How Twilio screwed us over
feint.me
feint.me
--
Hi Anthony,
I'm sorry that you feel we were't straight forward, although our goal was to be completely transparent in our communication with you about the status of this feature. Since the time of launch, we thought we had clearly messaged that international SMS has been an unsupported feature, and that the entire product was in beta status. We had published this fact on our SMS product FAQs, as well as in the public GetSatisfaction forums. Our apologies that we weren't clear enough about that fact, which caused you to roll international SMS features into your app.
We do pride ourselves on putting developers and their applications first, and in being open in our communications, product capabilities and limitations. Please feel free to email me at jeff@twilio.com, I'm always open to feedback, questions, comments or concern.
Sincerely,
Jeff Lawson Co-Founder & CEO Twilio.com
Fortunately I don't have a live app that depended on your service, so all it cost me was $50 deposited to start playing around.
SMS is just a huge pain in the ass. The carriers have a strangle-hold over this and can cut messages off on a whim. We had an experience once with Verizon where they "accidentally" cut-off our short-code to Poll Everywhere on a Friday evening due to some "clerical errors". 72 hours later they finally had somebody "fix" the situation (support doesn't work on weekends there), and in the meantime we were left in a really bad place trying to explain to our customers how Verizon screwed us over.
I've dealt with some total dopes at Celltrust (I need to call these guys out, they're atrociously bad. Please don't do business with them unless you absolutely have to) where our service to Canada was entirely out for a week because they changed how their API worked. They failed to communicate that to us and nobody at their company could tell me how it worked so I had to fire up nc -l, point their gateway at that server, find somebody from Canada who could send a text, and see what came in. After the fact, I told them how horrible the experience was and provided recommendations on how they could improve it. They threatened to cut us off as a client because they felt I was being mean to them instead of acting on the feedback that could have made their service better.
Twilio thinks this is assine and are trying to change how this all works. We've shared with them the pains we've experienced with SMS aggregators and they share the same thoughts with us that its a totally unnecessary amount of pain that developers face to build SMS applications. They're really trying to make this easier and lower the bar.
Anthony, while your experience wasn't that hot, it was far and above what you'd experience elsewhere. At least Jeff acknowledges that Twilio could do better. That is so far above and beyond anything else in the industry. Throw on top of that an amazingly well executed API and the drive to make a difference in this industry and you're in good hands.
You are not really taking responsibility here. You are good guys. Just say "I am sorry" and ask how you can make him feel better.
http://37signals.com/svn/posts/1528-the-bullshit-of-outage-l...
Better would have been:
"Anthony -
We're glad to have you as a customer, sorry we screwed up here. We posted notices to the status API and support email - but if you didn't get it, we didn't do our job! I definitely gained some perspective from reading your post.
We'll work even harder on future product updates to get notice to all our customers. If there's anything I can personally do to make the situation right, email me: jeff@twilio.com.
Again, very sorry. Thanks for being our customer and letting us work to earn back your confidence.
Jeff Lawson - Twilio CEO
P.S. Task.fm looks really cool. Best of luck!"
Danielle from Twilio wrote a much more sincere and helpful apology in the thread above...she should probably handle the "responding to blog rants" duties from now on.
Predictability is critical to B2B services. It is more important than price or performance. It allows customers to make an informed decision. It prevents dissatisfaction and PR disasters.
I also think this should be a lesson for younger businessmen. Customers will even get angry if you cut undocumented features. Or fix bugs they relied upon. Any change that impacts customers needs to be communicated in advance.
IMO, if you truly "put developers and applications first", you would have done this. I humbly suggest some soul searching.
> You may send SMS from a Twilio number to an international number and find that it works. However, this is an unsupported feature and service outside the US may change at any time.
Can't get much clearer then that.
http://webcache.googleusercontent.com/search?sourceid=chrome...
The reason he is upset is that Twilio didn't give him a warning window.
I just want to lay out what we're currently doing to achieve our goal of being one of the most accessible API companies in the world, and encourage you to tell us what we can do better. Right now we communicate status via our API, documentation, FAQs, forums, and social media channels like Twitter, Facebook, and IRC chat. We can also be reached by email 24/7 as well as over the phone from 7am to 7pm Pacific Time. Finally, we publish the status of all our services at status.twilio.com
In the case of international SMS, we saw high rates of undeliverable messages starting on August 16th and we have communicated this via our status page, forums, and email support channel. When it became clear to us that we wouldn't be able to restore service, we emailed customers who were impacted (yesterday) to let them know. While the service hadn't been working for awhile, people were checking back with us regularly, and we wanted to keep them in the loop so they could find new providers if needed.
As others have noted here, international SMS is very hard, we're trying to be open and keep everybody in the loop. We're very sorry for any trouble this has caused. You can always reach me at danielle@twilio.com
> The email from twilio informed me they had, without warning switched off all international SMS functionality. ...they were able to massively degrade the functionality of our app – an app which not only do people pay for, they rely on.
Which I suppose is interesting enough, but then:
> But when it comes what really matters, being able to offer my customers a reliable service, you need a written agreement.
Which isn't exactly profound wisdom. If you're charging people for a service, you better make sure that you can hold your backend providers accountable for their end of the chain.
Is twilio better suited to internal use rather than for building SaaS apps?
These are serious matters.
I can fully understand their decision to go US only for now, but I really hope that at some point they'll be able to roll this out reliably world-wide, that alone would lay the world literally at their feet.
So many services are now hard to impossible to implement because there is no easy way of doing this, in my own site I would have several applications for it right away but only if I can guarantee international delivery.
May be interesting to keep bookmarked (I don't write SMS-enabled apps myself).
I used one such service from 2000 or so when I ran a ringtone website.
It's a solved problem. Just search Google for "bulk sms gateway" etc and you'll find literally hundreds of providers with various APIs and so on.
We have paid for an "SDK" with one provider, they came to our office and installed a hardware device which we also had to buy, and they still can't manage to deliver 10% of our messages. That was in Australia.
After that mess, I personally wrote Twilio a letter asking them if I can help them establish themselves in Aus and the Middle East; which they declined, saying they were getting around to it themselves, which is good.
Bottom line: I would chalk this up as a slight oversight by a bunch of very capable, if very busy people, bless their heart.
They are one decision away from being fucked over by Skype and Google, if those two ever decide to compete with them. So I am rooting for our kids, even if they have to make gut-wrenching decisions like this on the spot.
Remember, just a year and a half ago they were nothing but a RESTful URL that accepted a few XML messages. They have come a long way, imo.
We're programmers, we're hired not to do everything ourselves.
Just find someone who provides smpp gateways and save yourself the pain...
When we launched the Twilio SMS Beta we tried hard to support sending SMS messages to both US and International destinations. When there were problems, we worked with our customers to collect forensic data on hundreds of carriers worldwide and pass it to our carriers partners to debug.
At Twilio we are dedicated to working with top quality carriers and technology. After months of working to fix problems, were not able to deliver the reliable International SMS service our customers have come to expect.
We apologize for any problems this has caused for our customer and we'll work to bring back International SMS service after were able to deliver on the quality we do the rest of Twilio services.
Cheers, -Evan
CTO and Co-Founder
You bring up a good point that the information might not be readily discoverable. We'll work to make the status page more findable and to extend the API (http://status.twilio.com/documentation/rest) with features such as RSS to let customer subscribe up-to-the-minute status information.
That's not to say that we don't want international bad, but we pretty much trust their judgement as to whether or not they say a product is ready or if it ain't. Twilio does good work, and they're amazingly responsive to developer requests/feedback.
That's really lame...
They have the obligation to inform users they're going to close a service, whether it's beta or not. The answers they gave is not the way to do it, giving an excuse because the service was unsupported and completely ignoring their fault in warning their users shows that they don't give a damn about the users.
There's nothing wrong in coming here and apologize for not giving a warning before closing the service, but they choose not to do it...
This is not an apology for not informing the developer before closing the service...
Reads like an apology to me.
Read on for a bit, they acknowledge that their communications were lacking and have improved on that by making it easier to find this information.
What do you expect of them? Groveling? A refund?
What he should have done is apologizing for closing the service without telling developers about it.
They can try to deflect this as they want but the fact remains, he should have sent a warning the service would be closed at date X, not closing the service and tell afterwards they're sorry for not making it clear...
The state of SMS is ridiculous. We go through a gateway. Certification was painfully slow, and the cost... OMFG the cost of a text. -It's as if the alphanumerals had to be mined, shipped, refined, and then placed into each message by hand.
I applaud Twilio for what they are trying to do. IMO, they are playing interference for an ugly beast.
As others have pointed out, SMS is an unsupported feature and service outside the US can change at any time.
So in conclusion, apart from twilio losing one of there early adopters (task.fm isn’t a big customer, but that doesn’t mean we can be screwed over) they are also damaging there brand. They used to be all about developers.
I think only getting miffed that your OWN service wasn't perfect and then posting to HN about how task.fm was "screwed over" is more damaging to twilio (is it?) than an SMS outage.
The author didn't read Twilio's own SLA and hardened entrepreneurs don't burn bridges publicly everytime something goes wrong.
I know trying to build an awesome web application is hard and its frustrating when supporting services go down, but that is the nature of our beast. We build on top of each other, so we must respect and support each other. Getting mad and cutting off your own hand because you are angry isn't helping build anything.
Google's API has bugs that my start-ups web application runs into all the time. Sometimes they change API behavior, which the visualization screws up the entire UI in the process. This effects thousands of our users and we have to release work-arounds as fast as we can. But I don't blame Google - it was our choice to use their APIs.
Generally, this sort of thing will continue to happen with all web service APIs. As a community of developers, I would ask them why it happened and provide constructive feedback about how important it was. You might just find they'll listen and you'll have more pull.
Great start-ups have a short list of early adopters or engaged users that they trust and continually go back to for product advice. Writing a post like this pulls the rug out from under your own service because you refuse to use it, it will cost you more, and twilio will suffer from lack of valuable constructive feeback.
Really just my POV, but hopefully you'll find why posting inflamatory articles like this only hurts entrepreneurs and our community at large. At most it sets an example of how we treat each other.
I had multiple email conversations with twilio team members over the past year and "caution" wasn't something I would say was strongly portrayed
Edit: J, maybe reimbursing him would not hurt.
If you buy a number from us outside the US, we can't do SMS on that yet, but we're working on it. In the meantime, feel free to send texts to and from anywhere from your US number. We won't even charge you any extra for the international sending.
Teleku is now owned by Tropo and as such can do international SMS as well. CloudVox doesn't do SMS worldwide, but I was talking to one of their founders recently and he mentioned they're working on it.
http://www.twilio.com/faq/international
Based on that I presume they switched off the feature globally and not just for you. Correct?
That doesn't mitigate your complaint.if anything, doing that without advance notice is worse. Did they provide any explanation in the email?
Twilio has between a very reliable provider in my experience so far, so this is worrisome.
waits patiently
Here are some challenges the industry faces: ———————————————————-
-Many intermediaries to reach end-users:
Between an App and an end-user you can get up to 4 or 5 intermediaries operating their own infrastructure. Adding intermediaries not only increases cost but also exponentially reduces quality – QOS is measured in terms of availability, delivery ratios, speed, features supported, etc…
Ask your provider about their sourcing strategy? Who do they buy from? Do they have a strict policy in terms of suppliers’ selection? Can they share delivery ratios?
-SLA does not really matter:
What if the providers of your provider do not guarantee any SLA? A provider can potentially guarantee an SLA on its own infrastructure but can’t control what happens afterwards.
-Non transparent fees:
Customers get charged on SMS sent to networks that are not or poorly covered but listed as LIVE in the provider’s reach list – even when you sign a written agreement.
-Tricky to reach all Networks:
Some networks in the world are tricky to reach and sometimes the only reliable way to deliver SMS is to connect directly. Other are installing SMS filters to cut certain type of traffic (e.g. Spam). For instance, if your provider mentions networks such as Comcell Colombia or IAM in Morocco, you’d rather check again because these networks are impossible to reach with full feature support/reliable service without direct connections. Currently there are at least 26 networks that are tricky to reach and 22 that are filtering.
Another rule of thumb is to check whether in Europe their prices are above $0.055, since European Wireless Carriers (also applicable to other regions/countries) require a termination cost for reliable SMS routes. If below that price, it could mean that your provider (or the provider of your provider!) is using unsolicited routes that are not reliable/has limited life time.
———————————————————————————————————
Delivering hassle free (and Low Cost!) global SMS termination is not obvious but we are working on it :)
Tony Jamous twitter: @nexmo e-mail: tony.jamous [@] nexmo.com
I'm glad it didn't work on my own mobile as I would have definitely integrated it into my site and no doubt run into the same huge problem here!
Sorry buddy; your post is basically a hit piece about an experimental service which stopped working. It seems they sent out an email and marked it on their status page.
So I'm going to categorize you with the other startup people who have pretty blog and hurt feelings for your own mistakes.
I hope I never do business with people like you.
But it is also a lesson from the other end. Twilio would probably not have done this lightly. By not offering agreements for there free service, they had the option of choosing to disappointing their users instead of loosing what might be a lot of money. It's never the easiest decision but it in some cases it is the best option.