Description

Twilio (TWLO - $28) develops and supports APIs/products that allow developers to embed SMS text, voice and video communications in their applications and is the largest stand-alone company in the Communications Platform as a Service (CPaaS) space. TWLO recently popped up on my accounting/fundamentals-based short screen. I think the business model will turn out to be weaker and it will take much longer to reach profitability than expected, though this will take some time to play out. ATM’s VG write-up on VIC and messages provide good background.

To summarize: At 7x 3Q17 run rate revenue, expectations are high for TWLO. But the precipitous decline in revenue from two of TWLO’s largest customers, increasing competition, and what I believe will be disappointing future operating leverage, suggest TWLO may have trouble reaching profitability and will disappoint investors.  The company has also become more aggressive with its accounting, something that we often see together with increasing pressure on the business. Because the Street has revenue growing 25% in 2018 versus the company’s recent 40% revenue growth in 3Q17, I don’t think there’s risk to Street revenue numbers next year, at least right now. Putting a valuation on the business is admittedly tough. I see this business as a difficult one that won’t achieve the margins investors expect. The company’s operating loss is increasing as it grows. I think if I’m right on the business model and increasing pricing pressure for TWLO, the stock goes a lot lower. For point of reference (but maybe not a great one), VG paid about 3x revenue for TWLO’s closest competition, Nexmo, which had $38m of revenue in 3Q vs TWLO’s $100m.

TWLO’s two largest customers in 2016, Uber and WhatsApp, have sharply curtailed their business with TWLO. Uber went from a 17% customer at $14m in revenue in 4Q16, down to a 9% customer at $8m of revenue in 2Q17 as Uber brought communications in house and moved business from TWLO to competitor Nexmo.  On the 1Q call, CEO Lawson stated “they [Uber] plan to move communications for some of their use cases in-app” and “their tremendous growth and the resulting magnitude of their communication spend has resulted in this change and approach”. It's hard to sugar coat "tremendous growth" at Uber and declining revenue at TWLO. Subsequent to the loss of business to Nexmo, TWLO in 2Q “readjusted” prices for Uber. In 1Q Nexmo also took Lyft business from TWLO in what Nexmo called a “competitive takeaway win”. While we’re on ride-sharing companies, Gojek was TWLO’s second largest customer in 2Q17 at 5% of revenue, but in 3Q TWLO failed to mention them and it sounded on the call like they were well below 5%. Gojek is an Uber competitor in Indonesia. It will obviously be informative to watch what happens with the Gojek business given what happened with TWLO’s other two ride-sharing customers.

TWLO’s revenue from WhatsApp fell from 17% of revenue in 2015 to 5% in 2Q17. WhatsApp is a “variable customer account” which TWLO notes are customers that decide “not to enter into contracts with us that contain minimum revenue commitments, even though they may spend significant amounts on the use of our products and they may be foregoing more favorable terms often available to customers that enter into committed contracts with us.” My interpretation is that variable customers decline more favorable terms because they don’t want to sign up for a long term commitment, thinking TWLO is a shorter term solution.

I don’t think these are isolated cases. Since TWLO has lost significant business from both of its largest customers (and Lyft, which could have been big) probably due to price, it begs the questions of whether any customers will scale with TWLO, and if they do, will it be uneconomic business? The CEO of VG states that Nexmo has 70% lower costs than competitors on the voice side (link first posted by Smokey - thanks):

http://www.nojitter.com/post/240171954/vonage-amps-up-with-nextgen-nexmo-voice-api

“Nexmo has long had a voice API, and global reach with the ability to provision local numbers to 85 countries. But now that's married to a huge backbone, terminating some 15 billion calls annually, with direct peering relationships with other carriers, Masarek noted. In addition, Vonage owns the phone numbers and so use of the Nexmo Voice API does not require a "rental" fee for a number.

"These bring all sorts of cost and quality advantages to bear," added Masarek, who said he believes Vonage's costs can be as much as 30% of what competitors without direct peering relationships charge.”

There’s really no leverage to be had on TWLO’s cost of revenue and there won’t be. And TWLO’s GM% is getting squeezed via price competition. Since TWLO services are priced higher than competitors and customers (e.g. Uber) have shown willingness to bring business back in-house, it seems to me that there is significant risk to TWLO’s gross margin. I don’t think the sharp drop in 3Q07 GM% is an isolated case. From the 10K: “In 2016, cost of revenue increased by $46.1 million, or 62%, compared to 2015. The increase in cost of revenue was primarily attributable to a $40.0 million increase in network service providers' costs, a $2.8 million increase in cloud infrastructure fees to support the growth in usage of our products and a $1.9 million increase in amortization expense for internal-use software.” So 93% of the incremental cost of revenue was variable and outside TWLO’s control. Incremental gross margin, vs the year ago period, dropped from 58% in 2016 to 55% in 2Q17 and a very low 42% in 3Q17. 3Q17 gross profit dollars were down 2% from 2Q even though revenue was up 5%. The company blamed the drop in 3Q GM% on mix, Uber (lower pricing I assume) and FX as if these were temporary issues, but then guided to similar GM% in 4Q.

Here’s the company’s description of cost of revenue (almost entirely variable costs out of TWLO’s control): “Our arrangements with network service providers require us to pay fees based on the volume of phone calls initiated or text messages sent, as well as the number of telephone numbers acquired by us to service our customers. Our arrangements with our cloud infrastructure provider [AWS] require us to pay fees based on our server capacity consumption.”

This is an informative read about pricing, among other topics, from a CTO that evaluated TWLO vs competitor Plivo at the beginning of 2016: https://www.quora.com/I-am-evaluating-Twilio-and-Plivo-Does-anyone-know-which-carriers-they-use-Any-feedback-on-the-quality-of-audio

Operating expenses have been growing faster than revenue, with opex at 75% of YTD revenue in 2017 vs 70% for the same period a year ago. Leverage, but in reverse. I get it, same story for many cloud companies. But I think this will matter if it becomes clear the business model is impaired.

Accounting and Financial Trends

Incremental gross margin and gross margin trends discussed above

We won’t’ have 3Q numbers until the Q comes out, but we can see that TWLO’s sales credit reserve (which is deducted from gross revenue) and its allowance for doubtful accounts (which is usually in G&A) dropped significantly y/y from 2Q16 to 2Q17. The sales credit reserve dropped from 6.3% of gross rev to 4.4%; had the company maintained a flat ratio, revenue would have been $1.8m (2%) less in 2Q. The allowance for doubtful accounts dropped from 3.3% of AR to 2.4% from 2Q16 to 2Q17. This isn’t a huge red flag compared to what I’ve seen elsewhere, but it is very unusual for a company that is not seeing pressure on its business to drop these ratios like this.

TWLO’s revenue from it’s largest customers went from growing 56% in 2016 to only being up 1% in 2Q.

Catalyst

Dissapointing earnings in 2018; continued GM% declines.