Hacker News Reader

Twilio went from $33m to $59 Q1 '15 to Q1 '16. That's huge growth - and is a lot to like.

Virtually every company that is growing at 80% YoY runs unprofitably.

Here's why: compounding. They're growing 80% YoY because of that spend and it's absolutely worth it.

Say they could run break even at 50% YoY growth. Straight line that growth through 2019 and you have a company doing just shy of $800m in annual revenue. Invest $40m extra a year for 80% growth (admittedly very simplified!), and you have a company doing $1.4B per year.

Growth costs money up front: it's marketing, it's customer acquisition, it's hiring and scaling so you're ready to do nearly double the volume next year and it's opening new lines of business.

You'd be nuts not to spend some extra cash when a) unit economics are good and b) you can turn it into significantly more cash in 12-24 months.

Twilio could run profitably tomorrow at the cost of hundreds of millions in revenue and potential profit 3, 4 and 5 years out.

Disclaimer: Jeff Lawson is a CEO I greatly admire.

There was a great article that gets posted here often about how it's ok to be unprofitable if you have high growth, and it's ok to have low growth if you have high profits. Anyone have a link?

Edit: found it. http://avc.com/2015/02/the-40-rule/

Net loss: Don't think that is a biggie. They expanded their cost base both in R&D, sales & marketing, and general and admin. With a little bit of cost discipline for one year, they are going to be profitable.

Voting rights: that is indeed yucky.

Competitive threat: Not really. The inertia for existing customers is high, especially since the average revenue per active account is about $580/month. For a big company, that is not worth optimizing. And switching to another (unproven) API from a running system... not likely.

>a little bit of cost discipline for one year

So, layoffs?

Or just continued growth without adding headcount.

A list of their competitors: https://www.quora.com/Who-are-Twilios-competitors.

As a Twilio customer that spends maybe ~$500 /mo i'll look at any alternative that allows me to send meta data or allows some process to have a foreign key allowing me to persist conversations in a sane manner (and I'm currently researching just that).

Twilio also does not have their own network, so competitors like Bandwidth that have their own network will be able to compete on price. That said current Twilio pricing is not prohibitive in our market and thus isn't a factor for us researching alternatives currently.

I suggest that you look at Tropo (https://www.tropo.com/). They provide the option to run your script on their servers. I find it's much easier to program non-trivial conversations that way.

Twilio is a Bandwidth customer.

Bandwidth launched a competitive product over two years ago: https://catapult.inetwork.com/pages/home.jsf

The fact that no one here has mentioned it is telling.

Also many of the SMS and MMS companies sign contracts for X messages per month with the telecom companies. If they have a slow month where they can't fulfill the obligations of the contract they'll resell their remaining capacity to another company in the space.

> This is surprising, since they don't have any big capital investments.

I haven't looked at the S-1 and I'm not sure if they even break this out, but user acquisition costs are all paid up front, while the value is realised over the life of the customer, depending on the average lifetime of a customer, this could be fine, if risky. And you often do prepay for AWS since it saves you money over the course of a few years.

Right but how expensive is that acquisition cost? So its about the average cost to acquire versus the average lifetime customer value. I think that's an important metric for them to break out

Sales & marketing is 49 million, R&D is 42 million. So, I guess the critical factor is the degree of user turnover. If they lose users too quickly, they need to keep spending big on sales & marketing. It doesn't seem like the sort of market where people jump ship quickly though, since once you've added the piece of code that integrates with their services, you'd only replace it if the other api is a lot cheaper or somehow more capable / reliable.

> And you often do prepay for AWS since it saves you money over the course of a few years.

I don't think this would show up as an expense. It would just shift assets from cash to prepaid expenses, and not show up on the P&L - until the expenses are realized, which would be over time.

Let's play the devil's advocate for a second.

                       FY13     FY14     FY15 
  Sales and marketing: 21,931	33,322	 49,308

Just turn it off and now you have a $11M profit. NetSuite and SuccessFactors did it in 2008 and they're both profitable, or at least were in NetSuite's case (NetSuite has recently pushed hard on S&M, hence a decline in profit in the last few years, probably because the market is dictating to do so and money is cheap).

EDIT: Grammar.

I think this is a smart move and a good observation. Specifically with respect to Twilio, because I already know about their services. Having needed this kind of integration as a business need previously, I am aware of what services they are offering. What I'm saying is that I don't think this is a discovery problem for Twilio. All of their marketing and sales efforts seem to be focused on asking me to get to know Twilio's services better. But I already know what they offer (I just don't want to buy a short code). I'm pretty sure most other people shopping for this kind of integration are in the same boat (already aware of what Twilio is selling). Spending money to make sure people know what Twilio is selling seems redundant to me. Just for one point on the graph.

I'd be surprised if they are targeting just developers and people knowledgeable about the tech.

A lot of that is probably focused on reaching PHB's in enterprises who don't know what Twilio is or why they should care until a sales person takes them to an expensive lunch to explain it.

Just want to note that yes, cost can rise in advance of growth. Whether or not that's "OK" depends on the unit economics.

Frankly, I share your skepticism. I think you hit the nail on the head - Twilio has few, if any moats, that can't be crossed by large competitors like Google or Amazon.

And if I'm Bezos and Twilio is making all this money off the back of AWS, why not just cut out Twilio? Heck, Amazon is doing that to friggin vitamin makers, why not do it to tech companies?

> Twilio has few, if any moats, that can't be crossed by large competitors like Google or Amazon

Here's one moat - with GOOG's annual revenue of $75 billion or AMZN's $107 billion how likely are either to chase an opportunity that, assuming perfect execution, complete market share grab and ultimate destruction of Twilio, will add $167 million of annual revenue to their balance sheet?

From the corporate politics standpoint this would be a hard project to pursue internally, as anything under a billion is unlikely to register on CEO's radar.

From the customer standpoint, chances of getting competent and prioritized support might be better with Twilio, where the telephony API is their primary product, versus Google, where it might be 108th project on the priority list.

This feels very much like something Amazon might pursue, although probably in a more piecemeal manner. AWS already has a lot of the components that would be needed for a Twilio competitor in the form of SQS, SNS, and Lambda - all they really need is the ability to receive incoming text messages and voice support to be competitive. A lot of people would jump on that just stay within AWS rather than add another provider to deal with.

I agree there are some fundamental building blocks they might pursue. There's already a parallel story to follow - SES and Mailchimp/Mailgun/Sendgrid. Competitive at the basic level, with Amazon actually validating the market and sending the clients off when they're ready to go beyond "simple".

Their large clientele is all sorts of spamming services. I receive tons of messages daily (currently ATT but Sprint before) for different type of products and services, mostly work by completing surveys. Everytime I trace number down - its Twilio. And their spam complaint team never answer to my complains, so unfortunately as long as spammers can use their services, they will be making money.

On the good side, I use their services anytime I need to "confirm my identity through a text message".

I think its impossible to determine whether a cellphone is a real number or Twilio number so it always works! I have a cash prepaid Visa card that I can purchase a Twilio number for $1 and have a text message with confirmation code come to my throwaway email address. Well worth a dollar for almost complete anonymity!!

It's not impossible. It's quite easy in fact to determine if a number is a real cell phone or not by doing an NPAC lookup to see if LNP is applicable (if so, you just check the service provider type, if not, a simple check of the NANPA allocation would do). Lots of companies provide it as a service. Just most places don't check and blindly try to deliver to whatever number you provided.

I don't understand this. Is it a land-grab focus?

Why do so many software companies focus solely on growth? A traditional business with $166 million in revenue is huge. Why are investors not asking for a return on their investments already?

Don't investors usually look for a return on their investments within 2-3 years?

>Don't investors usually look for a return on their investments within 2-3 years?

no

Return can be two things: profit now or future profit.

A traditional business with $166m in revenue has few places to invest extra cash that turn $1 into $4 in < 24 months. Twilio has this.

What would be crazy is not investing in growth.

The stock price is a result of many factors, future profit and growth is a big part of it. Having a good long term outlook is better than slashing everything for the short term.

I am a Twilio API customer. It definitely is a land rush. I am far less likely to switch to its competitors as my integration with their systems grows deeper. There is a finite number of large enough companies that would materially affect Twilio's revenue. 166million at $0.015/minute is not that much of business compared to what they are aiming for, which is global domination of computer-controlled telecommunications.

Personally, I'll stick to a rule I found very valuable with the first dot-com bubble: don't touch a stock until the company has had 4 quarters of profitability. Simple rule. Amazing how companies these days never pass that filter.

So you don't touch any growing company until they plateau? Doesn't that defeat the whole point of buying stock?

What makes you think that a company would plateau after 1 year of profitability?

I don't think Google is much of a threat, they seem to hate Gvoice internally having had several opportunities to make it a platform and passing (yes they used to have voice APIs on code.google.com). Bottom line, I don't think Google gets it (kind of like Facebook and web search)

And since the Unicorn Feeding stations have all been shut down :-) unprofitable companies of this size have three choices, IPO, die quickly, or try to shrink into a profitable chunk and bootstrap slowly into larger growth. I commend them for the IPO route, I expect to see more companies in this spot. This path also gives employees some liquidity but if they did a big reverse split prior to the S-1 its possible the employees will have underwater options at the IPO, anyone working there care to comment on that?

Given the way you laid out those scenarios, would you agree then that the goal then of IPO is finding a greater fool on Main Street and passing the buck while they can?

Is there anything here that indicates that the cash infusion from an IPO would get them on the track to profitability?

I have no experience with IPOs but as a Twilio customer they've been pushed a revamp of documentation, client libraries, their user interfaces, as well as several new products in the past ~6 months. My guess is an IPO would enable them to continue that. That or they just did it for a little hockey stick in revenue before the IPO.

Yes, in general the bargain is that in exchange for accepting some additional constraints on behavior and a required level of transparency, you can sell securities to non-qualified investors (aka the Greater Fool in your comment).

That said, I'm not sure I would be willing to invest at the IPO but if Twillo can show they are getting traction on their business plan and reach profitability, then I would consider investing in them at that stage.

>if they did a big reverse split prior to the S-1 its possible the employees will have underwater options at the IPO

Any past examples of this?

Lots. One I was personally familiar with was SSNI (Silver Spring Networks)

Can you elaborate on the exact financial implications of a reverse split prior to IPO? This Brad Feld article [1] seems to say the aversion to this is simply emotional.

[1] http://www.feld.com/archives/2005/06/when-are-350-million-sh...

Company, running privately for years and years, marginally cash flow positive but requires additional capital to expand, or develop the next product, or build inventory. Employees get refresher grants annually with an ever increasing strike price from .10 - 3.50 over say 8 years. Now during the road show the bankers say "You're company is really worth about 1/10th what you think its worth according to this feedback, we either cut off the roadshow or we recaptialize at a lower valuation." So the company does a 50:1 reverse stock split to get its outstanding stock numbers in line with the expected valuation of the company.

I'm not an accountant, but I've listened to a lot of accountants explain what is, and what is not, taxable. In this case, if you let your employees take a haircut (so their $1/share strike price is now $50 a share on a company expected to go out at $18) that isn't taxable (and it bites to be an employee). If your employee exercised their shares at the lower cost (to avoid capital gains etc) they now have a 'basis' value of $50 a share so if they sell shares at $18 the can claim a $32 / share capital loss. If you take back the vested but unexercised shares and issue vested shares at a new lower price, that price can be no lower than the pending IPO price (FMV) or it's a taxable event. And then when your company makes it out the gate and your long suffering investors cash out their funds, it pushes the price down around $8 once again putting your employees in a position to exercise at a loss or leave them on the table.

The key though is the company went through a period higher valuation, and employees are issued shares at the higher valuation and at IPO time the company is worth less than it was when you got your option, so your strike price is "high" relative to the company value.

tesla is a good example of a company that did this at ipo.

really, there are no financial implications. the valuation of the company stays the same with a split, but the price of the stock changes proportionate to the quantity outstanding. valuation = # shares * $ per share

what the reverse split does (or split - as twilio did a 2:1 split just over a year ago) is put the cost per share in a range that appears more attractive to buyers. in today's market, that is ~$15 +/-5. so if the internal valuation pre-IPO has the stock at $8, the company often reverse splits 1:2 and then you have an initial price back at $16. that way you don't look like a penny stock going out the door.

IANAL but citation needed. An ISO is granted against a certain point in time, and would be affected (as in, included) by a split at a future time. So in a 2:1 split you'd get double the shares at half the price. Other way simply doesn't work as that would screw your cost basis up and IRS would get all green and angry at you :)

The easiest way to screw people is to issue another set of shares and dilute.

Re the options, the S-1 says: "As of March 31, 2016, we had outstanding options to purchase an aggregate of 16,704,752 shares of our Class B common stock, with a weighted-average exercise price of approximately $5.57 per share, under our equity compensation plans. After March 31, 2016, we issued options to purchase an aggregate of 671,550 shares of our Class B common stock, with a weighted-average exercise price of $10.30 per share, under our 2008 Plan."

Assuming they're going to IPO at more than $10 per share (which is usual) and that the option strike price has not gone down (so all the options issued prior to 3/31/16 were at a strike less than $10.30 per share) it looks like almost all the options would be in the money to some extent.

> It's unusual to try to pull that off when you're unprofitable.

It may be unusual, but to me it seems even more important to have uncontested control when the common shareholders may be angry.

Here's what that looks like. Twitter's stockholder meeting was yesterday.[1] The stock is down 80%, Twitter is losing money, even Instagram now has more users, and the shareholders are very angry.

Twitter (TWTR) only has one class of stock. But so far, there hasn't been a major effort to oust Dorsey.

[1] http://www.ibtimes.com/twitter-inc-twtr-shareholders-meeting...

My first question would be: how hard would it be for them to become profitable right now?

That translates to: how manual is their business? How much cost in the form of wages, contractors, etc. could they shed?

If the answer is that they could easily become profitable, that means they're leaning into growth. If the answer is 'no' that's a bad sign.

> So they're still not profitable

Why would a company want to raise money (IPO) if they had a high net profit margin? If they're profitable and happy about it, the insiders could just dividend out the money.

If you had to guess what the net loss for Facebook at IPO was, what would you say? More or less than Twilio?

Per their S1, Facebook did $3.7 billion dollars in annual revenue in 2011, and had a net income of $1 billion. And they were growing that at 100% annually.

I'd say the comparison is pretty apples/oranges.

Uncommonly good distillation there. Thank you.