Description

Bandwidth Inc, BAND, has durable organic growth riding a secular uptrend, operating leverage, and a moat from a uniquely scarce asset - it is a CPaaS orchestration platform provider that is also a CLEC and tier-1 network backbone. Through a long history of events they are deeply ingrained into the global fortune 2000 and provide services that are now moving into a dramatic global upsurge. Their platform’s network effects and scarce assets impose switching costs on users resulting in over 100% net dollar retention.

BAND used to be a very different business and much of its low valuation is based on people still thinking of BAND as a commodity voice or sms provider. Up until a handful of years ago BAND was mostly a service for allotting and managing phone numbers and similar communications, whether those communication channels operated like phones or more unified like Teams/Zoom. Corporate customers of BAND would log into an online service platform and assign phone numbers to employees or users. That was the BAND of yesterday. Today, they have a communications-centered cloud orchestration platform and other companies are writing services on their platform, which is quickly becoming a standard for communication services in much the same way that people use an Amazon Web Services account to buy various database, search, and analytics services. Google Voice and Microsoft Skype both internally use BAND APIs for their telephony features. If a company wants to manage multiple call centers and be able to implement and rapidly iterate on fraud detection, multi-channel communications, and plugging AI into the customer support experience, BAND has the most integrated platform and the only platform for certain features. If you’re starting an AI company for sales, UX, or customer service, your first stop has to be BAND so you can get on their platform where customers can plug their data flows into yours. Every quarter, more companies are joining their platform making it more desirable for additional new companies.

Business evolution

The beginning: CPaaS/UCaaS

For much of its early life, BAND was primarily a provider of telecom management as a service (Communication Platform - CPaaS or Unified Communication - UCaaS). This was the sort of thing where human resources could log in and apply new numbers for new incoming employees. The service was also used by Zoom and Skype in their connections to the telephone systems (e.g. when people dial in). As BAND began to offer more unified communications (calls that roam from IP to telephony, linking together of IP calls to legacy telecom connections) they slowly began to embed more “tech” into their offerings. 

These communication services had $360mm in 2022 revenue and are growing a low single digit percents faster than GDP.

It was this business that led BAND to acquire their own tier-1 network backbone during the great financial crash when they could do so for pennies on the dollar.

Competition has declined significantly in recent years with the purchase of Voxbone giving BAND their pick of enterprise clients. Most competitors in the next category try to have an offering in this category just because BAND has shown so many buyers want the two linked.

Up the stack - Programmable services (e.g. messaging)

BAND grew with their customers and as they began to include more “tech”, their API offerings expanded. One of their main offerings in this category is API based text messaging. This text messaging is similar in some regards to what Twilio offers (indeed, Twilio was a large BAND customer until they decided they didn’t want to be buying services from a competitor and moved to another provider). Sinch is another big name that is mainly a competitor at the programmable services level. In talking to a lot of people about BAND many seem to only know BAND as “a Twilio competitor”, which is a poor understanding. As their own CLEC and tier-1, BAND has *much* lower costs and focuses on providing these services to large companies (including Twilio for years). BAND is the owner of the assets who monetizes them, Twilio is buying from someone like BAND and trying to make money with application services on top. Of course, BAND has introduced their own application services for the 20% of traffic that makes up 80% of the market, and is happy letting folks like Twilio spend resources on the rest.

These programmable services had $91mm in 2022 revenue and are growing quickly, the company gives a projection of 21% CAGR within the projection window -2026.

Whole stack - Cloud service provider - direct enterprise solutions

BAND has now reached the top of the product “stack”, providing a cloud platform for orchestrating and implementing everything in telecom. A company can sign on to BAND cloud and order communication services, both IP and classic telecom, run incoming calls through Pindrop for security and fraud protection, run calls through Genesys for customer support, plug in AI, run secondary paths for certain calls, and include backup failover destinations - all without leaving the BAND orchestration view. This is a new kind of cloud service. Some of the cloud providers have pieces of this but not the whole thing, and the pieces they do offer are often provided by BAND behind the scenes! 

BAND has a first-mover advantage here and it’s going to take a while for anyone to compete even if they could. Someone like Verizon or ATT has the network layers, but would have to conduct a moon-shot level effort to build a separately operating business and build the technical foundations - taking years and facing entrenched network effects of services already at BAND.

Direct enterprise solutions had $21mm in revenue in 2022 and are growing fairly quickly, the company projections 14%  CAGR through 2026 projection window, although recent data points higher. This area is in a strong secular uptrend as more and more companies want to add AI, fraud protection, and unified management to their call centers and phone infrastructure.

The big picture

BAND is growing out of its history as a commodity provider of UCaaS and CPaaS and into a role farther up the value stack, including a new role as the provider of the value stack as a cloud provider where the other participants bring their products as “apps” on the BAND cloud at much higher gross margins. As more customers join the BAND cloud more service providers also join, providing classic two-sided market advantages and network effects. As the programmable services and direct enterprise solutions grow so does gross margin as those two services have much better margins than the legacy business.

Competitive position

BAND provides about 10% of the phone numbers in North America, spread across all its services. Bandwidth has long touted being the only cloud communications service provider that was also a CLEC (Competitive Local Exchange Carrier)*. CLEC status gives access to unbundled network elements (UNE) through the various ILECs (Incumbent local exchange carriers).  BAND talks a lot about this, but just being a CLEC isn’t automatically going to make a huge difference - if it did then a competitor could buy one of the other 15,000 CLECs in the US. The more important aspect is that being a CLEC lets you control infrastructure that you connect to the communications network. For BAND this means their owned tier-1 network, transfer exchanges, and the specific data centers they have built out and interconnected using their CLEC status. BAND leases its data center space in colocation-friendly datacenters that match its tier-1 peering points, mostly from Digital Realty and Equinix, so that customers and providers can colocate with them. Major sites include Digital realty Atlanta, Dallas, Los Angeles, and New York. Equinix sites in Frankfurt and possibly London (not completely sure if London is still active). A Brussels Interxion site came along with the Voxbone acquisition as did Hong Kong (with host MEGA-iAdvantage), and additional space at CoreSite LA and Telx New York. Together with the global telecom assets they’ve built up over the last decade, it would take quite an effort to replicate the services and network BAND offers. Many of the CLEC tier-1 type assets aren’t even for sale any more and there is a lot of regulation to navigate. It would almost certainly be cheaper for Microsoft or Google to buy out BAND at three times current value than to try to assemble the assets on their own.

The biggest competition is really companies leaving the phone network completely. If you dial in from a phone BAND has an advantage with tier-1 presence and direct integration that can power fraud detection and redundancy. If the customer comes in via chat (including voice or video) through an app then the customer could avoid using BAND. To us it seems like people will be calling phone numbers for at least a while longer and the services that are used on chat are often starting in the telephone call servicing world so BAND winds up with many customers who are entirely IP call based. Additionally, the flows being set up by customers and vendors in the BAND communications cloud include so many desirable services that one could imagine even non-voice chats being routed through them in the future.

Finances and Economics of the business

BAND is currently being run for growth, with enough control asserted to keep expenses pretty close to revenues and operating cash flow minus CAPEX near zero. It can be difficult to tell if the underlying business is actually profitable when OPEX and CAPEX are both being run for growth. While the cash flow from operating activities looks fairly good, all of that and more is currently going into CAPEX.

Thus a critical question is whether the business is profitable investing in growth or not. BAND has had significant upswings in both expenses and revenue, especially after 2017. Growth can hide a lot of cost sins and it can be hard to estimate how much of the cost increase is due to the growth. We have data since 2015, thus rates of increase since 2016.

One view of business growth operating efficiency I like is the revenue increase versus increase in SG&A expenses:

Figure: YoY increases in SG&A and resulting revenue increase. Note that only 5 data points are used here. In 2016 and 2022 there was growth of US$14.3mm and US$82.2mm respectively accompanied by *decrease* in SG&A. That decrease shows that the company can exert discipline and remove SG&A that may not have been efficiently contributing to growth and can do so while still growing. Combine that with the slope here significantly greater than 1, here $1 of SG&A growth corresponds to $3.18 in revenue growth, and we have good business economics where the average gross margins of the business are more than good enough to generate long term operating leverage.

BAND also has the cost of filling out data centers (and capitalized software R&D). We can thus also look at OPEX increase combined with CAPEX compared to revenue increase. This is a harder number to build a heuristic around but a long term number I like to think about is a slope of 1 - this is the point where each added dollar of OPEX increase + CAPEX brings in at least one dollar of revenue. A similar fit for this metric for BAND is at 1.05. Assuming that the CAPEX can be used for a while and the numbers above show good SG&A discipline (which they do), I would consider anything around 1.0 or greater as being within the “reasonable” range for growth. There are a ton of qualitative discussions to be had around this number, for example how to account for growth due to acquisitions or how to evaluate the lifespan of the CAPEX. Overall, however, many companies don’t look very good in such tests but this management team appears to be watching the growth horizon and efficiently putting in new customers on high gross margin lines of business with high net dollar retention.. At any time they could ease up on the CAPEX and SG&A (as they did in 2022) and drive profitability gains. As a shareholder, I want them to keep optimizing for growth until their revenue is mainly in the higher margin businesses. In 3-4 years I would expect a multiple rerating as the perception of the main business changes, the debt is retired, and we’ll all be holding shares in a high margin SaaS company with excellent financials.

We can also look at the business historically and see the strong earnings and ROIC up until they went all in on spending for growth. Since they are adding growth in areas with even better margins and very high net dollar retention this is a business where cash will flow freely if and when they take their foot off the accelerator.

Valuation

Debt

The first thing that jumps out of the balance sheet is their debt. Most of it is convertible notes that convert at crazy high prices and pay very little interest. This debt is a neat accomplishment of management and should only be retired by the company when they can do so at a substantial discount to face value, and they have been doing exactly that as the opportunity presents itself. For example, after issuing some convertible debt for an acquisition in Feb 2020 (literally days before the pandemic) management went back to the well and issued $250mm in cheap debt in March 2021, then spent some time making deals and retired $160mm face value of the 2020 debt using $117mm from the 2021 issuance (closing in 2022). Moving forward I will be watching to see management using spare cash to retire debt opportunistically. One consideration is that they will likely not be able to refinance the debt on such beneficial terms.

(Their revolver is at SOFR plus 1.5-2% (or base rate plus 0.5-1% at companies choice), which is also quite nice.)

Rerating on multiples

On the qualitative side there is a major transition in category going on. BAND has historically been a lower gross margin commodity services provider (the first business in the business discussion above). However, the two higher margin businesses are growing significantly faster than this historical business and should emerge as the main business within a few years. Rerating of valuations based on this category shift offers a significant upside. Interestingly, one can see a couple years ago when the company was being valued on the “sizzle” of such services the stock price was more than 10x what it is today. Ironically, the valuation has dropped since then while the actual plan has materialized and the business change has gone from theory to demonstrated. 

DCF

There are lots of decisions to make when estimating a discounted cash flow value and I typically use them to educate myself on what the market is already pricing in and what the value might be in different scenarios. For the DCFs below I model the loan repayment deadlines as negative cash flow on the due dates - as long as they refinance at a rate lower than my discount rate of 9% they will do better than this cautious treatment below. (Also if they convert the debt to stock the valuations go up despite the added shares.) BAND’s cash flow from operations averaged 5.5% of revenue during 2015-2018 before they were heavily investing in growth, and averages 4.5% over their whole public filing history. While they have operating leverage and improving margins, for the first scenario the cash flow is based on the lower of those two numbers.

This assumes management guidance is approximately correct (management has historically guided well below what they deliver), and margins do not improve from their long term average including the heavy hiring period for growth. This does not model investment expenses in datacenters and the like, but those were negligible before the growth period and our analysis above suggests dollars reinvested for growth produce more growth than they cost. Still, the conservative margins and debt treatment result in all the value effectively being net present value of the terminal value so it’s worth noting we are assuming a 6% terminal growth phase after 2030 while keeping a 9% discount rate - whereas this industry is probably in for secular growth well into the 2030s. Even with all these conservative steps we get a cash share value about 100% up from today. In practice we would expect the market to value the final state of this company much higher in Price/Sales or other financial metrics than it does today, which would raise this price significantly.

In our second scenario we assume management growth guidance again but no cash flow at all until 2026 when they start yielding 10% (which would still not be above their historical max, and aligns with their increasing margins. Many of the previous observations apply. The present share value in this case is about $59 (vs trading around $13).

Finally, we can adjust the numbers to get a value closer to the recent trading range to see what is priced in. We wind up requiring growth well below secular addressed market growth, thus requiring significant loss of market share (vs. the increasing market share they have seen every year).

Ways to win

BAND has demonstrated that it can be highly profitable before it went into hyper-growth mode. At any time they could ease up on the investment and produce a significant cash flow yield (high single digit to low double digit). Of course, I don’t expect management to do this while they still have the combination of a strongly secularly growing addressable market and the sort of technology head start that venture capitalists can only dream of…

Execution success, even moderate execution success, will rapidly result in fundamental momentum: rising revenue, margins, and a changing narrative from commodity CPaaS provider to platform hoster in the spirit of Amazon Web Services but with no immediate competition. Considering we are currently buying the company for a total enterprise value less than 2x sales, even slight changes to the narrative have the potential to dramatically change the valuation.

Acquisition is another real possibility. It looks like the tier-1 network plus communications platform BAND has would be very difficult to put together from scratch today and would definitely cost more than the enterprise value of the company. That makes BAND an extremely attractive acquisition for larger companies that might want to strategically take over their platform position. Indeed, one concern I have is that an acquisition that happens too soon might not allow the company to reach the full gains that it should.

In the short term, management has guided for a mediocre 2023, largely because they don’t expect a repeat of enormous political SMS spending in 2022. However, this is also a management team that makes it a practice to constantly beat expectations, and so far have already been doing so in 2023. In any case, I would expect more of a breakout in 2024 than 2023.

*  Strictly speaking I found public records showing a few other CPaaS providers who hold CLEC status in at least a few states, but they don’t have the network, scale, or offerings of BAND. Most of the ones I found only have a CLEC status in one or two states and none had significant fiber/network assets.

Catalyst

The current EV/Sales is less than 1.5. At some point the market will note their growth as a central market and orchestration platform of advanced telecom services and the valuation will move to reflect their increasing margins, operating leverage, and greater than 100% net dollar retention.