Cogent Communications Holdings, Inc. (CCOI) Earnings Call Transcript & Summary

Okay. Good afternoon, everyone. Again, it's John Hodulik, the telecom, media and communications infrastructure analyst here at UBS. I am happy to announce our next speaker is Dave Schaeffer, the Chief Executive Officer at Cogent Communications. Dave, thanks for being with us.

Hey. Thank you, John, for hosting us. Thanks, UBS. Thanks to the investors for your time. And please, everyone, except for our apologies for starting a few minutes late. I think I was maybe a little more technically challenged than I should have been.

Not at all, Dave. We're getting a sort of first-hand sense of just how difficult this networking business that you have really is from the user end? But we appreciate everybody bearing with us and -- sorry -- yes, sorry, for the late start.

So Dave, let's jump right in. Obviously, 2020 has been a surprising year to say the least. Maybe we'll just start off by discussing how Cogent has been impacted by the COVID pandemic and what your priorities are as we look out into 2021?

Thanks, John. So first of all, our thoughts are out for everyone who has been personally impacted by the pandemic and hope everyone is safe. Our view is that this has lasted much longer than we had originally anticipated. And I guess, our current thinking is, we're about halfway through the pandemic, meaning we probably have about another 9 months to go. GDP was clearly impacted throughout the developed world. We have seen massive government stimulus to partially offset that, but a number of businesses have been profoundly and may be permanently impacted. We are fortunate in that we sell a necessary utility Internet access and have been much less impacted than others. In a world where GDP growth is 10% or 20% negative, we have been able to achieve top line growth of about 5%. That is substantially below Cogent's long-term average, but still better than most companies and better than other wireline telecom companies. Disaggregating our business and looking within the constituent demand sectors, in our NetCentric business, which is 30% of our total revenues, we've actually seen an acceleration in growth with growth returning to long-term averages of about 9% year-over-year. We've seen an inflection in the rate of traffic growth. And Cogent continues to grow about twice as fast as the Internet in general, due to our aggressive pricing strategy. And that gain of market share has allowed us to expand margins with 90% of our NetCentric business being on net, carrying a 100% gross margin and $0.95 of EBITDA contribution. Our corporate business is a more mixed beneficiary of the pandemic. At primary locations in the central business districts of major cities, we've seen a increase in customers taking 1 gigabit connections in order to support remote work-from-home employees. The need for non-blocked and symmetric bandwidth and a larger pipe has helped that portion of our business, even though many of those offices do not have employees in their offices at this point. Where our corporate business has underperformed is at secondary locations, branch offices. Many of those locations are off-net, some are on. We typically sell a dedicated Internet access service at the primary location and from that sale, we often can sell secondary locations, and we sell them a combination of either a dedicated Internet access or VPN services delivered over SD-WAN or VPLS. The pandemic has made many companies shutter those secondary locations not needing VPNs and not needing Internet access at those locations. As a result, our total corporate business, which has historically, organically, grown at 11.5% for the past 15 years, saw its growth rate decelerate to low single digits, 2% to 3%. And I think that's probably going to continue throughout the pandemic and then as businesses return to their offices, we will see a reacceleration in that corporate business.

Got it. Great overview there. Maybe, Dave, we can take both of those segments, maybe starting with the corporate segment that you finished up on. So what I heard was continued strength in sort of major businesses in sort of the HQ, but sort of a fall-off on the branch side. Could you give us a sense of what the mix is of your -- of sort of your current either connections or maybe more importantly, sort of revenue streams that come from sort of those 2 portions of the corporate segment?

Yes. So first of all, we characterize every dollar of revenue 4 ways: On-net versus off-net, corporate versus NetCentric. We also characterize it U.S. versus rest of the world. And then finally, product Internet access co-location and VPN services. Within our corporate segment, which is 70% of revenues, 80% of the connections are on-net, 20% are off. And by revenue, it is 60% on-net, 40% off-net. Within that on-net portion, the majority of it comes from those primary locations in the Central business district. Those companies then can have remote offices that fall in 1 of 2 subcategories. They can either be a secondary office within the same MSA or metropolitan service area or they can be in a remote setting. If they are in a remote city, they are more likely also to be on-net as companies typically gravitate to the same type of real estate across their footprint. The portion of the business that's been most negatively impacted have been those secondary offices in the same MSA, where we're not selling off-net connections in the suburban campus environments. An example would be a few out of business in New York, just as UBS has its main operations on Saks Avenue, you have a major facility out in Stamford, Connecticut. It's likely that you would shutter that secondary location in Stamford, you would keep the primary location. And as people return to work post-pandemic, they will use a combination of work from home, coupled with a mandatory requirement to go to the primary location on certain days. If, however, you were connecting to the UBS office in Houston or in Los Angeles or in San Francisco, those offices will probably stay as they are and I have to admit I don't know where UBS offices are in those markets, but I would suspect they're in big, tall, shiny fancy buildings, just like your office on Avenue of the Americas.

That's right. I think that's right. So just -- is it right to assume, given how you laid it out that where you're seeing the pressure in these branch offices with a lot of off-net connections is a sort of lower-margin business, so even looking at the exposure to this trend that you can sort of peel the onion back first in terms of sort of central business districts and HQs versus branches. And then on-net versus off-net, it would seem that the impact to overall profitability would be -- it's easier to sort of whittle it down, does that make sense?

Yes, it sure does, John. So first of all, to remind investors, our long-term guidance has been a 10% top line growth and 200 basis points per year of margin expansion. Cogent has been a public company for 15.5 years. And we've, in fact, delivered those numbers. We've grown 10.3% in total, and averaged 210 basis points of margin expansion. Last quarter, with growth at 50% of our growth target we still were able to deliver 150 basis points of margin expansion. And the reason for that outsized expansion and margin is twofold: One, even prior to pandemic, we had been slowing the expansion of our footprint, which allowed our contribution margins to go from 44% and up to 62%. That has continued. The second thing that has helped is that a much greater percentage of new sales are on-net with the NetCentric business being 90% on-net, 10% off; as opposed to the corporate business, which is 60% on, 40% off. So with the slowdown in corporate and the acceleration in the NetCentric business, we've actually been able to deliver contribution margins in the mid-60s even off of a lower growth rate.

Got you. So how -- I mean, you started off by saying you think we're sort of halfway through the impact of the pandemic or maybe just the course of the pandemic. Does that apply to sort of your impact -- the impact that the corporate business has seen or you expect it to see? I mean, my thinking here is, you would imagine a -- the work-from-home environment has already kicked in, obviously, really starting in March for every major company, and they would have gotten through that right away. And now we may be going into a new set of lockdowns and work from home whereas some parts of the country, maybe internationally as well, have gone back to work and now are going back home again. Are you feeling any incremental pressure from this second wave or has -- is it been a sort of a steady sort of glide path of impact given the trends you were just talking about?

So first of all, my observations on the pandemic are as much as just a citizen as they are as the CEO. And just looking at the practicality of getting a vaccine distributed in sufficient quantity to make people comfortable to return to social settings where they will be able to function without mask and without special distancing and restrictions on occupancy. I think we're still 9 months away from that. Even with the good news on several vaccines being developed, just the logistical challenges are so great. In terms of our business, again, let me just aggregate it. On the corporate side, I think we have troughed in terms of the worst impact. I think the paralysis that initially overtook corporate decision-making has now morphed into how do I best deal with a remote workforce and a prolonged period of work from home. That has actually been a positive catalyst to our sales and multi-tenant office buildings and central business districts. It's the reason why people who have not bought from Cogent are now buying and it's a reason why our corporate customers who had 100 meg connections are upgrading. In fact, over 80% of our new corporate on-net sales are 1 gigabit sales replacing the 100 MEG product, which had been our standard product for 19 years. That is a positive. I think what will continue to be a drag is sales of either VPNs or Internet in those secondary offices. If those offices are going to remain open, I think we'll continue to be able to win business and sell on that. If those offices are going to be shuttered, that will end up being a headwind. Now in addition to that, we took advantage of what we viewed as a slower selling period in the depth of the summer, July and August, and implemented a new CRM system. That also had some drag on sales force efficacy. And then finally, and we've commented on this extensively, we pivoted quickly to a work-from-home environment for our sales force. Cogent has 600 quota-bearing reps. They normally would physically sit in 38 offices around the world, and each rep was expected to do about 100 outbound calls a day. We quickly, in mid-March, pivoted exclusively to work from home. We equipped our sales force with laptops. We put cellphones in place, monitoring systems to be able to measure the efficacy of CRM activities of e-mail, and phone activities. All of that worked out well. The second thing we had to do was establish a onboarding system for new hires. Our model has historically experienced about 5.3% per month sales force turnover. Anyone who has ever worked in an outbound telesales organization knows how difficult that job is and how prone it is to turnover. Well, we needed to come up with a mechanism to hire, to onboard, to train and then ramp those people without ever getting them into an office. I think we did a pretty good job of doing that in the late spring and early summer, and we continue to hit our hiring goals. Where I think we underperformed is in terminating the less productive reps as quickly as we needed to. We saw our sales force productivity decline and our cost per dollar of revenue acquisition costs go up. We, in the fall, started to focus on that management of underperformance out of the business, and we saw an uptick in turnover in August and September continuing into October and reverting back to kind of more normalized rates of sales force turnover. As a result of this, even with the pandemic, we expect to see an improvement in sales force productivity in the fourth quarter and probably continuing through the first quarter, even though the pandemic will still be raging.

Got you. That's a good update. So I guess, does that also mean that you think that the issues at the new CRM system are behind you? You -- the disruption that you saw from that end and the sales force is some -- is in the rearview mirror or even within the better sales productivity that you just called out, are there still some lingering effects in the fourth quarter from the new systems?

Yes, I think we're well adapted to our new CRM. And just to remind investors, we are not a software company. We believe in commercially off-the-shelf software. We ended up deploying initially Siebel. We kept that for 13.5 years. We then migrated to Salesforce when Oracle acquired Siebel and ceased enhancements on that product. We were on sales force for nearly 6 years and then concluded that it was not effective in integrating into other systems that we had. It was requiring our sales force to toggle back and forth between multiple systems too frequently. So as a result, we developed a CRM system that allows our reps to remain in 1 screen all day long, and that improvement has resulted in our ability to be much more effective. And I think those benefits are in-place now. So I think we will continue to see improvements throughout this quarter and throughout next year, even without the pandemic being resolved.

Got it. One thing before we move over to the NetCentric side of things, I guess it's really 2 things. First of all, can you talk a little bit more about that 100 meg to 1 gig sort of transition and is this the start of a -- I guess, it had started before, but you've seen -- it sounds like you've seen an acceleration of that -- those conversions. How far are we through that? And can you talk about the sort of ARPU/margin differential between the 1 gig product and the 100 meg product? That's number one. And then two, David, this is one of the things I wanted to ask you, but just because I know that you're sort of an expert on the sort of corporate real estate market, I mean, do you think that what we're seeing in the pandemic is -- will have long-lasting effects on sort of multi-tenant office buildings. Is that -- or do you think we get through this and then everything goes back to where it was?

So 2 very different questions. Yes, let me take the real estate question first. I think we will see a return to the central business district, but it won't be exactly as it was before. Cogent has lived through several economic downturns, the dot-com crash, the great financial crash. And each time, it was predicted that real estate would permanently change. I think what happens is rents compress, equity in many real estate projects get flushed out and tenants tend to migrate from B and C buildings to A buildings at lower rents. The central business district of major cities remains desirable to businesses. I also think the shared office business model, WeWork, Regus was already in trouble, pre-pandemic. This further accelerated the decline of those shared business models. Finally, there had been a migration away from doored offices to open floor plans and cubes. Banking being maybe the most extreme example. Banks used to occupy about 325 square foot per employee. We have seen that decline to about 240 square foot per employee over the past decade through cost-cutting and the migration to cubes. I think the pandemic does 2 things: I think it returns people to doored offices; and two, some employees will work some days remotely. So I do think there will be a permanent shuttering of some of the suburban offices that were meant to limit the amount of commuting, and that will be done through telecommuting and work from home. Now to the question around port size and the impact on Cogent. I'm going to show both of our ages, John. We've known each other a long time. And the -- when Cogent cut started, people laughed at us for offering a 100 meg non-blocked and non-oversubscribed internet service to our corporate customers saying they would never use it. And the reason we chose 100 meg connection is that most local area networks inside of the office were 100 meg, the interface to the wide area network was capped at a 100 meg, and that became the highest speed, which we could connect. Over the past 5 years, there has been a significant increase of customer premise equipment that can support 1 gig WAN interfaces for the 2 quarters prior to the pandemic. So in Q1 of this year and Q4 of 2019, we actually were selling more gigabit than 100 meg because the customers could accept it, and it was a relatively modest upcharge at $200 premium. With the pandemic and the need for more bandwidth to accommodate those work-from-home employees, you have to remember, there are 2 very different architectures for a work-from-home employee. For a very large company like UBS, you are using your broadband connection to build an ad hoc VPN and you're connecting through a data center to the UBS network. That is not normal. For most businesses, their firewall at their primary location act as that VPN concentrator, that's a very different model. And because of that, those companies need more bandwidth, and it needs to be symmetric. So we actually have seen an uptick in our gigabit sales where today, almost 80% of our new sales or 1 gigabit, and we think that will continue post pandemic because the customers can accept it. And in a perverse and somewhat ironic back pattern, it actually costs Cogent less to deliver a gigabit to a customer than a 100 meg. We actually bring a gigabit to them, and we're stepping it down just to meet the needs of their customer premise switch. And now we no longer need that stepped in.

Okay. Interesting. Well, that's great. I think that's a great background, great update on the corporate side of things. Maybe moving over to the NetCentric. Previously, you cited that increase in traffic growth in September and October -- in October. Can you talk about the sources? What's driving that acceleration? Has it continued to date? And do you think -- again, it gets back to what's COVID-related? And what sort of long term? How do you expect traffic growth coming out of the pandemic to evolve?

So the primary driver of traffic growth on the public Internet has been and continues to be streaming very well. We have seen 3 things happen: One, just the total minutes of video consumed have gone up with people sheltered in place without having other entertainment sources; two, the percentage that is streamed versus linearly delivered increase; and third, we have seen a proliferation of streaming services to initially all streaming was kind of Netflix. And now there are many other alternate services with alternate business models, whether it's Peacock or Disney, HBO Max, Hulu and a bunch of smaller services. We're also seeing a globalization of streaming, like many technology trends, they begin in America and then expand around the world. Cogent derives half of its traffic outside of the U.S. and our traffic is relatively balanced with roughly 55% of our traffic coming from streamers of content and 45% going to the other side of that transaction, the 7,200 access networks that we connect to. And over 70% of the instances Cogent gets paid on both sides. That is a benefit to us and what we are seeing with streaming are 2 things: the streaming window has widened from a traditional window of 7 to 10 p.m. to now expanding to a peak window from 3 in the afternoon till midnight; and two, streaming has gone from early adopters to being mainstream. Those trends spiked up the rate of Internet traffic rate. Even though the base of the Internet continues to grow, traffic growth continues to compound at about 25% per year. We have seen an acceleration in that growth rate as a result of this transition from linear to streaming. Today, only about 20% to 25% of all content is streamed with the vast majority still being delivered linearly. That trend was expected to take 10 years before the dominant form of video consumption was streaming. What we did is pull 2 or 3 years transition -- forward into 2 or 3 months. And I don't think we go back, I think we just continue to grow off of a larger base. And with the broadening of the base, we also have the benefit of getting better effective pricing per megabit. We offer discounts based on volume and contract term and with more streaming services that has helped our effective pricing.

Got you. Can you talk about sort of competition on that side of the house? Who are your main competitors? Has there been any real change in that list or in their sort of how aggressive they're being from a pricing standpoint? And then maybe given the increase in traffic, any comments that you can give on sort of utilization of your network today? I mean, our -- and this gets to sort of CapEx and capital intensity. As -- are you able to -- I know you -- obviously, you keep a lot of overhead in the network. Is there an issue where you might have to spend more to keep up with the growth that you're seeing?

Got it. We continue to grow about twice as fast as the Internet gaining share. We guarantee to underprice our competition by 50%, and that's allowed us to become the most connected network in the world and today, the second largest carrier of Internet traffic. Our entire business is based on the Internet. The companies we compete with, the Internet is a small subset of their total business. So on the NetCentric side, our competitive landscape is bifurcated into 2 groups. There are 4 large players that compete for the bigger opportunity. Lumen, ourselves, Telia and NTT. We're the second biggest of those in gaining share. The other 3 are losing share. There's also a tier of smaller regional players, you'll see companies like AT&T and Verizon, Telefónica, Deutsche Telekom, Telecom Italia, PCCW in that second-tier they tend to focus on smaller accounts that can only provide competitive services regionally. We compete there as well and gain share. I think we are continuing to focus on a product segment that other parts of the industry don't view as profitable. If you look at all of those competitors, Internet transit is an afterthought, they disincent their sales force from selling it, and we continue to gain share and focus on it. Now we can do that profitably with lower capital intensity. And now to the second part of your question. Our network architecture was more efficient day 1 than others. We've built a network-based on IRUs, running IP directly over DWDM, protecting a layer 3 using Ethernet, effectively a corporate land on a global scale. That is a very different architecture than any of those other companies use. They are primarily phone companies that allocate bandwidth to the Internet, and their networks are apparently less efficient. But the most important factor has been the improvement in technology over the past 20 years. We cheated at the beginning by buying a bunch of distressed assets, but those assets are long since depreciate. We have benefited because we capture more of that technology advancement as economic profit than any of our competitors. The 2 incremental technologies are wave division multiplexing, which is on an 80% compounded price performance improvement and optically interfaced routing, which is on a 40% price performance improvement. Our architecture allows us to deploy that incremental technology much more efficiently than our competitors. As a result, we're running at about 32% utilization in our network. We generally add about 50% to our network capacity annually, and we do that within our $35 million maintenance CapEx budget. We do spend about $55 million, but that incremental $20 million is for footprint expansion, new data centers and multi-tenant office buildings attached to the network and fiber into new markets. But on the existing footprint, as technology wears out, we moved the less efficient technology to the more peripheral regions, and then deploy the next-generation in the densest parts of our network. Our capital intensity, both in absolute terms and as a percentage of revenue have been consistently declining and should consistently decline. We have been fortunate that we produce more free cash than we can consume. And that free cash is used to grow our sales force, deploy in our network and grow our footprint, but increasingly, we're using that cash to return capital to shareholders. It's why we have returned $900 million of capital, and we have returned it through tax-efficient means with roughly $675 million being used as a dividend, and roughly half of that dividend has been classified as a return of capital. And then the remaining has been used as stock buybacks.

David, that was actually, my last question. Given all the cash you guys are generating -- I wanted to hit on the cash return. So you started to stop in the right direction here to finish up. Obviously, generating a lot of cash, dividend, you've slightly accelerated the dividend increase this year sort of on an absolute dollar basis. I mean, how should we think of these cash returns to shareholders in the future? The stock has been under some pressure, given the issues with the pandemic. Also how do you choose when to be active with the buyback and sort of toggle between that and bigger increases on the dividend?

So let's start with the fact that we have 34 consecutive quarters of sequentially growing our dividend. Only a handful of public companies have that track record. Two, we have a balance sheet that is substantially under-levered and we have $400 million of cash on our balance sheet for more than we need to run the business. Third, we have a very durable revenue stream. Even with a financial crisis as profound as the pandemic, we are growing and expanding margins, growing our cash flow still in the mid-teens. We expect to be able to continue to do that. We are going to continue to be tax-efficient in the return. We are committed to the dividend. We are also doing buybacks. We bought back stock in the third quarter. We publicly commented on buybacks in the month of October. We will opportunistically use excess cash to accelerate those returns of capital. And we're trying to balance an efficient balance sheet with kind of having the optimal cash return policy and what we found is if we do our returns too quickly, they actually don't create lasting value. So it's kind of steady wins the race, we're gradually increasing our returns, and that's part of the reason why we decided to accelerate the rate of growth in the dividend going from a sequential rate of increase of $0.02 a share to $0.025 a share earlier this year.

Got it. Well, that's great. Dave, I think that was a great overview. Thanks for your time. And again, thanks for bearing with us on the technical side today. We really appreciate your time.

All right. Thanks, John. Thanks to all the investors for joining, and we'll all talk soon. Take care. Bye-bye.

Okay. Take care. Bye-bye.