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

Hi. My name is Phil Cusick. I cover the comms services and infrastructure space here at JPMorgan. Thanks for joining us. It's my pleasure today to introduce Dave Schaeffer, the CEO of Cogent Communications. Dave, thanks for joining us.

Thanks, Phil, for hosting me. A pleasure to talk to everyone over the video. And while I enjoyed Boston, I have to admit, I enjoyed not having fly up there for the conference, Phil. So thanks for hosting me yet another year.

Thanks, again, for taking the time. Just given everything going on with the pandemic, maybe you could take a minute and talk about what Cogent is doing to take care of customers and employees through all of this.

Yes. So let me start with employees. In early March, we went to a voluntary work from home. We quickly modified that and went to mandatory work from home, so virtually all of our employees in our 38 offices around the world are working remotely. We did have to scramble as 288 of those employees only had desktops, so we had to get laptops for those employees shipped to them and get them configured with cellphones so they could connect into our VPN. We also modified our hiring process. We are different than most companies in that we are continuing to hire throughout the pandemic, and we had to come up with a program to get new employees onboarded remotely. So we would ship them laptop, but through FaceTime or WhatsApp or Skype, walk them through the setup so they can then begin their online training. And we anticipate that at least for another month or so, all of our employees will be working remotely with the exception of field services. For field services, we had to get them certified as essential workers so they could move throughout markets while there were mandatory shelter-in-place orders. For our customers, we have done a few things. One, we've continued to offer compelling value. Two, we rolled out a promotion for our corporate customers allowing them to increase their bandwidth by tenfold, going from 100 megs to 1 gig for a flat increase of $200 a month for the remaining contract term. Since many of our corporate customers are running their work-from-home VPNs through their firewall, this additional bandwidth was very helpful to them. And then for our NetCentric customers, we saw a real scramble globally for more port capacity as we experienced a real acceleration in the rate of traffic growth, a surge in traffic for about 4 weeks at very unusually high levels. We saw roughly 7,000 access networks around the world, adding additional capacity and our 5,000 content producers. In our NetCentric business, we do get paid by the bit, so it is in Cogent's best interest to accommodate this additional traffic requirement. So it's really been a busy 6 weeks for us.

Sounds like it. Let's follow up on a few things you just mentioned. So you had said on the conference call that there were a few weeks where areas were shut down and you couldn't go into buildings. So you have those field services techs certified now. Is the sort of install and network conditioning pace up to regular at this point?

Almost. There are probably less than 10 buildings globally that still will not allow our personnel into the buildings. They won't allow anyone in the buildings. They're effectively shut down or quarantined. There are still, at least, a dozen data centers in Europe that are not allowing tech access, and we must use helping hands from the data center operator. But out of a footprint of nearly 2,900 end buildings, 1,800 multi-tenant office buildings and about 1,050 data center buildings hosing -- housing about 1,230 data centers, having a dozen or so problems is pretty good. And in all of our markets, we've been able to work with regulators and get our employees certified as essential services. So even in the markets with the most stringent lockdown conditions, our employees have documentation that they can't move around freely because the Internet is really viewed as a lifeline to society.

Okay. And on your corporate sales side, you mentioned the hiring issues. Should we think of further disruption this quarter? You saw a little bit of disruption last quarter in terms of sales. Should we think that more this quarter is customer -- as employees just got into homes, but people for them to call even, how is the overall sales process looking today in mid-May?

Well, let's first of all recap last quarter. We saw our sales force productivity increase from 4.1 installed orders per rep to 4.5 sequentially. In the quarter, we lost 8 days of selling time due to our sales conference. And then in addition to that, we also lost some momentum on installs at the very end of the quarter, yet we saw a 10% increase in productivity, while our head count increased by nearly 4% sequentially. Going into Q2, I think those issues are behind us. We continue to hire additional reps. We continue to onboard those reps and train them, and we also manage out underperformers. Our sales force turnover has averaged over 5 -- about 5.7% of the sales force per month historically. We did have to modify our review process, but we are able to still manage out those underperformers. We've been monitoring sales force efficacy, looking at leading KPIs. Call volume per rep is up, e-mail activity per rep is up and CRM activities per rep are all at elevated levels. Now to hear the sales force say that they just all want to work from home and think it's a better environment, I'm a little more skeptical, and I believe that improvement in productivity is because the people on the other end of the phone actually want to talk to us as opposed to hang up on us. The crisis in work from home has caused people to reexamine their local area networks and their connectivity in their wide area network, and that has them receptive to our sales pitch. As we've discussed in the past, 72% of our sales people are corporate and most of what they do every day is ineffective because the customer is not interested in talking about more bandwidth. In a normal environment, we would get moved -- when a customer moves, that's a window of opportunity. When a current provider fails, that's a window of opportunity or a change in their IT infrastructure with a move to cloud or SaaS, all being catalysts for the customer being receptive. The COVID-19 crisis has caused yet a new catalyst as all the employees no longer report to the office and work from home. Most of the companies need to upgrade their connectivity to the Internet at their firewall, and that makes the IT staff very interested in talking to a Cogent rep and taking them up on the offer that we have. So this has been, I think, net beneficial to our corporate sales efforts.

Okay. So as I think about -- it's hard, we're, what, 40 days into the second quarter, should we think about overall it's net beneficial for the second quarter? Or is there still like a couple of weeks of delay and so things overall net out about even?

I think, overall, a net positive, Phil, I think the only area that we saw a concern is on our off-net installs. We've seen those installed windows lengthen. Most of our off-net installs come from companies that use union labor, and some of those union employees have not often willing to go out and do the fieldwork necessary. But in general, I think we're doing pretty well. On the corporate side, I think we're still getting through to people. We're still selling new business. We're upgrading customers, so I think we have some tailwinds and should do better in Q2 than we did in Q1. And then on the NetCentric side where it's a volume-based business, we saw this really elevated rate of growth for about a 4-week period. We've reverted back to a more normalized growth rate but still at an elevated level, and it's really being driven by streaming video, and I think that trend has continued to improve. What we basically saw was in a month, we saw a year of -- 1 year of streaming conversion in 1 month. And now we're just a year ahead off of a higher base.

Okay. And as I think about churn through the quarter, should we think that businesses that are open, for the most part, aren't leaving you at this point, that is, the higher rate?

So again, let's disaggregate our customers and talk about both customer bases. For our corporate customers, we have the benefit of having those customers pre-vetted by the building owners. Our footprint only touches the largest and typically most expensive office space in any of the markets that we serve, roughly 110 North American markets. And as such, even though we sell small- and medium-sized businesses, our customers tend to pay their bills and stay in business. We did not have the traditional pizza shop or a carryout or a hairdresser as a customer, and therefore, our bad debt expense has remained low. It's about 0.8 of 1% per month. We have been monitoring our daily collections. From April 1 through May 11, our collections were up 3% over the same period in 2019. Our revenues were up about 5.5% during that same period. So we are seeing a slight degradation in collections. Our DSOs went from 23 days to 24 days. That's nothing to sound the alarm about. Most wireline carriers run with about 60 to 65 days of sales outstanding. So I think we're doing well on that front and do not expect elevated churn. Our on-net churn in the first quarter was 1.1%, identical to the fourth quarter. Our off-net churn ticked up ever so slightly from 1.1% to 1.2% in the quarter, both are within historic norms, and we expect kind of similar churn in Q2.

Okay. And you mentioned the $200 upgrades. How effective has that been in driving people up? And will we notice that in the numbers?

So we've already been noticing the migration to gigabit speeds in our ARPUs. Our on-net ARPUs have plateaued, even gone up a couple of quarters as the majority of new sales have been geeky at 900 versus a typical capacity or a 100-meg connection at about 650. The customers upgrading plus the fact that most new sales continue to be gigabit should help on-net ARPUs, that is a blend of corporate and NetCentric. And then in the off-net segment, our ARPUs have been declining for 2 reasons: one, longer contract terms, just like on-net, but also lower loop cost, as we've seen a lot more competition between cable and telco for those local access loops. I think going into Q2, should expect similar rates of ARPU decline as you saw in Q1.

Okay. Okay. That makes sense. Let's talk about the NetCentric business a little bit. You talked about demand really accelerating for the first month and then still fairly high. Talk about the breadth of that increase in demand. And has the -- we talked about it on the earnings call that it's broadened, has that continued? And is this really the catalyst we've been waiting for, for the NetCentric overall revenue growth to start to return to a more normal level?

So we clearly saw an improvement in NetCentric revenue growth in Q1 versus Q4. We think that will continue going forward. We're still not back at the historical averages of 9% year-over-year, but the 1% sequential growth was a good step in the right direction. That's been helped by accelerated traffic growth, particularly in the last 2 weeks of the quarter. That resulted in full quarter traffic growth growing 12% sequentially versus 9% the previous quarter. We've also seen year-over-year growth growing at 36% in Q1 versus 30% in the fourth quarter of '19. Those are all very positive trends, and they should continue going forward. The other positive trend is we've seen that growth coming from a much broader group of customers. We carry traffic for every large provider globally. And for the past couple of years, the vast majority of traffic growth has come from a handful of household names. Those names are continuing to grow, but we're seeing some new entrants in the market, particularly some of the new over-the-top streaming services become significant contributors to traffic growth. But because those providers are still at a smaller scale than some of the legacy providers, we have been able to get better effective pricing per megabit.

Okay. Any change in the competition in that NetCentric space? What do you see from other providers of that service?

Yes. So 20 years ago when Cogent was founded, there were nearly 200 global backbones competing in that market. Today, there is about a dozen of us in that market. For the larger customers, CenturyLink is the largest player. Cogent is #2 in the market, Telia #3 and NTT #4. There are really only 4 players for a large NetCentric customer, whether it be access or content. For the smaller, more regional users, there are 8-or-so regional players, companies like Telefónica and Telecom Italia, France Telecom and Deutsche Telekom. But I think every one of our competitors continues to deemphasize transit because of the 23% year price compression and the fact that it is a static addressable market. Cogent has been able to gain market share by underpricing our competitors by 50%, that policy continues going forward, and we anticipate that we'll continue to grow substantially faster than the Internet with price declines similar to market-based price declines but from a lower base. The net result of that dynamic should be that Cogent's NetCentric revenues continue to improve and converge back to that historic growth rate of about 9%.

Okay. How -- you've seen a little less than that 9% for quite some time. As you look at the trends over the last few years and again, the recent broadening we've seen, is your confidence in returning to that 9% as you even -- not just in the near term, but as you look out over the next couple of years?

So first of all, it's difficult for us to have that level of visibility. Even our customers don't have that level of visibility because it is a usage-based service. Our NetCentric business has underperformed for the past 7 years. First, with the loss of meg upload; second, with some of the violations of net neutrality constricting the Internet; and then third, this heavy concentration of growth with a few particular customers. All of those trends are receding, and we're returning to a more normalized NetCentric growth pattern. I fully expect that, that will help revenue growth accelerate, and we should get back to that kind of historic run rate of 9%. And we may even have the opportunity in some quarters to exceed that. NetCentric revenue has been volatile over the years, and there have been a number of quarters where we would materially outperform the average. And I think we're trending towards at least average performance, if not, outperformance.

Okay. Another connectivity provider called out some delays from customers taking time to reassess their network investment. Has Cogent seen customers take any kind of wait-and-see approach? It seems like that's been the opposite lately.

I worry about that. And clearly, we have a lot of both health uncertainty and macroeconomic uncertainty. What we have seen so far is people acting in a crisis mode, meaning they have a problem, they just need it fixed. And their question is, "How quickly can you get it fixed for me?" I think going forward, one of my biggest fears on the corporate side is the sales force just won't be able to get the buyer on the phone to have a meaningful conversation. That's just not been a problem yet, but it is something we're monitoring. And at some point, IT managers will no longer forward their phones, so their personal cellphones be as responsive to e-mail and LinkedIn. Right now, the sales force is extremely effective because of calling someone who wants to talk to them. At some point, either employees go back to the office when the crisis is behind us or people don't perceive this to be a continuing problem. And therefore, the problem goes away.

Okay. Let me just switch gears and talk about your data center business. I heard from one of your peers today who talked about their sort of legacy carrier locations as potential edge data centers as customers need storage closer to the edge. How do you think about that? And to me, it seems like a number of facilities that didn't work out as colocation and now repurchasing those to edge in low latency. Does that make sense to you?

It does not, to be honest. I know that's a very hot topic in the market. We have seen CDNs, which are really some form of edge compute try various models. Maybe the initial CDNs had as many -- as nearly 1,750 locations. That proved not to be efficient. And you were not getting adequate utilization of your compute and storage capacity. Most of the more modern CDNs, whether they be proprietary to an OTT player or commercial, the third-party players have converged on a model of about 30 locations, globally, getting the right balance between latency and facilities utilization. One of the things that made the Internet so successful is the fact that it is usage-based but not distance-based. So there's no distance sensitivity to the pricing model. Now there are advantages in performance as you get closer to the customer but those advantages start to diminish as you get more sites out and you start to find the cost of housing that content in those widely distributed sites to become prohibitive. The space at the base of a tower or in an edge market is fairly precious real estate. And what we're seeing is more of a concentration of compute and storage into hyperscale facilities that are typically located in million-plus square feet buildings that are built with subsidies, both in terms of real estate tax subsidies and power subsidies. You won't have that kind of market leverage in an edge environment. I just see a limited market opportunity for edge data center space.

And so not -- maybe it's directly to your business, but if there is a world that needs much lower latency, how is that addressed if it's not in a number of very close, relatively small data centers?

So we today connect to over 1,230 carrier-neutral data centers in 47 countries around the world. We connect to over 7,000 access networks. We have a footprint that gets content very close to the networks that distribute them to end users, both in terms of geography and direct AS, or autonomous system, adjacency. One of the things we do routinely for our NetCentric customers is a traffic analysis. We will show them where their traffic is entering the Cogent network, what geographies it's exiting in and to what networks it is transiting to. We do a similar analysis for access networks. We show them where they're pulling traffic into their network and what networks it's coming from. That analysis with our sales team and sales engineers allows a consultative sale and allows our customers to optimize their footprint. So they are taking more capacity at locations that are more relevant. But getting beyond one of those carrier-neutral facilities to a proprietary edge facility has never really made economic sense, and it has not materially improved performance.

Okay. A different topic we've talked about many times is the current ecosystem of VPN and MPLS and how new technologies, like SD-WAN, could over time cannibalize that. So first question, how do you see the current ecosystem of MPLS and VPN? And how has that changed in the last year? And second, give us an update on your...

So let me start with a little bit of a primer on what we are talking about. The SD-WAN that we offer is a point-to-point service. That is the same thing we offer with VPLS. The demand for that service is probably actually down in COVID-19 because there's less need to connect office-to-office. Now a counterbalancing force is the legacy MPLS services are rigid and very expensive. So -- and as companies look to optimize their IT infrastructure and reduce cost, they will look to migrate to one of those 2 technologies. The technology that is most in demand right now with work from home is a dynamic VPN, not based on either of those technologies. So it's much like the Zoom call that we are on today where there is an ad hoc connection being produced. So if you're a work-from-home employee, you're using your local broadband connection, you're effectively tunneling in to a firewall to get into the corporate local area network and then retrieve information and possibly use computing power. That is a very different type of VPN and is completely ad hoc based on clients in the end user's device, either a phone or PC. For the dedicated VPNs, MPLS continues to decline. We've already seen about a 20% reduction in that marketplace. We're seeing that rate of decline accelerate as customers look to jettison those older networks. SD-WAN continues to be the primary technology that customers ask for. But when you go through the pluses and cons of those networks, coupled with the current state of the equipment vendors' architectures, most customers opt for a VPLS solution and not SD-WAN. So we continue to sell those products. I would say that virtually all of our customer inquiries are for SD-WAN, but the majority of the sales continue to be VPLS. Now as the customer premise equipment becomes more robust and prices come down and feature sets become really usable, I think SD-WAN will continue to gain market share. But I think for the legacy wireline industry, we are on that sharp price performance decline part of the curve, just like we were with voice when voice moved from plain, old telephone to VoIP. There's still some cost services left, but they're declining at a slower rate. In MPLS, we've gone from the flat part of the S-curve to the nearly vertical part, and we're probably going to be on that for several years, and it's why wireline service providers, enterprise revenues globally, had been in decline, and I think that rate of decline is accelerating. And we are a net beneficiary because we don't have a declining revenue base and have the opportunity to sell either SD-WAN or VPLS alternative.

Okay. And the last topic that I wanted to hit, and this is maybe even further afield from you, but consumer broadband, we've seen a massive acceleration in cable taking share this quarter, mostly it looks like from DSL and for people who are relying on their wireless product. And we've also seen in the last 6 months a couple of different, let's say, either private or emerging bankruptcy transactions where people are talking about building a lot of fiber to the home over time. So with broadband usage and revenue per user rising, have you had any change in your thinking around the value of building fiber or overlaying copper with fiber?

So first of all, I believe strongly that fiber is the correct technology and the outside plant in the developed world needs to be rebuilt. About 16% of the U.S. today has fiber connectivity. There has not yet been a fiber provider whose returned a return on capital greater than their cost of capital and fiber overbuilds, whether it be Verizon with FiOS or Google with its product. I think the regulatory environment that promotes facilities-based competition has hurt fiber deployment. Secondly, 67% of Americans connect broadband via cable. The deficiencies of that cable plant are becoming more and more obvious as the plant is more extensively used in this environment. So I think post COVID-19, there will be an examination of the cable operators, and their lack of investment in plant. And I think there may be new regulations put in place that promote private capital, overbuilding plant with fiber and getting an adequate return. So it's the right thing to do for the consumer, it's turned out to be the wrong thing, so far, for the investor.

Understood. Well, that's a good place to leave it. Dave, thanks very much for joining us. And everybody on the line, thanks for joining us today. We'll see you tomorrow. Have a great day, Dave.

Thanks a lot, Phil, for hosting.

Thanks.