use Stripe because my business falls into the "Scam" area. Am I a scammer? No because otherwise my business would have failed years ago.

I happily use another competitor. I haven't had any issues with them and since I am not a scammer, I don't foresee any either!

Remember. One mans ITS A SCAM, HE'S A SCAMMER is another mans very successful business which people don't have a problem with.

It's hard to take this comment seriously if you're not using your main account and refuse to say what your business actually is.

In addition to pc86's point, my comment wasn't targeted at you personally. There are certainly businesses that will be in the Stripe exclusion set. However, I was arguing against the implication that they've got some sort of really extensive set of exclusions that lots of people will need to worry about. For all the entries on it, it's not truly that large. Mostly it's just the fact that while they need the "don't do illegal things" clause, it's way easier to exclude specific things and then tell people that they fall under that specific rule than to get into lengthy and fruitless arguments about whether or not a particular activity is legal.

> Am I a scammer? No because otherwise my business would have failed years ago.

You may not be a scammer, but that's a horrible argument. Bernie Madoff's scam started in the 1970s. He wasn't caught until 2008.

I stopped recommending Stripe when they stopped accepting my card. Despite me pointing this out to them, nobody at the company has done anything concrete to fix it. Sadly, they are no longer a company by developers for developers.

In addition, they are pretty useless in some regions where they do not support locally dominant payment methods.

Hey there, I work at Stripe. I'm sorry for trouble—could you email me at edwin.wee@stripe.com and we can look into this?


This comment is so off base, I can only imagine it might be a sleazy competitor with a throwaway account, or the single most unlucky Stripe customer in the history of Stripe.

Perhaps you could elaborate on the card provider or what is unique about your card.

Did that last week to no avail, emailed a responding employee, got nowhere.

This is the list for Austria, most likely from whatever partner bank is helping their roll-out in Europe.

https://stripe.com/us/prohibited-businesses <-- US list is much more clear-cut

Developers, developers, developers, developers, …

I competed in the telephony space between 2006-2009. Us and Twilio all launched at the same time approx. We lost, they won and I'm happy for that. Today I'm their customer (and they are non the wiser about our somewhat parallel history). Everything about Twilio is top notch, from the tech all the way down to customer service. I only wish them success and growth.

Curious to know of your company's experience. Why did you lose?

People. Lethal choice of CEO. I'll avoid names if you don't mind. Our product was essentially like Twilio, to start with at least, until the company was dragged into a more social arena (with deals with then big, now defunct social networks). We had some pretty cool features like the ability to trigger calls, sms chats, voice blasts etc, via text commands ("call humpty, dumpty and numpty tomorrow at 5pm", or "text johnny, bonny and sonny: Hi guys"), so you could communicate with the system via IM, email, web etc. We then essentially took all that lovely stuff and instead of being the API behind Telephony, or even the Telephony behind the social, we did stupid marketing stunts for some singers and stuff like that.

The whole story is a fascinating tale of human nature more than technology. Maybe one day...

Wow, very interesting. Thank you for sharing. It seems like the moral of the story is to follow the money as Twilio did. B2B may be boring but at least they pay you.

What was the thought about where the "marketing stunts" would lead? More consumer-focused product would lead to your company becoming the next AT&T or something?

There was no thought behind it. Not rational anyway. Hence it was death by Lethal CEO. We the founders were kindly removed from the company not far from its actual demise.

And yes - I would take B2B any time. B2C is hard and requires twice the luck B2B does. When you provide a service to businesses, if your service adds value, they will buy it. When you deal with consumers, you have to be trendy, at the right place and time, cool, exciting, amazing, talked about, add value or not - who cares, etc. That startup started as a B2B idea, and morphed fairly arrogantly to a hip wannabe consumer app.

In my experience, minimalism and cohesion are also one of the best way to sell products.

Just show how it's done. Be simple and clear. Don't try to spin. Don't inflate the future or past.

I've never thought of pitching like this. Demoing, yes, but pitching investors with a live coding demo? Wow.

Definitely going to dig up some of Twilio's demo videos and study them.

Great story! And congrats to Twilio on the IPO, 90% first day of trading pop, and creating a great (and essential) service ;)

Interesting. That's also how I felt the first time I ever used Twilio, having come from a background including telephony (including wrestling Asterisk). I can easily tell Fred has also been exposed to the hellish nightmare of telephony in his past at some point. You can't really imagine just how bad it used to be, even with the Asterisk project. For example I have a fairly large library of computer books (perhaps five or six shelves) in my office ("laboratory"). The thickest book in that place by far is the Asterisk animal book from O'Reilly. The next thickest book is Sendmail's animal book. Twilio really did do something totally amazing.

Twilio, like Stripe, should be commended for their fantastic marketing efforts. Looked to use Twilio as really liked their offering but too pricey here in the UK.

This site was denounced as unsafe by my Windows Phone Nokia 930 IExplorer and it blocked my access. No idea why, this is the first time for avc.com

This should not be downvoted. I'm getting "This website has been reported as unsafe" on IE11 on Win7 x64. Even if it's a false positive, it's disconcerting to see this on an article linked from the front page of HN.

https://www.stopbadware.org/clearinghouse/search?url=http://... doesn't show anything.

Is there any further information on the warning?

Microsoft Smartscreen says it's been reported as a phishing site. However, if I strip out the URL parameters and just go to http://avc.com/2016/06/best-seed-pitch-ever/, there's no problem. It's only when you go to the full linked URL http://avc.com/2016/06/best-seed-pitch-ever/?utm_source=feed... that there's a problem.

A long time ago I saw how my friend writes code. Since then I started to learn how to write code too ) Live coding has something special..

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