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

Okay. Great. Thank you, everyone, for joining me. This is Sami Badri with Crédit Suisse, covering the communications infrastructure space. And right now in this session, we have Dave Schaeffer, the CEO of Cogent; and Sean Wallace, the CFO of Cogent Communications. Thank you both for joining us today.

Sami, thank you for hosting us, and we would like to thank Crédit Suisse for a great venue and thank the investors for their time and interest in the company.

All right. Absolutely. So there's a lot of interesting things going on with Cogent. We're going to cover a huge variety of them. If you have any questions -- if there are any audience members have any questions, please feel free to click the Ask a Question button, and then that would just prompt an e-mail to me. And I'll filter through the questions as to the conversation. So Dave, I think first big question for you is we are around midyear of 2021. Could you just provide an outlook for us on the corporate dynamics that you are seeing, and in particular, churn decline month-on-month, if any kind of commentary or anything topical from that perspective you could share with us? And then we'll roll from there -- some of the questions.

Yes. Sure. Thanks, Sami. So let's start with the level of activity that our corporate sales organization is seeing from customers. We're seeing, in large portions of the country, businesses return to offices. In other portions of the country, we're seeing companies start to plan for that return to office. So I think across the U.S., we're seeing a higher level of sales engagement, planning for new larger Internet access connections to support work-from-home and also beginning to see companies start to deploy new VPN architectures, many of those replacing MPLS networks that have been basically on hold for the past 15 months through the pandemic. We also saw an elevated level of churn in 2020, peaking in November of 2020, with many of our smaller customers probably having the most economic pain from the pandemic. And then subsequent to that, each sequential month, we've seen a reduction in the rate of corporate churn. So all in all, we're seeing a much more healthy environment. While we're not yet back at our historical growth rates in corporate, we have definitely hit bottom, and we're starting to see improvement across the entire corporate sector.

Okay. Great. And then one kind of pivot I did want to take is, as you see corporate come back, have you seen any further indication of a greater share of flexible working models? And as you see that play out, has that actually translated into 1 gig demand coming in? Or how would you describe that process or that trend evolve and how it impacts the business?

So I would say based on the sample size that we have in our engagements with customers, it's anecdotal information. I think virtually all companies that we are talking to plan to return to the office, and that spectrum ranges from 100%, 5 days a week in the office to those that will allow flexible workspaces and flexible work schedules for their employees. I would say that the majority -- more than 50% of companies, at least initially, planned some kind of flexible period. Whether or not this is a transition or a permanent work scheme is yet to be seen. I think also as companies realize that they will have some portion of their workforce remote, the need for symmetric larger connections, I think, becomes even more important, and it's why the vast majority of our corporate sales are today 1 gigabit connections. And secondly, virtually all of our new discussions with potential customers is to skip going to a 100 meg connection but directly to a 1 gigabit connection.

Got it. And then when we think about how that all kind of translates into pricing, what should we expect the effects of all of this on -- from a pricing perspective, at least on the average -- at least the ARPU that you are seeing in the business?

So within our corporate customer base, we have both on-net and off-net corporate customers. We're seeing really 3 different factors at work. The first is an increase in ARPUs due to the migration to larger connections. Two, for our off-net circuits, our input costs have been falling at about 7% to 8% annual rate, and we're seeing that trend continue as we're able to create competitive tension between our various off-net group providers. And then finally, with the passage of time, for a like-to-like connection, there is generally a 2% to 3% price erosion. With those 3 factors, we've actually seen corporate ARPUs remain basically flat. For corporate customers, they buy a fixed connection. And then they can use as much or as little of that bandwidth 24/7 is completely reserved for them. That is very different than our NetCentric base, which does pay on a metered basis.

Got it. Got it. I did get a question in from the audience. The question is asking on as you see things open back up and your expectation is that a lot of corporate offices will open back up and people returning back to the workplace, have you had to change your go-to-market at all to capture some of that -- this return to the workplace?

So our goal in reaching out to potential and existing corporate customers is to have our sales force have one touch point per month per either our existing customer or prospect, probably through a phone call, if not at least through a bidirectional e-mail exchange. We still do not have enough Cogent sales reps to reach out to our on-net footprint with that level of curiosity. But our go-to-market strategy has always been dependent on our direct sales force with over 99% of our revenue coming from our direct sales efforts and only 1% coming from the channel. So I think the strategy that we've used for the past 18 years since selling services has remained unchanged. The only thing that has changed is we continue to grow the sales force to make sure we have enough reps to meet those outreach goals.

Got it. Got it. I wanted to shift a little bit to the NetCentric business, which I think is probably going through a bit of an evolution of its own. So after we've seen the significant growth that has played out, especially in 1Q of 2021, do you see demand trends continuing to be solid through 2Q? Or is there something in the trend or in the growth rate that we should take into account as we look at the growth that you've been able to achieve so far in NetCentric, especially in that acceleration?

Yes. So the acceleration in NetCentric growth that we saw in the fourth -- in the first quarter of 2021 year-over-year was the second best growth that segment of our business has ever delivered. That is a bit of an aberration. It was driven by the globalization of streaming coupled with the increased volumes of streaming due to the pandemic and then the broadening out of our customer base. All these factors contributed to this outsized performance. As people go outside more in summer months, they tend to consume less bandwidth, less minutes per day on streaming. That doesn't mean streaming is not going to grow. It's just going to grow slower during the summer. As we think about our NetCentric business, it has delivered an average growth rate of 9% year-over-year. That's substantially below the 24.5% that we delivered in Q1. And I think we should see a reversion to that more normalized growth rate over the next several quarters. The acceleration of the transition from linear video to streaming, I believe, will continue. We entered the pandemic with about 18% of video consumption in the developed world coming through streaming and 82% being linear. Today, about 40% of minutes spent consuming video is coming from streaming, 60% is still linear. So we still have a long way to go in this transition, coupled with the fact that the streaming phenomenon is becoming much more of a global trend than it was 15 months ago when streaming was predominantly a U.S. business model.

Actually, a really good segue there is, there is something functionally changing from where the streaming is coming from with North America already having streaming as a big factor or a big network dynamic, but now that, that dynamic is extending into Europe, could you just walk us through how Cogent is capturing or indexing to the growth of streaming from the European region?

Yes. Sure. It's both Europe and the rest of the world. Cogent has more regional access networks as transit customers than any other provider in the world. We operate a global network of 97,000 miles of fiber connecting to over 1,300 carrier-neutral data centers in 48 countries. In that footprint, 7,400 access networks buy their upstream from us, from companies as large as China Telecom and China Mobile, Jio in India to small regional players. We've also seen the proliferation of new streaming service providers. So a couple of years ago, basically one name, Netflix, dominated the market. While Netflix continues to be a strong streaming partner and continues to grow, we've seen products such as Disney and Peacock and HBO Max broaden out that market segment. Cogent has several unique advantages. One is we have this geographic ubiquity, a footprint; and two, we're able to lower our customers' cost because in the majority of cases, in fact, approximately 70% of the cases, we're getting paid on both sides. That number is up from only 50% 2 years ago. So that means the party sending the traffic is paying Cogent to enter our network and ride our global backbone. And the company receiving the traffic is also paying Cogent. That two-side of your transaction is made completely transparent to our customers. It's one of the great competitive advantages we have, and it's the ability -- it gives us the ability to lower pricing because we're splitting that cost among 2 parties, and only a 30% of the instances is traffic going from a customer to a peer. Because Cogent's North American footprint is predominantly peered, our European and global footprint has a much lower percentage of traffic that is peered and a much greater percentage of traffic in which access network is also our customer. So we benefit 2 ways from this globalization, increasing the percentage of 2-sided traffic and the fact that we have the ability to deliver this traffic while many other carriers can't.

Got it. Got it. I wanted to go to just some of the pricing trends. Sorry, before we get into that, actually, let me just make sure -- yes, actually, pricing trends. You also noted pricing on megabits that were positive in 1Q. And I just wanted to just make sure that the durability of that pricing was going to hold for the remainder of the year, or should we expect any kind of real change happening there as the regional mix changes?

Yes. So 55% of our NetCentric business is outside of the U.S. It is typically charged in local denominated currencies. Secondly, the average price per megabit over the past 20 years on the Internet and for Cogent, this declined at about 23% per year. Investors should expect that as our long-term rate price decline. That price decline is possible because of the advances in technology. The 2 key technologies that we utilize are optically interfaced routers and DWDM, wave division, multiplexing optimal transport. Those continue to improve at substantial rates, and our network captures those efficiencies more effectively than any of our competitors. It is why we can continue to have the types of price declines coupled with margin expansion and declining capital intensity. In the first quarter, we had a bit of an anomaly in that our installed base was effectively flat. That was driven by the fact that many of our customers that we were selling were smaller customers rather than large. Over the long run, our installed base and our new sales should both decline at about 23% per year.

Got it. Got it. I wanted to pivot the conversation a little bit to just longer term trends you're seeing in NetCentric and where you think this business will be in the next couple of years. I was hoping if we could just touch on that. And one other factor is we've seen a lot of these new applications come to mind with streaming. We've seen Peacock. We've seen Disney+. We've even seen Netflix's business continue to grow. Are you still waiting for the next big killer app to kind of see the next step function increase in Cogent's business? Or where would you say we are in that process and that discovery?

Yes. So the NetCentric market is becoming more internationalized, which actually helps Cogent because of our footprint. Two, we have the largest sales force dedicated to this market segment of any carrier. Three, it is an important strategic business to us. And for that reason, we continue to focus the resources and gain market share. I think traffic will continue to grow on the Internet, albeit from a much larger base at a compounded rate of about 25% per year. That's in line with historical trends. That's in line with third-party studies such as Cisco's visual indexing projections. We typically grow somewhere around twice as fast as the market, gaining market share. We have seen most of our competitors deemphasize transit and pivot to other products. I think that bodes well for Cogent. I think there are really 3 ultimate drivers of traffic: number of people connected to the Internet; number of minutes a day that they are connected; and the bit intensity per minute. For the next several years, streaming will continue to be the key driver, but we will also see an increase in resolution rates going from 720K to high def to 4K and eventually to 8K. We're also beginning to see a convergence between gaming and traditional video with early examples of augmented and virtual reality. So I think there will be a continuous evolution in how video is consumed and what video means. I think that's different than a step function from some new application, but rather just a continued evolvement in the immersiveness of the video experience and the customization. So one thing that unicasting, which is how the Internet distributes video versus broadcasting, will accommodate is the ability for more customized content that will create greater consumer engagement and a more immersive experience. So I think we're still at the very early stages of how the Internet will be used for video consumption.

Got it. Got it. So I do have a list of peering questions, but we did get 2 questions that came in regarding your dividend. First, I want to ask another question that I have prepared, and we can kind of include the follow-up questions. So could you just speak to the company's capital return program that you have in mind and how maybe that capital return program has evolved or changed given the most recent refinancing you guys executed? And then I'll nest in the questions I got in.

Yes. So Cogent invests in its business first and foremost. But because we are so efficient, we have had excess capital since 2006. We have bought back about 20% of our flow. So we implemented a dividend and grew that dividend through our 35 consecutive quarters sequentially. We believe we will be able to continue to grow that dividend. To supplement our return of capital, we took a virtually unlevered balance sheet and have slowly added leverage to that balance sheet. And we've been very cognizant of the cost to capital. Our most recent financing of our senior secured debt lowered our cost of capital from 5.375% to 3.5%. On a blended basis, our cost of capital is below 4% today. That gave Cogent approximately another $10 million of free cash flow that can be diverted to equity and not payments to our debt holders. So we will be selective in using buybacks, but we're committed to a long-term methodical growth in our dividend as our 35 sequential quarter experience.

Got it. Got it. So here is the interesting question I got in my inbox. What has been the biggest pushback you have gotten from investors regarding increasing your dividend and maintaining such an elevated capital return program?

Yes. I think there are actually 2 concerns. The first one seems somewhat counterintuitive, which is if we raise the dividend materially, our dividend yield would be too high, therefore, creating some apprehension among maybe less knowledgeable investors. The thought is that when companies pay too high of the dividend, it's really an indication of the weakness of their business. Now my belief is that eventually, that dividend yield would normalize as the stock price increased. The second has been that we do use leverage prudently. We have a stated net leverage goal of between 2.5x and 3.5x EBITDA. It is lower than what our peer group typically has, and we naturally delever at a fairly accelerating pace. We've actually seen that delevering occur in the most recent quarter. We have been growing our dividend at roughly 15% per year, and some investors think that we should grow it faster. Others think that we should be more conservative and hoard more cash, thinking that there may be an opportunity to be more aggressive on buybacks. We really try to weigh all of our different inputs, both cost of capital and what investors thoughts, and we think we're taking a pretty measured approach of return to capital.

Yes. Another question came in regarding some of your assets. And the question is actually probably referring to what Zayo did, which is divest some of the data centers to another data center operator. Would Cogent ever consider that type of strategic maneuver where you're shedding assets that are, I guess, considered less strategic compared to some of the multi-tenant office buildings that you guys are a bit more strategic about?

We're always looking to maximize value for our shareholders. The only asset that I would consider nonstrategic would be our 54 data centers that we operate, but they only represent about 3% of revenues. And looking at that Zayo transaction, at least from the outside, it was not a completely arm's length transaction where the holding company was consolidating data centers under one operator and networks under another operator. In our case, we are not part of a holding group. So any transaction would have to stand the test of the market. We, I think, have shown a pretty good growth rate across all of our business segments. And as a result of that, I'm not sure there's really a trading arbitrage that occurs by separating our businesses. And those data centers are integrated into our network in that they provide hub locations and sales offices for routine. So I just don't see a clear set of assets that we could bifurcate out of our unified network.

Got it. Got it. So we can go back to business questions now that I've prepared. So the first one is, can you just talk about the peering business that you have now officially launched? And can you provide an overview of this offering and just how it compares to the existing peers or competitive landscape?

So when companies decide to connect to the Internet, there are really 2 methods to do that. One is to buy transit. That's what we sell to 7,400 networks. And it's, in fact, the way in which virtually 90% of traffic is exchanged. However, there is a second method, which is peer. Now traditionally to peer, you had to do 3 things. You had to build a very large global network. You had to get the counterparties to agree to interconnect to you, and then you had to have a very sophisticated staff to manage that. What we did is leverage our network to take advantage of Cogent's ubiquity and its relationships with customers and opened a global peer exchange. Prior to that, exchanges were geographically concentrated. So you exchange in Frankfurt or in Amsterdam or in London or in Dubai, but there was no exchange that offered a global footprint with flat pricing. And we've put that in place to help that 10% of the market that wants to have direct connectivity. We made it easy to operate. It does still require the 2 parties to agree to exchange traffic with one another. So we created an automated website that allowed customers to go in and quickly see who was available, send them a request to peer and get an answer quickly, much like a dating site. And if the 2 parties agree to exchange traffic with one another, because of proprietary technology, we're able to automate the provisioning of that circuit. Without that automation, it would be virtually impossible to build this on the scale that we had.

It does look like we lost a connection with our moderator. But if you have any other comments to share, David, you can go.

Well, I think we're nearing the end of our session anyhow, maybe Sami have to jump for another fireside. But I think Cogent is uniquely positioned due to its network architecture, its global reach, its on-net footprint in North America to continue to grow top line. Our average growth rate over the past 16.5 years as a public company is 10.2%. With that, we've been able to expand margins of 200 basis points a year. We think we can continue to do that. And with that expansion in margin and growth in top line, we'll be growing our cash flow between 15% and 20% per year, and we're committed to returning that to investors. So because we're focused only on the fastest growing segment of the telecommunications ecosystem, the Internet, that gives us a unique position compared to other telecom service providers. And the architecture that we build using dark fiber IRUs from 270 suppliers and then IP directly over DWDM allows us lower capital intensity and better operating efficiencies than any of our competitors. So I think we're in a great place for as the world reopens, as Internet traffic continues to grow at an accelerating rate due to streaming, Cogent is the company that is probably most uniquely positioned to benefit from these trends. I'd like to thank everyone. I'd like to thank Sami and Crédit Suisse for hosting us.

Thank you so much, David, for your time here today.

Thank you.