top | item 36382361

Goodbye, Twilio

545 points| joaopaulomcc | 2 years ago |blog.miguelgrinberg.com

336 comments

order
[+] AndrewKemendo|2 years ago|reply
There’s nothing surprising here.

This is a typical cycle for a venture backed technology business. There’s just nothing else to say here other than this is 100% expected outcome if you decide to build a product on venture capital money, which requires an exit and an increasingly large exit to the point where you IPO.

Unless you avoid this structural pathway, you will be 100% guaranteed to do this.

I am unaware of a venture capital funded technology company that has maintained the core of what they do, and the value proposition, but didn’t push most of their money into paying for sales marketing executive compensation and eventually finally, stock buybacks, or other things that directly enrich investors at the cost of employees.

Having had a couple points with Jeff Lawson I believe he’s a good person who wants to do the right thing for the most amount of people and do it ethically, which is why he jumped into this thread. However, he faces the same pressures as everybody else, and so it’s honorable that he is attempting to find ways to mitigate the downside harms of this new direction but at the end of the day the arrow of history is clear.

[+] jeffiel|2 years ago|reply
Thank you Miguel for all of your contributions to Twilio over the past four years, and I hope your next gig is just as rewarding!

For all those interested in why we acquired Segment, and are focused on the integration of data and communications -- several years ago, we came to the conclusion that the world doesn't need more communications, it needs better communications. More relevant. More effective.

As a developer, I know that's really hard to pull all the threads together to make realtime personalization of every communication hard -- and Segment is so good at it.

So that's what we're focused on!

As an aside, the fraud and scam vectors of email, sms, and voice have grown a lot since we started the company 15 years ago. We are always fighting that cat and mouse game with the bad folks of the world. Are we perfect, no. But are we here to make money off those bad actors? Hell no. That's why we just launched fraud guard [1] for free to all Verify customers, and soon to all SMS customers as well. More to come like this.

Happy Father's Day (in the US) HN!

[1] https://www.twilio.com/docs/verify/preventing-toll-fraud/sms...

[+] fallat|2 years ago|reply
This person businesses.

And to others reading: Miguel didn't say anything bad about Twilio really, just that the alignments for them aren't there - and that's ok. Maybe it back fires and Twilio adjusts back to its dev-focused strategy. Businesses just evolve as they need to. If Twilio ends up being "bad" it just means there's now a spot for someone else to form "The Good Twilio" :) See: the many Google competitors, the many smartphone competitors, the many VPN competitors, etc...

[+] ocdtrekkie|2 years ago|reply
Question: What does Twilio do with the profit it does make off those bad actors once it discovers they are bad actors?

Probably my top issue with companies like Google which make a ton of money off of crime is... they keep the money from the crime! That's a perverse incentive to at least do a poor job preventing it.

I remember visiting the Twilio offices when it was still tiny during a Google Glass related thing. I still have the T-shirt.

[+] spondylosaurus|2 years ago|reply
Only tangentially related—and I know the odds are low that you worked on this personally—but I want to give sincere thanks for the stuff you guys have shared via Twilio Labs. The netlify-okta-auth package in particular was exactly what I needed to complete a recent project, and the documentation it came with was nearly perfect.
[+] swyx|2 years ago|reply
first of all, big admirer of you.

if you feel inclined, would really love your comment on OP's observation:

> Sadly, us developers are not at the center of everything anymore at Twilio.

it does seem the recent messaging has de emphasized that in favor of "Customer Engagement Platform". as the originator of "Ask Your Developer" (I read your book!) that has to sting a little bit. would love to hear your thinking on how Twilio continues to also engage developers in its next phase.

[+] holografix|2 years ago|reply
What better example of Twilio becoming a completely sales driven organisation than this opportunistic and self serving response, with a paragraph directly lifted from you sales 101 deck no doubt.

Well done!

[+] matanyal|2 years ago|reply
Hey Jeff,

Good to see that y'all are still pushing in the right direction! I left between the layoffs to pursue some innovation in the semiconductor space, but I look back fondly on my time at Twilio (Sendgrid)!

Building APIs into businesses is no easy task, but I'm glad that there are those still fighting the good fight, even if the field changes. Best of luck to the Twilio of the future!

[+] dools|2 years ago|reply
Hey there, this is tangential but you just launched limited MMS support in Australia, but no voice and not 2-way and from what I can gather only on the Optus network.

I offer a virtual mobile number that does voice, txt and mms in Australia which could be globalised, interested in chatting?

[+] alberth|2 years ago|reply
Hi Jeff

If I understand the page for Fraud Guard correctly, this appears to cost money.

Seems like Twilio should be filtering spam (a) by default / auto-enabled for all and (b) at no cost to the user.

Much like how spam is included & auto enabled for all (and free) from Gmail/Outlook/etc.

[+] rsync|2 years ago|reply
Haha.

Yes, fighting the good fight - doing god’s own work.

[+] paxys|2 years ago|reply
"The company was great when it was just engineers, but then they hired account managers and sales reps and business people and marketers who ruined everything."

Common refrain from developers at SaaS companies who don't realize that the party they were enjoying all these years was directly funded by VC dollars, and these other people who they hate so much in fact do critical jobs and are necessary for converting all their work into a viable business.

[+] gamesbrainiac|2 years ago|reply
I think you read a different article. Miguel is aggrieved by the fact that it is no longer a developer focused company, which was the reason for its success.

Sure, you need sales, marketing and the rest of it; it does not mean that you need to forget what made you great in the first place.

[+] rewmie|2 years ago|reply
> (...) and these other people who they hate so much in fact do critical jobs and (...)

Just because someone is hired for a position of a sales rep/account manager/business people, that does not mean they walk over water. Those positions attract plenty of types whose primary skill is to leverage their soft skills talent to latch themselves to a organization while delivering no added value at all.

[+] afavour|2 years ago|reply
> viable business

A developer run and developer focused business could absolutely be a viable business. But it wouldn’t be the 10x hockey stick growth company venture capitalists demand and ultimately they hold the power here. So we get the inevitable.

[+] Quarrelsome|2 years ago|reply
the interests of account-managers/sales-reps and developers are often in conflict over resource allocation. Its natural for those interests to compete and I feel like brushing off one side of this coin as juvenile whining is a partisan position to take.

Its a common problem in tech and finding the balance between the two is arguably a big part of long-term success for an org in this industry.

[+] buro9|2 years ago|reply
There is a certain point at which the engineering is no longer what makes a company successful, it fuels an initial growth and can sustain it longer, but then the deciding thing on whether the company survives and thrives becomes the more operational aspects, the traditional sales & marketing.

With the product-market fit nailed, a solid offering out there, Twilio now occupies that later stage, and the buildboard reflects it. Everything changes at this point, because it's also not the top priority of the company to keep their own engineering happy at this point.

[+] kypro|2 years ago|reply
I'll try to add some perspective to help people understand Twilio's side here.

Twilio is has always been a very unprofitable company. Unlike their SaaS counterparts Twilio's PaaS model has significant costs in the form of fees which they pay to network providers. They're also loved by bad actors who can leverage their service for fraud (search "toll fraud" if you're interested) and the ongoing effort and investment needed to counter this growing fraud problem is significant. What's more there's very little moat in a communications API so Twilio has little pricing power which only compounds the cost issues they face since they cannot easily increase prices to offset growing costs of an already unprofitable business model.

Basically the business model sucks. I'd argue it's one that can only exist in a world where investors are not interested in profits. Now interest rates are rising it's becoming harder for companies like Twilio to attract investors by posting revenue growth alone.

I don't think Twilio wanted to shift their focus, they simply had to. They had to find something that had more of a moat, and therefore better margins. Segment is the answer, but ultimately it's a different product which understandably requires a different focus. If you sell a communications API then developers are the people whom you must evangelise, but Segment is a product used by business folks to grow their business.

I don't envy the position Jeff and Twilio's management team are in. They've had to make some really tough decisions, but if you like Twilio then I'd see this simply as them doing what they need to do to survive and continue providing the awesome developer focused products they provide.

I'll also note that unlike Facebook's mass layoffs Twilio's layoffs weren't simply done to increase profitability, but needed to right side a business that's currently burning over $1 billion a year.

[+] joshstrange|2 years ago|reply
> They're also loved by bad actors who can leverage their service for fraud (search "toll fraud" if you're interested) and the ongoing effort and investment needed to counter this growing fraud problem is significant.

And Twilio is wholly incompetent when it comes to battling this. Not only are they unable to keep bad actors off, they punish good actors. They require "approval" of what text messages you are going to send (never mind if part of it needs to be dynamic) and limit your sending ability to arbitrarily low levels. The company I work for just switched off them after high rejection rates and them randomly blocking us or limiting our send volume. They also bungled the whole number verification thing that recently went into effect, or at least dropped the ball on the communications side. Very glad to be off them and kind of sad, I remember when Twilio first hit the scene and it was truly amazing and so cool to work with.

[+] Animats|2 years ago|reply
Yeah, it's sad. They started as simply an SMS gateway.

I still have a Twilio account, for my steampunk Teletype setup. It's a pure reply system - you text to a phone number, that's printed on a Teletype machine, and the sender gets an acknowledgement back. It's inbound SMS. Years ago, at Twilio's request, I demoed this at a Twilio convention.

Twilio now wants me to "register my marketing campaign" and pay an additional monthly "campaign" charge for the service. Their business model no longer comprehends a pure request-reply service. They now assume their customers want to spam.

[+] firebirdn99|2 years ago|reply
Working for any company is a "job", and not a "family" as many have found out in the last few years (but happens at every downturn). It's hard to draw a line, between personal and professional identities, but it's something I feel everyone should do.

Corporations simply are built to amass profits, and outcompete others. They do not care nor are they built to care about you.

[+] shakes|2 years ago|reply
I had the privilege of working with Miguel at Twilio. His work had a meaningful impact on our goal of empowering developers around the world and I’m unsurprised to see such a thoughtful take from him has he departs. A huge loss for Twilio but I can’t wait to see what he does next.
[+] rsync|2 years ago|reply
I built an entire personal telco for myself using twilio apis and twiml bins, etc.

This was encouraged by helpful “hacker” blog posts and official howtos detailing integrations with alert systems and home automation, etc.

Twilio sold itself as useful infrastructure.

All of this falls apart next month when A2P (or whatever it is) begins.

Twilio polluted its own environment so completely that nobody can exist there anymore.

(Unless you’re spamming, which is what customer engagement is)

[+] smallerfish|2 years ago|reply
I've been playing with Twilio recently, and it's certainly not the lean platform that it once apparently was. In the Java client SDK for example, I count 7 separate Message classes. There are multiple APIs with overlapping functionality; you can send SMS and whatsapps several different ways each. The documentation is likewise byzantine and bloated.

I get that scaling tech organizations is hard (Stripe is another company with abysmal bloat in their APIs) but jfc, get a competent chief architect who is opinionated, please, and aim towards coherency. If you launch a new API that's intended to replace older ones, then stop exposing the legacy APIs in the default SDK.

(That said, better than MessageBird, who don't even have SDK support for the APIs they're promoting as the correct way; on the other hand, their documentation is markedly better than Twilio's).

[+] rabiddmeese|2 years ago|reply
> If you launch a new API that's intended to replace older ones, then stop exposing the legacy APIs in the default SDK.

They might want to alienate and break things for many of their existing clients who use the older APIs. Plenty of companies contract out dev work until they have a working product and keep it as is. They don't have devs on hand to update them if they deprecate older APIs leading to a completely broken app despite all the old functionalities being perfectly adequate for the task.

[+] mightybyte|2 years ago|reply
Anyone else share my thought that Twilio (and any other companies like them that I'm not aware of) is likely the primary driver of the massive epidemic of spam text messages and phone calls? Phone and text have been rendered almost completely useless to me. I often get close to double digit spam texts and calls per day. Obviously the drivers of this stuff are many and complicated, but it seems like the automation layer is the main thing that enabled it all.
[+] coderintherye|2 years ago|reply
The really sad thing is the executive team / board / investors not being able to comprehend that they could vastly increase their revenue (and profit) by just providing expanded technical services to me (an existing customer). There are a ton of things I would pay Twilio for beyond SMS and WhatsApp Messaging. You don't need a MBA to figure that out, you just need someone talking to your customers and an engineering and product team that can turn that into real products.
[+] cj|2 years ago|reply
Anecdotally, I think one result of the layoffs of the past couple years has been companies

1) really focusing and doubling down on existing product lines, less experiments outside core competencies

2) double down on focus on top 20% paying customers, pull back hands on support for smaller clients, focus on keeping (and upselling) existing large clients with less focus on the bottom 50th percentile

[+] erhaetherth|2 years ago|reply
FWIW I never thought Twilio was good as a developer. At least not coming hot off of Stripe. Which is glorious. The dev mode is fantabulous, and so are the docs. Twilio... I.. it's just atrocious in so many ways. Firstly, it needs a proper dev API key and then all the SMS's should just be collected into a dashboard on their site instead of going to phones. And their Connect option is very half baked. Customers should be able to re-use their existing Twilio numbers, not re-buy under a connected app, and they should be able to manage their numbers on twilio.com, not force me to rebuild the entire Twilio UI.
[+] predictabl3|2 years ago|reply
The writing was on the wall when they got acquired, forced two-factor auth with SMS only and forcibly locked me out of my account by turning it on when I didn't have a phone setup. If you want to use SMS as spam prevention, then just do that. Breaking your product and forcing bad security to harvest numbers is, well, just shitty.
[+] arnvald|2 years ago|reply
When who got acquired? Twillio? They're a public company, aren't they?
[+] wilg|2 years ago|reply
1. I could not tell you what that billboard means, despite also looking through the related blog post. Perhaps it is something only marketing teams understand, which would be an interesting failure mode.

2. I tried to sign up for Twilio Sendgrid the other day to send some emails from a Heroku app but I couldn't complete the sign up because I never received the activation email. Classic!

[+] ds0|2 years ago|reply
Tangentially, I'd like to shout out both Miguel's work both in explaining the Flask web development framework for Python as well as his work developing Flask-SocketIO, both of which I've used extensively.
[+] fastball|2 years ago|reply
I've had to move away from Python-SocketIO due to countless regressions and subtle failures of the library.
[+] miguelgrinberg|2 years ago|reply
Thank you so much. Glad to hear you've found my work useful.
[+] kelnos|2 years ago|reply
I was at Twilio for 10.5 years, and left at the beginning of 2022. It's interesting to read things like this, from someone who joined in 2019, as the change OP is talking about in the company started well before his arrival. We all saw it, with increasing emphasis on enterprise sales (we hired someone from Salesforce as COO specifically for this). Courting individual developers and small startups stopped being a priority, even though those sorts of folks are the ones who formed the backbone of Twilio's mind share.

I'm not saying this was bad thing for the business; Twilio's growth would have slowed well below what would have been acceptable for a VC-backed (and the newly-public) company without this move. I'd be lying if I didn't say I was happy with where the value of my equity went. On top of that, Twilio was life-changing socially: I have many dear friends who have remained dear friends well after we all left the company, something that's never really been that case for me at other companies. As for the work itself, there were many fun and challenging problems to solve over time as the company grew, despite our frustration with the growing pains.

In the end, I left for very similar reasons as OP: the company seemed more focused on helping brands increase engagement than on doing anything novel with communications. Sure, a lot of the older, interesting use cases (the ones that got me stoked about Twilio in the first place) were still running on the platform, but those kinds of things felt like a minuscule percentage of the business.

The change of the billboard to the "How can I reduce my acquisition cost" branding makes me feel sad any time I drive by it. "Ask your developer" was brilliant, but Twilio isn't that company anymore, and hasn't been for a while.

[+] 29athrowaway|2 years ago|reply
It seems the MBAs are at it again.

https://www.youtube.com/watch?v=P4VBqTViEx4

Twilio is now another Xerox.

MBA programs should be illegal. You can thank MBAs for spreading PFAS everywhere, single use plastic, profiteering from insulin, making every American a diabetic by adding sugar into everything, selling glyphosate for residential use, killing all the pollinators, moving all manufacturing to countries with no environmental laws, and other consumer and planet-fucking initiatives.

The number of extinct species, diabetic patients, number of employees on minimum wage with no health insurance, policitian revenue from lobbying activities, dead bees, cut trees, lost topsoil, former Roundup users with non-Hodgkins lymphoma, revenue from weapons, atmospheric temperature increase, ocean acidity, plastic waste, pollution, every form of ecocide and every other measure of things going wrong is directly proportional to the number of MBAs.

Why? because if shareholder value is everything, then the employee, the environment and society at large becomes irrelevant.

[+] SamuelAdams|2 years ago|reply
Alright fine. Twilio has largely saturated the communications space. You can now send an API request for a phone call, SMS message, and WhatsApp notification in a lot of countries.

If you were calling the shots, what would your next growth move be? What should Twilio expand into?

[+] cyberax|2 years ago|reply
> selling glyphosate for residential use

What's wrong with glyphosate?

[+] INTPenis|2 years ago|reply
As a Swede I naturally gravitated towards the 46elks.se service instead of Twilio. Been using it for years now, no issues. They even called me once to find out what their clients were doing, like a survey.
[+] taf2|2 years ago|reply
This is unfortunate that he got disillusioned but the shift is for really one simple reason - segment. The original twilio that went on to acquire sendgrid, was increasingly in a low margin commodity market. Segment is hugely profitable in contrast and is not dependent on increasingly competitive upstream providers (carriers) providing lower cost similar apis as twilio voice/messaging. It still stands that Twilio’s apis are better than the competitors but that will only last so long. Segment will keep twilio relevant for a much longer time.