At UT, I was part of an student investment team called TUIT, where among other things, we pitched public equity investing ideas. In April 2020, me, Emory, and Carson (two other members of my pitch team) pitched long $TWLO at ~$84 a share. Below is the report we wrote on why.
Company
Twilio (NYSE: TWLO) is a CPaaS (communication platform as a service) provider that sells API and turnkey products that companies use to interact with their customers across the following channels: phone/voice, SMS/text, online messaging (eg. Messenger and WhatsApp), and email – added most recently through the acquisition of Sendgrid in early 2018
Thesis
Due to its continued reinvestment strategy and recent SendGrid acquisition, Twilio owns the best-in-class product suite in digital communications software and is also a stated market leader in the cloud communications industry. Its existing customers across SMBs and enterprise have demonstrated continued annual spending growth, reflecting a developer-centric approach and continued upselling opportunities as client cohorts expand.
Accelerated by the rise of 5G, cloud communications-platforms-as-a-service (CPaaS) models are rapidly taking share from legacy on-premise communications solutions, a market worth upwards of $35B. Twilio’s turnkey and API products place it in a leading position to champion this growth and maintain a copycat-resistant moat.
Background – Cloud Transformation & Infrastructure-as-a-Service:
In the context of the growth of cloud infrastructure, it is important to understand Twilio as a direct parallel to AWS, which emerged early as the leader in infrastructure-as-a-service (IaaS).
In the early 1990s, on-premise computers/servers were required to set up a website. By the mid-1990s[1], a web developer could run sites on many web hosting services. Around this time, the development of the VPN (virtual private network, which allows access to a private server network over public internet) allowed for shared computing power over public internet which was termed the “Cloud” [2]. In the early 2000s, Amazon created a set of APIs[3] in response to development challenges with Merchant.com, which was an effort to allow retailers to build online shopping sites on top of Amazon’s e-commerce platform.
Management realized that developer teams were “reinventing the wheel” each time instead of leveraging shared resources. This led Amazon to create APIs for “compute, storage, and database” internally. Amazon management soon realized the broad marketability of these APIs and set out to build an “internet OS.” This initiative became AWS, a pioneer of internet infrastructure as a service.
AWS allows developers to build applications with APIs that access pre-built functionalities and powerful cloud computing. This saves on development time because developers don’t have to reinvent the wheel. Moreover, it eliminates the capital and operational cost barriers of compute infrastructure, giving users easy scalability through a pay-as-you-go (OPEX-, rather than CAPEX-based) business model. Many novel and highly scalable applications/companies have been built on AWS that would have been impossible – or significantly delayed – otherwise, including Airbnb, Lyft, Expedia, Netflix, and Dropbox.
AWS is a powerful business model because it provides it gives companies easy access to scalable infrastructure. Companies simply lack the ability to build comparable infrastructure on their own, and even if they could, they could not match AWS’s economies of scale. Several years after AWS was created, competitors grew to include Google Cloud, Microsoft Azure, and IBM Cloud. However, AWS maintains a market-leading position and has rapidly become Amazon’s most profitable business, responsible for 63% of Amazon’s operating income in 2019:
This raises an important question: why does Amazon still dominate AWS, and why is it so profitable? Shouldn’t cloud IaaS have become a commodity by now with such well-capitalized competitors? When applying economic perspectives from the industrial past, this makes sense – but digital IaaS also carries with it some fundamental differences:
- Infrastructure platforms are inherently highly entrenched. Developers would have to rebuild applications built on AWS if they wanted to switch, unlikely for existing client giants from Netflix to Twitter.[4] Because of the relative success of these early adopters and the periphery businesses they’ve enabled in cloud consulting and AWS tool management, newer builders are favored to choose AWS for its larger network of support and robust feature set.
- AWS’ early and aggressive investment created barriers to entry. Amazon was the first mover and invested heavily, leading to a virtuous cycle of reinvestment, improvement, and economies of scale. Despite being some of the most well-capitalized companies in the world, Microsoft and Google still lag because they cannot account for all the lost user-iterated time.
- The space has become undoubtedly more competitive. Even with Amazon’s well-positioning, newer competitive markets are transforming some of the current supplier bargaining power and control into consumer benefit, through both server diversification and price negotiations. For example, Zoom uses a combination of AWS, Azure, Oracle, and its own servers.
[1] https://biztechmagazine.com/article/2012/02/history-web-hosting-infographic
[2] https://www.wired.com/1994/04/general-magic/
[3] https://techcrunch.com/2016/07/02/andy-jassys-brief-history-of-the-genesis-of-aws/
[4] https://www.contino.io/insights/whos-using-aws
Thesis – Twilio & Communication-Platforms-as-a-Service:
Cloud transformation in communications follows the Amazon-pioneered growth story in infrastructure, leading up to the development of communication-platform-as-a-service (CPaaS).
Historically, companies interacted with customers largely through physical venues or cumbersome phone-based contact centers. Contact centers required massive capital investments in servers, networking equipment, and licensed software. Moreover, they required armies of developers and IT personnel to keep them up and running. Building and maintaining communication infrastructure was expensive and challenging. From 2004 to 2006, Jeff Lawson (Twilio’s founder and current CEO) worked on AWS products at Amazon. In 2008, he founded Twilio based on the importance of “integrated communications to produce a “differentiated customer experience.” Around this time, he said this of software:
“Software is now moving from the backoffice to the business model,” he said. “Every company needs software to build a differentiated offering.”
Twilio’s goal was analogous to that of AWS: to make well-documented APIs that would enable developers to easily implement communications in their apps without having to reinvent the wheel. Like AWS, Twilio’s early success gave rise to several competitors that also do communications API: Tropo, Nexmo and Plivo, founded in consecutive respective years starting 2009. Tropo was acquired by Cisco in 2015. Nexmo was acquired by Vonage in 2016 as part of an acquisition spree that included 8 other CPaaS, CCaaS (contact center as a service), and UCaaS (united communications as a service) companies.
Just as IaaS allowed companies to build better software products on an OPEX model, CPaaS allows companies to build better customer communication channels on an OPEX model. This happens in two ways:
- API Implementation. An API is a tool, or a building block, for software. CPaaS APIs allow developers to integrate a feature like SMS communication (allowing a program to send/receive SMS messages) with just a few lines of code. This requires far less work for the developer and allows him to focus on building features specific to his product.
- Turnkey products. Turnkey products are prebuilt solutions for specific applications, namely call center. Turnkey solutions are naturally less flexible, but a company can get one up and running in little time and with little-to-no developer man-hours.
In both cases, companies choose CPaaS because it offers better functionality than legacy products and a better cost structure than developing those capabilities in-house. Why are these communications capabilities so important? Every industry has become increasingly customer centric, led by the rise of technology aggregators [1] that have commoditized once-siloed processes on the supplier side. Now, companies must compete less on their supplier relationships and more on their communication and understanding with customers. In this era of true customer-first business, having digitally enabled, efficient, and scalable communications channels is a requirement, not a luxury. Twilio offers solutions for the bulk of these use cases:
- Migrating contact centers & corporate telephony to the cloud: Companies are replacing legacy on-premise customer contact centers with cloud-based solutions (API-based and turnkey) that offer features like better voice recognition and more efficient call routing. Cloud costing for these processes is more affordable as it scales with usage.
- Lyft built a cloud call center with Twilio flex (turnkey solution) to provide support to riders and drivers
- Interacting with customers through apps: Existing companies are using APIs in mobile applications that increasingly serve as their primary medium for customer interaction/service. As these channels scale, companies will grow their recurring spend.
- New Software-native business models: Developers are using APIs to build new products that operate across various communication channels.
- Zendesk, a helpdesk CRM SaaS provider, uses Twilio voice APIs
- Uber sends ride status SMS notifications to customers with Twilio
This market falls into two distinct categories: API and Turnkey products. API is still Twilio’s core business, and it is facing increased competition from copycat providers. Like AWS however, Twilio’s early leadership and investment creates a larger barrier to entry:
- Twilio is highly entrenched. Developers would have to rebuild Twilio-based applications if they wanted to switch, making it a high stickiness product. Twilio has also focused aggressively on developing mindshare within the developer community.
- Twilio’s early and aggressive investment creates barriers to entry. Twilio developed a best-in-class product line[2] over several years of aggressive investment and has become the “worldwide leader”[3] in CPaaS. As it grows, this scale will compound: SendGrid alone sent over 60 billion emails in 2019, and Twilio has a 10% penetration of companies in the Global 2000.
Twilio will dominate CPaaS if it continues to invest ahead of its competitors. Like with AWS’s cloud infrastructure, eventually companies will try to replicate Twilio’s direct product offering. However, the commonalities end here; while AWS’s competitors are well-capitalized platforms backed by tech giants who dominate sizable respective markets, Twilio’s primary three copycats are more antiquated (two were acquired by dinosaur businesses) and therefore not meaningfully threatening.
[1] https://stratechery.com/2015/aggregation-theory/
[2] Bloomberg Intelligence
[3] https://www.twilio.com/press/releases/idc_marketscape_2019
Industry – Why CPaaS is Superior to Legacy On-Premise Alternatives:
In 2011, Marc Andreessen coined the phrase “Software is Eating the World”[1] and predicted that advantages of digitization would fundamentally change critical industries, from already ongoing disruption in commerce and manufacturing to long-standing legacy sectors like education and healthcare. This insight was more than prescient, and “software-powered insurgents” have demonstrated their competitive ability from now common stories like Uber and Airbnb to rising tides of fintech providers.
As companies continue to undergo this “softwarization”, they will need more application-to-person (A2P) solutions to manage interactions through their products and customers. Rather than build each of these capabilities in-house at restrictively high costs, they will become clients to companies like Twilio that provide a pay-as-you-go service that scales with demand. This competitive edge positions Twilio to take advantage of the scale of growing communication markets and establish itself early on as a “tollbooth” by becoming CPaaS provider of choice.
CPaaS is taking over A2P solutions because platforms make it easier and cheaper for companies to access what legacy, on-premise solutions do[2]. They also unlock new capabilities: under the larger theme of digital transformation, companies prefer cloud compatible communications that can adopt for new media channels[3].
Products like Twilio Flex consolidate and integrate traditionally disparate channels at a usage-based cost structure, disrupting the high capital expenditure barriers to entry in legacy products and serving as a conversion funnel into cloud-based communication. For call centers, the most prevalent customer interaction channel, this transformation to the cloud has been exponential. Twilio lies as the crucial API provider and CPaaS player that services different types of cloud-based call centers:
[1] https://www.wsj.com/articles/SB10001424053111903480904576512250915629460
[2] https://www.prnewswire.com/in/news-releases/service-providers-increasingly-turn-to-communications-platforms-solutions-to-quickly-deliver-real-time-communications-capabilities-to-enterprise-customers-855841301.html
[3] Melissa Knox, Morgan Stanley Software IB: https://www.morganstanley.com/ideas/rise-of-the-tech-super-platforms
This leadership in cloud contact centers indicates large growth opportunity as the market itself develops over the next few years due to migration from on-premise services. While this shift is more gradual, Twilio’s access to 7 million developers and existing clients that are already looking to migrate can enable consistent and targeted leadership that will help it take the lion’s share of this transition.
Below, we see that Twilio’s current addressable markets do not rely on this rapidly growing cloud-based market, despite ongoing growth at the cost of on-premise alternatives. Yet the diversity and scale of Twilio’s total TAM provide it the revenue security to be a leader in capturing the growth of cloud contract markets and email applications, even with programmable messaging constituting 48% of revenue today.
Risks & Mitigants within CPaaS:
- Large customers like Uber could insource CPaaS solutions and wean away from Twilio revenue contribution
While this has always been possible, developing a comparable, complete set of software in-house is difficult for several reasons, not just because communications regulations and frameworks are wildly different internationally. Creating software packages and communications standards that apply and fit within these different regulatory landscapes is a time and capital-intensive task that often strays far from company’s core capabilities. At any rate, Uber’s decision was fueled by cost optimization relevant to specific geographies, as opposed to a reaction to the Twilio platform. In fact, the company remains a base Twilio customer and still commits to related annual contracts[1] with Twilio and its competitors.
- Adoption in near-term will likely be slower due to economic downturn
Predicting this pandemic’s impacts is a multi-trillion dollar question, but Twilio’s value doesn’t come from a short-term year-long horizon anyways. As costs are cut across the board, companies from SMB to enterprise will internally account for inefficiencies with their existing suite of services and tools. This might actually propel a shift to a less leveraged (operating-wise) cloud solutions and platforms like Twilios CPaaS. As customers have increased attention to their devices and digital personas, communicating across all tech-enabled media will become a priority, creating a long-term growth opportunity.
- Platform specific improvements on messaging apps could reduce CPaaS TAM
Developers of leading messaging apps are also trying to expand vertically for different use cases. Recently, Apple introduced Business Chat, which poses a substantial threat as an emerging type of competitor in the CPaaS market. Due to its innovative nature and existing access to a large consumer base, this threat is perhaps the most credible among all messaging players. But, it is unclear how likely other messaging developers are to enter into this space.
[1] https://www.fool.com/investing/2017/05/06/twilios-uber-bomb-why-investors-shouldnt-panic-yet.aspx
Diving Deeper into the CPaaS Market:
Although Twilio does not break out their revenue distribution between product types in any investor relations materials, the usage-based fee structure of APIs and the relative infancy of pre-built, turnkey solutions makes their APIs the likely dominant portion of revenue share. The other pre-built solutions likely include small segments of somewhat “commoditized” communication solutions like general mass communication or simple voice API solutions, as they do not include Twilio’s flagship cloud platform service.
Generally, API solutions are very difficult to sell to decision makers at enterprise partners. APIs are also a very popular complementary product for CPaaS providers, and they are actually used as a differentiation factor against Twilio in the case of many companies such as RingCentral or Vonage, while also reflecting the overlap with unified communications-as-a-service (UCaaS) services of larger vendors. As a result, developers are constantly rolling out new types of APIs – that by nature require constant updates – with increasing use cases. These include APIs for team-building collaboration control, contact center frameworks, messaging systems for appointments, and two factor-authentication.
For telecom and the cloud services they migrate to, APIs are the future; the companies that can effectively roll out the most use cases will dominate this industry. Current telecom leaders that have worked on this facet include Oracle, Huawei, HP, Ericsson, Aepona, ZTE Soft Tech, and Cisco, several of which are rolling out UCaaS platforms as part of a larger umbrella that includes CPaaS. Yet, Twilio APIs stand out in the fragmented product listing for API use cases, despite being from a smaller but equally pervasive market in cases like SMS development. Anecdotally and through developer mindshare, Twilio is consistently both the product that developers use – tautologically, because other developers use it (a testament to their robust sales and marketing spend), and product-leadership wise, because like Stripe and GitHub, their product stands out in a field of far more rudimentary and poorly documented API services. On Stackoverflow, the home page of Internet developers fielding feedback and Q&A’s, Twilio’s tag results in over 7,151 questions as compared to 238 for Nexmo and 160 for Plivo[1], its closest competitors.
On the other hand, contact centers and pre-built solutions are a seriously fragmented and rapidly growing (22% CAGR to 2022) segment that includes inbound/outbound call centers. Often, big companies run their own operations which consist of many different interconnected components like phone systems, automatic call distributors, recording software, predictive dialers, response software, and others. Legacy players here are giants like Cisco, Genesys, and Avaya, but because these solutions are so antiquated, they are ripe for disruption by an easy to use and modular solution like Twilio’s.
[1] As of May 3, 2020: https://stackoverflow.com/tags
Company – The Economics of Twilio’s Business:
How are customers billed?
Usage-based fees come from things like Programmable Voice/SMS APIs where pricing links to the volume of usage, such as number of minutes of phone calls, number of texts sent/received, number of emails, etc.
“Our Flex contact center platform is generally offered on a per user, per month basis or on a usage basis per agent hour”
Non usage-based fees come from subscription-based solutions. For consistent customers, Twilio often enters yearlong minimum revenue contracts to continue to drive revenue. Variable customer accounts do not do this, however. Regardless, Twilio’s pricing matrices are available to companies so they can see up-front the costs of using any combination of services.
How much does it cost to acquire a customer?
Sales and marketing costs are set to scale down in line with the company’s long-term targets of 20% of revenue, and even currently, they are already some of the lowest in the industry. Since S&M spend slightly outpaces new cohort growth the next year and is in fact a consistent roughly 1.2x multiplier on new cohort revenue, it is likely a significant contributor to Twilio’s high dollar-based net expansion as well.
We see this cost multiplier remaining static overtime, as despite gaining market share, Twilio will not incur higher spend to grow its new client cohorts because of increased enterprise exposure. Additionally, Twilio’s developer mindshare is sure to ease customer acquisition, with a registered developer database at currently 9x (7 million) the size of its closest competition.
Going forward, Twilio’s main customer targets include more enterprises, given their current penetration of only 10% of the Global 2000. Its developer-first sales strategy should help it compete due to lower customer acquisition costs in that area. However, revenue expectations for these enterprise customers depend entirely on the type of account they end up becoming. Active accounts (contract-based) have historically high dollar-based net retention/expansion rate, implying high sales growth and upselling over time. Variable accounts are not broken out as clearly, so while there is little granularity, it’s feasible to assume their spend is fairly consistent given the relative size of example variable accounts like WhatsApp. Because they consist of such a small portion of total revenue anyways, this cements Twilio’s revenue as less fragile to individual customers and their heightened inconsistency.
In generating revenue, Twilio’s costs are distributed unevenly between carrier fees and S&M. Because carrier fees for things like programmable SMS messaging carry over directly to consumers and are treated as COGS, Twilio’s gross margin is lower than similar enterprise-facing SaaS companies. Yet, on the SG&A side, Twilio’s previously mentioned lower customer acquisition costs enable higher and faster efficiencies of scale. Our specific model projects Twilio leveraging high upsell revenue from its existing customers that have industry-low churn in order to calculate lower SG&A and S&M expenses as a multiple of exiting customer accounts, rather than directly as a percentage of revenue.
1] Consistent commentary across the web: https://www.reddit.com/r/webdev/comments/2bjv2h/best_sms_api/
Main Takeaways
- Twilio’s developer focused approach has gained the respect[1] from developers and created a best-in-class product suite, as shown by online commentary in developer communities and by sheer references and questions in leading forums. Product-wise, their specifically created markup language, TwiML, contributes to this as a super-efficient advantage. While no product sells itself, the developer-internal push for Twilio cements the rationale behind their low customer-acquisition costs.
- Revenues are generally reliable, especially with minimum revenue commitments created by active account contracts and the minority contribution from higher risk variable accounts.
- Twilio’s product suite taps into all parts of the CPaaS market, including voice, text, video, and email (after its SendGrid acquisition, enabling it to specifically benefit from the macro trend of CPaaS gradually capturing shift from on-prem solutions.
Twilio continues to rely on many very small (but growing) accounts, creating an uphill battle but large growth opportunity in penetrating larger enterprise partnerships and posing a significant threat to legacy product suites.
Long Twilio.