Hacker News new | comments | ask | show | jobs | submit login
AWS Pinpoint Launches Two-Way Text Messaging (amazon.com)
181 points by artsandsci on Sept 26, 2017 | hide | past | web | favorite | 84 comments

"helping to power"

You confused me for a second, because you're swapping the subject and object with your quote.

AWS Pinpoint powered by Twilio -> Twilio helping to power AWS Pinpoint

Not my initial interpretation of "AWS Pinpoint helping to power Twilio"

Can someone provide clarity on this pls?

Twilio's stock started dropping [1] with this announcement. Jeff Lawson wants to protect his stock price and restless employees (I am sure there is an internal email forthcoming as well).

The worst (possibly realistic) reading of "Helping to power" is that Twilio confirmed that they are just one of many companies that are powering this solution for Amazon. The implication is that Amazon will keep playing these companies against each other to drive down rates.

[1] https://finance.google.com/finance?q=NYSE:TWLO

It sounds like this was tweeted to quiet the peanut gallery.

Over the last year, I've been exposed to a lot of the nitty gritty of the SMS and carrier world and I was surprised to learn how at scale competition for SMS volume is insane and extremely complex aggregators and the commercial arrangements seemingly incestuous.

The cost savings of easy integration and the appeal of a strong developer ecosystem are less and less attractive as the volume of SMS being managed make the marginal cost savings of going with a direct relationship with Amdocs, Syniverse etc....

Disclosure: Short TWLO.

If you look at Amazon's playbook (e.g. in Retail and Transportation), Twilio should be concerned:

1) Partnering with Twilio and other aggregators gives Amazon a fast path to market with broad reach. This is consistent with how they have approached groceries, transportation and other verticals.

2) Amazon is willing to be aggressive on price to win market share. They have already undercut Twilio and they will no doubt continue lowering rates as needed. The current rates are still very high relative to SMS cost and they will come down.

3) Amazon is establishing the relationship directly with developers. Others have tried to do this on price (e.g. Plivo) but they haven't had the developer base and Twilio has been great on developer marketing. AWS is very formidable here. Note that Twilio had to come out and say that they were a partner.

4) The reality is most of Twilio's business is on the most basic APIs for making phone calls and SMS messages. Twilio may have a lot of other capabilities, but 99% of them don't matter. In addition, AWS will keep adding capabilities.

5) AWS will keep asking Twilio for price concessions at each negotiation renewal point and RFP aggressively. With more volume they will have more power. Given the commodity nature of SMS, there is not much Twilio can do about it.

6) You can look at the position UPS is in to get a sense of how this can play out. Amazon will chip away at the most lucrative parts of the business and look at going direct. Leaving UPS (or Twilio) with the less interesting edge cases.

The one counter I have read is that porting out of all the markets will be complicated for Amazon. If Amazon has done a reasonable job with vendor contracts, they will have the right to selectively renew or port-out on a per market basis.

There is a scenario in which Amazon acquires Twilio. It might happen. But never bet on a company because you hope they will be acquired.

EDIT: It may actually be worse for Twilio given Jeff Lawson’s tweet. “Helping to power” seems to indicate that Twilio is just one of several vendors as opposed to “powered by Twilio”. Can someone from AWS or Twilio confirm?

Don't use Plivo. It's the worst provider ever. 5 months ago they forgot to renew their domain and all services were unavailable for 2 days. and this was not the first time we experienced such issues with them.

We use both Twilio and Plivo in our text messaging products in CallHub. Both are great vendors to work with. We'll not move to Amazon just for a lower price. The things that we value are,

1. Quality of delivery

2. A consistent API across all our telephony needs.

3. Great support when we need it

I especially love the fact that these vendors are constantly innovating and give us early access to those innovations as well. The most recent was Twilio's release of studio.

CallHub probably has more demanding requirements for a call and SMS platform than most customers. If I understand your business correctly you provide robo-calling (!) services for political parties. Most companies use Twilio in very basic ways.

In terms of innovation in SMS and telephony, there are also numerous ways you can see Amazon surpassing Twilio. For example, combining speech detection, automation, etc. from Alexa with their platform.

Hmm. Interesting.

Cheaper (~0.65c) per SMS than twilio (~0.75c) in the US, great if you're already on AWS and support for sending to 200+ countries[1] but no MMS support that I can see...


Amazon got volume discount I’m sure, and provides marketing for twillo

No need for volume. SMS is dirt cheap on its own. For my clients we implement custom SMS solutions and our wholesale rates are free inbound and 0.0018 outbound. And we can activate any off-net number, don't need to own it even.

0.0018? What country?

What country?

SMS rates vary depending on which country/telecom it's sent to/from.

The 200+ countries is slightly misleading, especially when the title mentions two-way messaging. I've done a lot of research into messaging/SMS providers like Twilio, Nexmo, Plivo, etc for supporting two-way messaging in developing countries. Most providers currently only support two-way messaging in similar subset of countries in North America and Europe. Since it is powered by Twilio, I don't think it adds any new countries and definitely not 200.


Pinpoint works with multiple partners and carriers. You can get the list of countries here: http://docs.aws.amazon.com/pinpoint/latest/userguide/channel...

Disclaimer: I wrote the post and work for AWS

Thanks for the reply. Are you saying that all of these countries support two-way messaging? It's my impression after signing up and looking at pricing pages that many of these countries only support outbound messaging.

You're totally right. We'll update the list of countries that support 2-way messaging in the docs in the next few days. It's ~50 or so.

Hey, one thing I spent a bit of time trying to find and couldn't was if the numbers would be designated cell phone numbers?

For example, in Thailand all cellphone numbers start with 08/09, in the UK with 07. There are some funny local laws about that. I have quite a few usecases in mind which require the provision of a cell phone number.

One can send SMS to (at least some) landlines in the UK.

When is SIP trunking and Hosted Business PBX getting added to AWS?

Which is weird because AWS Pinpoint is powered by Twilio[1]


It’s probably in the api lock in. Amazon can switch to another provider once they’ve got a better deal elsewhere, and customers need not notice. They’re commoditising twilio, which is not a great place to be in, for stock.

(Alternatively: Amz ends up buying twilio and the market was wrong :) )

I don't think this is what's happening but would be an interesting acquisition strategy. Spook the market into offloading Twilio, driving the price down before they make their move.

They can't really easily move elsewhere for 2 way texts. It'll require porting numbers worldwide in every country away from Twilio without any downtime

Algos going nuts again?

Twilio is massively overvalued.

The UI is garbage but that's par for the course with AWS I guess.

Any specific feedback you want to share on how we could make it better? Is it just the UI or UX (flow of info) as well? I'm happy to collect any notes and get it back to the service team. Alternatively you can click the feedback link in the lower right hand corner of the console page.

You can email me any feedback or DM me: randhunt@amazon or @jrhunt on twitter

Given how AWS tends to move to a policy of "we do what we want", I'm not surprised. I've had EC2 and SES send out spam for weeks without mitigation despite complaints.

EC2 has limits for how much SMTP they can do by default. SES has similar controls in place. They take the fight against spam very seriously. If you wanna let me know specifics randhunt@amazon.com I'd be happy to help take care of the issue.

No offense, but that's rich. I've emailed abuse, sent public notifications through Twitter, and more. Received silence.

I've had spam campaigns -- open spam campaigns, complete with scraped emails, dubious subjects, and more -- last for 3+ weeks before finally getting shut down. I wouldn't call that "taking the fight seriously".

If you received no response at all from the abuse alias then something is wrong there and I want to fix it.

You’re welcome to email me or link to the tweets or ping me directly on twitter @jrhunt. I personally take it seriously and I will get things taken care of. HN isn’t a great forum to dive deep on individual support though so feel free to take the convo to email or twitter DM.


Delivery rates on AWS are not great in my experience, and there's no easy way to automate some kind of event triggering when a delivery has failed (ended up creating a lambda that reads the fail events from cloud watch). Plus, you have to specify a monthly spending limit that must be manually reviewed and enabled by Amazon, if you go over the limit, Amazon starts failing the SMS delivery, even when SNS responds with a successful status.

AWS is great, but the biggest advantage of Twilio and other SMS providers is that they are way easier and straightforward to integrate.

One thing I like about AWS is they keep pushing forward. You'd think they would have held this until AWS Reinvent, which is only 2 months (edit: was originally "weeks", doh!) away.

I think it's a simple way of figuring out potential products and services that are cross-selling opportunities for their existing customer base in order to increase their Average Pocket Share. As it turns out, everything that they push forward is something someone is already offering to their customers and hence 'Why not us' while all it takes is some new devs or existing devs to work on copying an existing service and offering a cheap refined version of it.

Edit: I'm not trying to undermine the challenge or business incumbence requirements. It's just not a risky innovative proposition.

I presume they can analyze any requests/traffic and figure out which services their customers are using. Talk about a competitive advantage.


AWS re:Invent is during the Thanksgiving weekend.

lol you're right - I was even thinking months as I typed that. Silly bit flips :-)

Put your fingers on a diet!

I wonder what happens if someone calls the phone number after receiving an SMS text. The FAQ/docs don't seem to address this.

In Australia at least, it doesn't connect/you get a busy tone

I thought this was supposed to be... 2 way. But an Israel SMS just says from "NOTICE" and I can't even respond to it.

"For countries that require an alphabetic sender ID, the message displays NOTICE as the sender ID." http://docs.aws.amazon.com/pinpoint/latest/userguide/channel...

Did you try using a local number? Many country codes only allow sending to local numbers. Example: Romania.

Looks like Amazon is gunning for Twilio

Just speculating here, but it wouldn't be surprising if Amazon ultimately acquired Twilio. CEO Jeff Lawson used to work for Amazon after all.

his is likely a dry run on how well it can integrate... and what adoption levels... then they will buy Twilio. Buy some twilio stock now.

Agreed. If you see how Amazon has done acquisitions in the past, they like to brutalize companies by aggressively undercutting them and then buying them for cheap after their value goes down. That's what happened to Zappos and Quidsi, it's a strategy that's saved them hundreds of millions. I like Amazon and I basically worship Bezos but I would not like it if they moved into my line of business.

This is powered by Twilio.

Doesn't bode well for Twilio that Amazon keeps rolling out these ecosystem-enhancing SMS features. They're still a standout solution for anybody who doesn't use AWS IMP but these features are getting a little too close for comfort

Twilio's advantage is that they have amazing API documentation, and extremely low barrier to entry.

Yea, I tried to use SNS today but couldn't find any example code for Python. I got frustrated and signed up with twilio. After signing up and doing their test I was able to click directly to the Python example.

1. I love Twilio and I recommend it over pinpoint most of the time. The API is objectively easier and the documentation is better.


    import boto3
    sns = boto3.client('sns')
    number = '+12345678900'
    sns.publish(PhoneNumber=number, Message='Hello, World')
Disclaimer: I work for AWS and wrote the post

OT: why on Earth would you call your library boto3, instead of something obvious, like aws?

I can't be the first person who was confused by this.

Here's the right twitter link sorry: https://twitter.com/garnaat/status/913093232917086209

I admit I don't use SNS but the boto3 docs are usually really well documented.

the integration between the python code and SNS is relatively well documented, but it is missing examples for how to actually achieve real-world results (such as sending yourself a text message).

Responding to this and the parent it was the lack of examples that made me change my behavior. Not the lack of documentation for boto3.

Also in my experience SMS is confusing (read: arbitrary and complicated rules per country) enough you want a domain expert available for support. Given the track record of other AWS products I find this unlikely.

That said, if you have an existing sms setup, this is quite exciting.

I have some Twilio stock. It has been under performing for the last year :$.

As a product, I hear that Twilio is great. But from what I've read, after their IPO last year, the have to close large Enterprise accounts and are having trouble. Amazon eating their lunch is not good news either. Also, is Twilio still using AWS?

Yes they are still using AWS entirely (as far as I've seen) in fact they've doubled down on their AWS commitment it seems by integrating with Lex and Connect.

Their stock isn't so hot right now but that's not terribly out of sorts for a relatively new company. Enterprise I've heard many stories about. Their commitment to starting with the developer is admirable though.

Twilio is in a very tough spot, because any relatively large account quickly realizes that they are big enough to roll out their own solution, and in 2017 is very easy to do, from stellar open source options to many vendors willing to sign up just about anybody as a wholesale account. And many have moved away, and will continue to do so. Even with the recent discounts, Twilio is way over priced, and its REST API is simply not in line with what is expected from a communications platform in this day and age. That's why they are trying to will continue to shift to hosted code. Corvisa (acquired by Shoretel) had it right when they had embedded lua.

Disclaimer: I do own some Twilio and several other telco stocks (but I still believe Twilio is way overvalued and buy on dips because a large carrier will eventually buy them out so I'm willing to go for the ride), but I make more money from companies dumping Twilio/Nexmo/Plivo and going with custom solutions.

What are some of the open source options?

Twilio is dead simple. An idiot like me integrated SMS alerts into my fantasy football injury notification app literally in the 15 minutes that I drink my coffee in the morning before leaving for work. And 10 of that was signing up for the trial and reading some docs.

>extremely low barrier to entry.

Isn't AWS' entry barrier even lower?

I think this makes the answer to your question "No"


I imagine this matters less for the larger customers spending a lot in this space.

True. But for every IBM spending hundreds of thousands of dollars, there are thousands of small companies spending hundreds of dollars. So long as you control the software, it's probably a wash in the end.

Not exactly, given that Amazon is using Twilio.

Just a heads-up: AWS Pinpoint text messaging cannot be paid with credits.

I just checked my bill and was charged $0.18 for sending three test messages to myself.

What is it that makes SMS so expensive to send in the UK?

Cost in the US: 0.645-0.75¢ per message

Cost in the UK: up to 7+¢ per message to send, inbound 0.9¢

I get that there are differences, but a factor of 10 is quite a shift. Twilio is 4¢ too if you want to send from an actual mobile number.

One more step closer to the end game... Business Hosted VOIP. They're already doing SMS, call center and a bunch of other stuff in this space. Hosted PBX and SIP trunking is coming. They are on a VOIP/WebRTC hiring spree since last year.

Yes, as I mentioned call center, and they also have Chime for conferencing. I expect trunking and business voip to be released very soon.

AWS is doing a great job producing more and more sticky features.

Is there a straightforward method for charging a customer for those texts?

Is this available outside of the US/Canada?

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