Hacker News new | past | comments | ask | show | jobs | submit login
Productizing Twilio Applications (kalzumeus.com)
121 points by joshuacc on Dec 19, 2011 | hide | past | web | favorite | 33 comments

That's a superb article re: real-world Twilio. We did a head to-head with twilio and tropo and twilio won hands down in terms of actual "production-worthiness" despite tropo having "more" features including speech reco. http://pardner.com/2011/04/tropo-not-ready-for-prime-time-we...

> We did a head to-head with twilio and tropo and twilio won hands down in terms of actual "production-worthiness" despite tropo having "more" features including speech reco.

Had the same head-to-head, and Tropo won for us. Tropo has substandard docs and higher prices, but the tech and API is top notch. We also needed the international coverage. Regardless, they are both great products and great companies. You can't go wrong with either, unless you have very particular needs.

Don't disagree. Tropo does some extremely cool stuff... having a single app handle IM, SMS, tweets, voice is kinda magical. Seems to have better local number availability as well. The lack of any rate limiting handshaking for handling sms blasts was a deal killer for my app however... imo if you try to push sms faster that they can send it they should either queue the messages (as twilio does) or provide handshaking that says "can't send that message" instead of silently discarding the message.

That said, our app uses a mashup of both tropo and twilio now, we use tropo to handle "tweet" dialog with consumers.

> imo if you try to push sms faster that they can send it they should either queue the messages (as twilio does)

Absolutely! I simply don't understand that decision. I mean, I can create my own internal queue to get around the problem, but that burden shouldn't be on me.

   Additionally, Twilio is the nascent, embryonic form of the first Internet that a
   billion people are going to have access to, because Twilio turns every phone
   into a smartphone.
So the total potential market for Twillio apps is the amount of mobile phones in circulation which are not smartphones. Will this number not be shrinking in a few years, even if the predominance of non-smartphones in developing countries is causing it to increase on the short term?

For arbitrary businesses reaching out to mass-market clients, it'll be a decade at least before "customer is extremely likely to be carrying true smartphone" will translate to "exploitable assumption". Local businesses do not as a rule write cross-platform phone apps. If you want to be the company that e.g. gets rid of those stupid flashing pager things that restaurants give people waiting for tables, it helps you very little to know that 1/3rd of restaurant customers will be holding an iPhone.

On the other hand, that application is more or less "hello world" with Twilio, and presents a fine customer experience even to people with iPhones.

One might say this would include landlines and to some degree existing phone systems inside companies. Those numbers are probably shrinking some (especially landlines in some areas) but I don't think they'll go away entirely for many moons.

The current hotness is apps, but you generally write a different app per device platform. If you think of smartphones as 'http enabled devices', Twilio is allowing a cross-platform experience.

I don't buy it entirely - much of the benefit of http-enabled devices is the display of information as much as the information itself. Not just 'hi res displays', but any display at all. Information transmittal via twilio is often audio/time based. For the same reason podcasts haven't replaced websites, twilio phone interaction isn't going to replace smartphones, websites or anything else.

BUT... it's a great adjunct tech, and does open up a lot of possibilities for new avenues of information exchange/interaction. One might go so far as to call this a 'blue ocean', assuming you'd read that book and agree with the blue/red categorizations. :)

The market for Twilio is every phone in the world. We also offer SDKs for iOS and Android phones to make and receive calls with Twilio Client. http://www.twilio.com/api/client

I've been looking at Twilio on and off for more than an year and would really like to make use of it and do something interesting. But they're still mostly US-centric that makes it kinda difficult for anyone not targeting a US market to use efficiently (I don't blame them - I'm sure the paperwork itself to run something like this in each country must be a nightmare)

We just added full support for SMS in Canada[0], have local voice-enabled numbers in the UK, many more European countries on the way[1] and international SMS in beta[2]. If we're not where you need us, most likely we're working hard to be there as soon as it's possible.

If you have a phone number you want to use to make calls from in a country we don't offer local numbers, you can verify it to use as a caller ID then call to any of our supported countries[3].

[0] http://www.twilio.com/CA/pricing [1] http://www.twilio.com/eu-beta-signup [2] http://www.twilio.com/faq/international [3] http://www.twilio.com/pricing/international-calling-rates

Thanks. For what it's worth, I'm in India - I know the bureaucracy around here will make it tedious - but on the other hand, everyone has a mobile phone and sms is very popular (and cheap) - the market for phone-based apps is potentially very huge.

We haven't announced any formal plans for expansion to Asia, but we are committed to serving developers worldwide. We'll make sure to let everyone know when we have plans we can share.

+1 for getting SMS and Voice support in Asia as well.

Checkout Tropo (https://www.tropo.com/) which is basically the same, except your application logic can be hosted on their servers if you want. It is by Voxeo so uses their existing network for international calls and SMS.

There is also Plivo (http://www.plivo.org) which is an open source implementation, except setting it up is far from trivial in my experience :)

Twilio just added support for the UK and they've said they'll continue to add support for more European countries.

I would like to do something with Twillio, unfortunatelly I cant find a idea...

Start with any industry/business you like. Find any business process which involves a telephone and a set of repetitive instructions involving it. Turn it into a Twilio script. Take that to business owners, office managers, etc. Demo the script for them. If their eyes light up and their checkbooks come out, you have a good idea.

To get you started you might want to implement a Google Voice like application. Though that sounds that might sound difficult, using Twilio it can easily be done in a weekend. It will teach a lot about how to create Twilio applications plus it gives you the freedom that Google Voice lacks.

For example I hated getting phone calls when I was in meetings. So I hooked my Twilio app into my Google Calendar API and bam, when I was in meeting Twilio would ask them to leave a message or press 5 if it was an emergency. Also greeting people by name when they call is pretty awesome.

And what might this "app" that you speak of be? :)

I haven't worked on it in a while, but you can still view the code here: https://github.com/ekosz/Personal-C-Assistant


There is a big list of them in Patrick's post, maybe that would be a good place to start.

There are tons of cool little ideas out there, and plenty of big ones as well.

I'll plug one of my recent creations: http://shouldbeback.com/

Twilio contests are a great place to see what interesting things people are building and the cool API Mashups that are possible http://twilio.com/contests/. The most recent was a Shopify+Twilio hack that had some really cool results http://bit.ly/vBrGRx

Write a power dialler, that is a system that takes a list of numbers and the number of agents you have available and will ring enough customers so that the agents are on the phone more often than not.

Sell this to people running call centres, make sure it stays up all the time and they will love you.

I've had a couple successes hooking Twilio up to a CouchDB instance for SMS based data collection. This might be an interesting way to start using it.

Plus, if you aren't getting too sophisticated, you can just use Couch to handle the API calls to Twilio directly.

Write an app that reminds people to take their medicine x times daily.

Shameless self plug here: my iPhone app JustRemindIt just does that - http://justremindit.com. Give it a try! It's based on Twilio only.

Thats not a bad idea. Going along the same lines, my parents phone an elderly neighbour daily to ensure she is alright.

Good write up. I had a distantly similar problem last year. Needed a poor man's transactional email and newsletter blaster. I wanted to leverage my ruby app code and make calls to a mail delivery engine. Problem is, things fail, and you don't want to send the same email under the same conditions twice. I ended up with a brute force failsafe by logging a "signature" of each message send to a mongodb collection and do a look up as a safety check prior to each send (or batch). Its not efficient, but works and when my mailer fails, which it does often, I simply restart it and don't worry about ticking off my user base. I end up with a rather large mongodb collection, but every few months, I prune it by removing log items that I know won't happen again.

I ♥ this throwaway line:

The end-game for Zynga’s take-over-the-world vision is the human race slaved to artificial dopamine treadmills.

Continuing that off-topic digression, I highly recommend the recent 'Black Mirror' episode (from the UK's channel 4), "15 Million Merits": http://www.youtube.com/watch?v=Z_c_Kf8NtOg&feature=resul...

That was like 30 straight minutes full of extremely useful information. Wow.

That voicemail deletion "bug" is awesome.

Thanks, that puts a smile on my face.

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