Cogent Communications Holdings, Inc. Q1 FY2021 Earnings Call

Company Representatives: David Schaeffer - Chairman, Chief Executive Officer; Sean Wallace - Chief Financial Officer; Walter Piecyk - LightShed; Tim Horan - Oppenheimer; Nick Del Deo - MoffettNathanson; James Breen - William Blair; Evan Young - KeyBanc; Frank Louthan - Raymond James; Angela Zhao - Bank of America; Bora Lee - RBC Capital Markets.

Good morning. And welcome to the Cogent Communications Holdings First Quarter 2021 Earnings Conference Call. As a reminder this conference call is being recorded, and will be available for replay at www.cogentco.com/. A transcript of this call will be posted on the same website when it becomes available. Cogent’s summary of financial and operational results attached to its press release can be downloaded from the Cogent website. I would now like to turn the call over to Mr. Dave Schaeffer, Chairman and Chief Executive Officer of Cogent Communications Holdings.

Thank you and good morning everyone. Welcome to our first quarter 2021 earnings conference call. I’m Dave Schaeffer, Cogent’s Chief Executive Officer. With me on this morning’s call is Sean Wallace, our Chief Financial Officer. Now, for a few comments on our results: As a number of COVID-19 cases in the United States have begun to decline and businesses have initiated plans to begin reopening offices, we have begun to see an improving business climate for our corporate segment. We have seen declining unit churn in our corporate segment sequentially for the last five months, and our churn levels for the first quarter were lower than the elevated levels that we saw in 2020. Our NetCentric business continues to benefit from the greater-than-expected growth in streaming subscriptions and the segment saw its third sequential quarter of accelerated growth. For the first quarter, our traffic was up 8% on a sequential basis and 36% on a year-over-year basis. Despite these improvements, we remain cautious in our outlook about the uncertainty of the current economic environment and the challenges that the pandemic has brought. Our first quarter results show a higher level of growth in revenues as our revenues grew sequentially by 2% to $146.8 million, an increase of 4.2% on a year-over-year basis. On a constant currency basis, we experienced a revenue growth rate sequentially of 1.7%, and year-over-year growth rate of 2.3%. We continue to operate an extremely efficient network. Our network serves a growing number of markets, including carrier-neutral data centers and multi-tenant office buildings in the central business districts of North American cities, and is able to handle the increase in traffic volume at a relatively fixed cost base. We’ve experienced year-over-year and sequential growth in our non-GAAP gross profit, our non-GAAP gross profit margin and substantial year-over-year growth in our EBITDA and EBITDA margin. Our non-GAAP gross profit grew by 2.7% sequentially and grew by 7.6% on a year-over-year basis. Our non-GAAP gross margin percentages for the quarter improved by 40 basis points sequentially to a record high of 62.5%, and grew by 200 basis points on a full year-over-year basis. Our quarterly EBITDA grew by 10.2% on a year-over-year basis. Our quarterly EBITDA margin increased by 200 basis points from the first quarter of 2020 to 37.8%. The performance of our existing customer base continues to be strong despite the impact of COVID-19. Customer churn, days sales outstanding, and most importantly, cash collections all improved in the quarter, as I noted earlier, and our corporate churn rate fell for the second quarter in a row.

Thank you, Dave. Good morning, everyone. This earnings conference call includes forward-looking statements. These forward-looking statements are based upon our current intent, belief, and expectations. These forward-looking statements and all other statements that may be made on this call that are not historical facts are subject to a number of risks and uncertainties, and actual results may differ materially.

Thanks Sean. Hopefully, you’ve had a chance to review our earnings press release. Our press release includes a number of historical metrics that we report on a consistent basis. Now for a few comments against our results, against long-term multiyear targets that we have been reaffirming. Our targeted long-term EBITDA annual margin expansion guidance is for approximately 200 basis points per year of margin expansion. Our targeted multiyear constant currency long-term growth rate is approximately 10%. Our revenue and EBITDA guidance targets are intended to be multi-year goals and are not intended to be used as quarterly guidance. Our corporate business represents 62.7% of our revenues. Our corporate business declined year-over-year by 5.1% from the first quarter of 2020 and declined by 1.8% from the fourth quarter of 2020. Our NetCentric business, which represents 37.3% of our revenues, had yet another strong quarter and showed accelerating growth, sequentially growing 9.2% quarter-over-quarter and growing by 24.6% as compared to the first quarter of 2020. Volatility in foreign exchange rates primarily impacts our NetCentric business as approximately half of that business is outside of the U.S. On a constant currency basis, our NetCentric business increased by 18.7% from the first quarter of 2020 and by 8.3% on a sequential basis from the fourth quarter of 2020.

Thanks Dave, and again good morning to everyone. Talking about corporate and NetCentric revenue and customer connections: We analyze our revenues based upon network type, which includes on-net, off-net and non-core, and we also analyze our revenues based on customer type. We classify all of our customers into two types, NetCentric customers and corporate customers. Our corporate customers buy bandwidth from us in large multi-tenant office buildings or in carrier-neutral data centers. These customers are typically professional services firms, financial services firms and educational institutions located in multi-tenant office buildings or connecting to our network through our CNDC footprint. Our NetCentric customers buy significant amounts of bandwidth from us in carrier-neutral data centers, and include streaming companies and content distribution service providers as well as access networks who serve the consumers of content. Revenue from our corporate customers for the quarter fell sequentially by 1.8% to $92.0 million and fell year-over-year by 5.1%. An increase in the USF tax rate had a $0.4 million sequential positive impact on our quarterly corporate revenues and had a $0.8 million positive impact year-over-year. The USF tax rate changes quarterly, and we cannot predict the impact of future USF rate changes on our revenues. As we have discussed in previous earnings calls, we believe that the growth of our corporate revenues was directly impacted by reduced building occupancy in central business districts of major cities as a result of the COVID-19 pandemic. We also found that as a result of the work-from-home environment and general challenges from that pandemic that many of our corporate customers delayed decisions about system upgrades and new network investments. The slowdown in sales, combined with normal historical levels of churn, has contributed to a modest reduction in corporate revenues for the past several quarters. On a positive note, we are seeing some signs that indicate corporate buying patterns are beginning to return to a more normal level. Sales of our largest product by revenue and connections, our one gigabit per second direct Internet access product, had its third quarter in a row of rising sales. We are also seeing some of our larger corporate clients begin to expand and reconfigure their networks. In terms of churn in our corporate base, we are encouraged that churn has fallen in each of the last five months, and most of the churn is derived from our older 100 megabits per second VPN and direct Internet access products, and we continue to see very low levels of churn from our one-gigabit connections.

Thanks, Sean. I’d like to highlight a couple of the strengths of our network, our customer base, and our sales force. As I stated earlier, we saw an acceleration in revenue growth in our NetCentric business during the last three quarters and an impressive NetCentric growth rate of 24.6% year-over-year. Streaming subscriptions, primarily in international markets, continued to outpace analysts’ expectations and we are a direct beneficiary of this trend. I want to highlight some of the key statistics that we believe reflect some of the strength in our NetCentric business. At quarter-end, we connected to 1,274 carrier-neutral data centers and 54 Cogent data centers. This is more than any other carrier globally, as measured by independent third-party research. The breadth of this coverage allows our NetCentric customers to better optimize their networks and reduce their latency. We expect that will widen our lead in this market as we project to connect to over an additional 100 carrier-neutral data centers per year for the next several years based on the planned construction pipeline of these facilities. At quarter-end, we directly connected to 7,470 networks. This represents a 6.1% increase from a year earlier. This collection of ISPs, telephone companies, cable networks, mobile network operators, and other carriers allows us direct access to the majority of the world’s broadband subscribers and mobile phone users. At quarter-end, we had a sales force of 239 professionals solely focused on the NetCentric market. We believe that this group of professionals gives us one of the largest and most sophisticated sales teams in this industry segment. Now for some comments on our corporate business. We are seeing some positive trends in our corporate business, but they have yet to fully materialize in positive revenue growth. As the work-from-home environment becomes established as part of the way people will work going forward, our corporate customers continue to look to upgrade their internet access infrastructure to larger connections and symmetric connections. Our corporate customers are aggressively integrating some of the new applications that have become part of the work-from-home world, such as video conferencing. This usage will require high-capacity connections, both inside and outside of the premise. As Sean mentioned earlier, this has been an aggressive push for us to increase our corporate customers’ connection size to either one gigabit, and in some cases 10-gigabit symmetric connections. Our sales force continued to experience improvement from the continued training efforts and the accelerated management out of underperforming reps. As a result, on a sequential basis, our total sales force headcount did decline to 547 reps and our full-time equivalent reps shrank to 522 full-time equivalents at quarter’s end. Year-over-year, our sales force increased by five reps or 1%, and our full-time equivalent rep count was flat. Our sales force turnover was 6.6% per month in the quarter, which is a slight drop from the 6.9% per month that we experienced in the fourth quarter of 2020, primarily as a result of our disciplined approach in monitoring our remote sales force and managing out those underperformers and improving the efficacy of our sales force. These factors resulted in a continued rebound in sales productivity to 4.3 orders per full-time equivalent per month, a 2.4% sequential increase from the 4.2 orders installed per full-time equivalent per month in the previous quarter. Overall, we believe our sales force has accomplished a great deal in the past year. The first quarter was our sales team’s best quarter in the company’s history. Much of that business is installing and will be fully reflected in the second quarter. I wanted to thank our entire sales force and the entire Cogent support team for all they’ve done and the hard work that they’ve had throughout COVID-19 and the continued improvements that we’re implementing in 2021. So in summary, Cogent is a low-cost provider of Internet access transit services. The value proposition remains unmatched in the industry, demonstrating our low-cost position is the fact that since 2016, we have lowered our cost of goods sold per byte mile transmitted by an annual compounded rate of 22.5%. We remain optimistic about our unique position in serving small and medium-sized businesses located in the central business districts of major cities in North America, with approximately 1,800 multi-tenant office buildings on-net and over 975 million square feet. Despite the recent drop in new tenancy in many of these cities, we are beginning to see landlords aggressively entice new tenants into their space with lower rents, shorter lease terms, and improved tenant amenities, which will keep these Class A buildings at high occupancy levels. As COVID vaccinations increase throughout the United States and in Canada, our tenants will begin to return to the multi-tenant footprint. We believe that our corporate business can return to its long-term historic growth trends. Our customers' churn, bad debt, and days sales outstanding are all better than our historic numbers. We believe this represents the unique, high credit quality of our customer base and the value of our service to these customers. Our targeted multiyear constant currency revenue growth rate of 10% and EBITDA margin expansion of 200 basis points per year remains our current best outlook. Our Board of Directors has increased our dividend for the 35th consecutive quarter to $0.025 per share per quarter, to $0.78 annually, representing a 14.7% annual growth rate in dividend. Our consistent dividend increases demonstrate our optimism in the business, the operating leverage of our business model, and most importantly, the accelerating cash flow generating capabilities of our business. While we did not repurchase stock in the quarter, we did have over $30 million available for a repurchase program to remain in place through year-end. Now I’d like to take a moment and just close on some of our capital markets activity. We are looking to optimize our balance sheet and take advantage of the current interest rate environment. We anticipate being able to lower our cash interest expense by over $10 million a year. In the first quarter, we repurchased $115.9 million of principal senior secured notes that are due March of 2022 with cash on hand. We also issued a conditional call for another $45 million of these notes. Assuming that we perform on this call, our pro forma leverage will fall below 4.25 times as defined in our indentures, both in our 2022 and 2024 notes. As a result of that, we will be able to transfer an accumulated builder basket from the operating company to the holding company, freeing up more capital to be returned to shareholders. As part of this capital management program, we are planning to offer $500 million of new senior secured notes due in 2026 under Rule 144A. Yesterday afternoon, we released our 10-Q for the quarter ending March 31, 2021, and our quarterly press release to facilitate this transaction. My apologies to our research analysts who expect this information early the next morning, but we felt it was critical to get this information into the market so our potential debt investors could analyze our improved business outlook. If we are successful in pricing this offering, we will complete our conditional call, retire $45 million of our 2022 senior secured notes, we will reclassify $121.5 million into our permitted restricted payments, and we can then move this cash from holdings to fund future dividends and share buybacks. Post this reclassification, we will repay the full balance of the senior secured 2022 notes and issue our new 2026 notes. Pro forma this transaction, we will have $390 million of cash on hand in total and $145 million available at the holding company for dividends and share repurchases. Again, I’d like to thank all of our investors for their continued support and the entire Cogent team for an excellent quarter. Now I’d like to open the floor for questions.

Your first question is from the line of Colby Synesael with Cowen & Company.

Hey, this is Michael on for Colby. Two questions if I may. Were there any one-time items we should be aware of in the quarter, particularly within the NetCentric business? As part of that, how should we think about the sustainability of the momentum and growth that we’re seeing in the NetCentric business currently? Also, regarding the mentioned improvement in the sales backdrop, is there any quantification you could give us on what you’re seeing or the improvement you’re seeing in your corporate sales funnel either quarter-over-quarter or year-over-year? That would be good. Thank you.

Yes, sure Michael, thanks for the question. First, there were no extraordinary items in our NetCentric business. This was actually a very broad number of customers, both content delivery and access networks increasing the utilization of our network. We benefited from the fact that our new sale price actually increased rather than declined. This is a result both of foreign currency translations since nearly half of that business is outside of the U.S., and the continued broadening of our customer base. Almost 69% of our traffic results in a two-sided payment, where both the customer sending and the customer receiving are paying Cogent. Only about 31% of our traffic exits or enters our network through one of our settlement-free peers. To remind investors, Cogent does not sell peering or purchase peering or transit as a service. To your second part of the NetCentric question, we think that this reflects an accelerated shift in viewership throughout the world, moving from linear video to streaming video, improving resolution, improving content, improving choices and subscription models. These trends first occurred in the U.S. and are now becoming more internationalized. That should bode very well for the NetCentric business, and in particular Cogent should capture a disproportionate share of that due to the global nature of our network, the carrier-neutral footprint that we described, and the sheer number of access networks that connect to us versus any other carrier, giving us a sustainable advantage. Now, this is a usage-based business. It typically does not grow as rapidly sequentially month-to-month because people spend more time outside and less time consuming video. This year may be a bit different, particularly because some of the lockdowns throughout Europe have meant people are not able to go out and enjoy the good weather the same way they normally would be able to. In the U.S. we are seeing a more normalized seasonal pattern. As I think about the corporate business, the 24% plus year-over-year growth that we delivered was either the best or the second best in the company’s history of over 21 years. That business ebbs and flows, but has averaged a 9% growth rate. We were underperforming ever since the loss of Megaupload and some of the peering issues with violations of net neutrality. We are encouraged by the current regulatory backdrop and the current administration’s commitment to an open Internet. All of this bodes well for this business performing at about historic trends for the foreseeable future. We may have several very good quarters, but over the next five to ten years, we think the NetCentric business can exhibit that high single-digit 9% or 10% year-over-year growth rate. Now to the final part of your question around corporate visibility: one, we have seen the decline in churn rates. We have seen most of that churn come from smaller, older customers who have built economic pressure because of the pandemic. Offsetting that has been the increase in larger connection sizes and the increased activity level of the sales force as measured by number of spoke-2s per day, number of spoke-2s converted in the opportunities and number of opportunities sold, and that is demonstrated by the increase in sales force productivity with over 65% of the sales force focused on the corporate market here in North America. Many companies are expecting to return to the office, either on a full-time or a hybrid basis. These companies realize that they need to plan for that return to work and still accommodate work from home. That has been a positive trend for us as companies now are looking for symmetric, larger connections and more locations. There will be some network grooming of some offices, but for the most part we anticipate sequential continued improvement in our corporate business and a return to its long-term average growth rate as corporate workforces return to the office in a post-pandemic world. Next question, please.

The next question is from the line of Walter Piecyk with LightShed.

Hi, Dave. So let’s just go back to the corporate, obviously it’s been down. Would you explain the USF of 2% or more in the last three quarters sequentially? Aside from the long-term returning to growth as people come back, how do the trends look for the current June quarter? Can you maintain flat revenue in the June quarter?

So I actually think we will see continued improvement. The rate of decline has troughed as I commented. The churn numbers have sequentially declined now for five consecutive months and preliminary data for this month indicates it will decline even further. In addition to that, based on sales activities, we actually believe that the corporate business will improve on a sequential basis in the second quarter. Whether that is enough to return to an aggregate positive number is yet to be seen. In many parts of the country, probably everywhere except the Northeast and Toronto, we’re seeing an increase in office re-openings and an immediate need to accommodate this new hybrid work model.

So, it’s improving? When you say improving you're just seeing some KPIs within the business, but it’s unclear if that can deliver flat revenue sequentially, do I understand that right?

That’s correct. I do have visibility to the fact that the rate of revenue decline will improve. Will it improve to a zero rate of decline; it’s too early in the quarter to tell.

Walter, I’ll just add into Dave’s insights. If we look at what happened in 2020, what’s interesting is we look at the two products: the GigE product which is 1,000 megabits per second versus 100. All that churn was in our 100 megabit product. We grew every month on our gigabit product and as we’ve discussed earlier, that mix between 100 meg to gig has changed: 100 meg was our biggest product in the first quarter of last year and is now much lower in proportion. The impact of that decline is going down. We also look at growth in our direct Internet access product. In January our units declined, but in February and March we saw the first increases in net units since the beginning of last year. So we see a bunch of green shoots all over the place: lower churn; companies beginning to look at how they’re going to operate in the new environment; great growth in our 10-gig VPN product, which is a small product but tripled in units. So we see all sorts of improvements: the churn that we had suffered over 2020 is going down, corporates are beginning to reconfigure their network and make decisions, and that typically means they’re going to increase their capacity and we’re the right company to do that.

Got it, thanks Sean. And can I just ask one quick question also on the balance sheet? When you cite your net debt numbers, you exclude the operating lease liabilities, which obviously increased over the course of the year. Can you refresh our memory in terms of why that went from $88 million Q1 last year to $109 million this year? And then is there an offset in terms of expenses that are not booked as expenses and are just depreciated, and if so, does that factor into the debt covenants that you’re citing when you’re talking about this 3.2x debt leverage?

So if you remember, last year FASB changed the rules for lease accounting. You also have to classify certain future contingent obligations as operating lease liabilities. As a result, we saw some of our expenses reclassified as capital leases, particularly some long-term O&M contracts on fiber IRUs. Secondly, we saw some future obligations now need to be classified as operating leases. That is why the operating lease balance increased. Sean, I don’t know if you want to add any more accounting detail.

Which helps EBITDA, right?

Which does help EBITDA, and that was clearly disclosed last year in the first quarter. The operating lease balance increased because some of those future obligations now need to be classified as operating leases.

I think that was clear. I’m just trying to get the right net debt leverage, whether we include it. Like if you’re using the 3.3x, I guess we have to put those expenses back in if we’re excluding operating leases. In any event, your leverage is up year-over-year. Do you think this is kind of the max on leverage? If you ignore the repurchase in the prior quarters, do you think we’ve reached the peak in terms of your leverage, however you’re calculating the numerator and denominator?

Yes, there is no P&L impact of this accounting change. This is purely a balance sheet gross-up. So there is no impact on our ratios.

And we expect over the next couple of quarters, given the growth in cash flow and given a modest reduction in CapEx, that our debt-to-EBITDA levels are going to come down marginally quarter-on-quarter?

Yes. And finally as I’d mentioned earlier in the call, we anticipate with this proposed refinancing and the effective constant level of debt when going back prior to our repurchase of the $115.9 million of debt, that we will end up saving about $10 million a year in cash interest expense.

Okay. Thanks Dave, thanks Sean.

Thanks.

Your next question is from the line of Tim Horan with Oppenheimer.

Thanks guys. I might have missed it, could you give the corporate customer connection total this quarter and last if you have it?

Corporate connections, is that what you’re asking? We had 46,719 at the end of this quarter, and we had 47,175 at the end of last quarter.

So a decrease of 1%.

Okay, down, got it. That churn was due to the off-net churn over the last couple of quarters, is that why it declined?

That is correct. The decline was in smaller connections. While we can have unit churn, because of the ARPU uplift from larger connections to bigger connections we can actually see revenue growth improve. We also get the benefit of the mix shift with more connections and bits being on-net than off-net. Off-net are those primarily secondary offices, so we also get that benefit to our cash flow.

Got it. So the corporate ARPU would have been down slightly year-over-year, just so we’re on the same page, right?

That is correct.

I guess how is that climbing if the mix is shifting to gigabit from 100 megs? I would expect that number to have increased?

You have two things going on. On the on-net side you are seeing an increase in ARPU. On the off-net side, it is declining very rapidly.

Because of all our on-net wins.

Indeed, that’s part of the strategy. Our ARPU on off-net has declined sort of 8% to 10% per year, much faster than our changes in our on-net connections.

Got it. And that off-net ARPU of about $1,700 for a one-gigabit connection — that seems a little high. Who are you competing with and what’s the market price point out there?

Most of our off-net business is sold in conjunction with on-net. We develop the relationship with the on-net connection, which is most likely a one-gig connection in the $500 to $700 price range depending on contract term. The customer then asks for a similar service in an off-net location. We look at our database of nearly four million fiber-served buildings from 90 different vendors, get the best loop price in an automated tool, and then price to get a 50% gross margin product to offer that customer. Our win rates are much lower in off-net. We compete with inferior products delivered over TDM or over copper, and we are selling an unbundled pipe. The off-net business is not a dedicated, viable business by itself, but rather an adjunct to our on-net relationship, and that’s why many customers buy from us in many locations.

It’s just simplicity. If you’re a company reconfiguring and you have multiple locations, it’s much easier to use one vendor, i.e., Cogent, versus having to use multiple vendors for different circuits.

Very helpful, thank you. Lastly Dave, do you expect the traffic growth overall to slow? The year-over-year comps could get difficult in the second half of this year.

It’s a bit hard to answer because it is usage-based. While we are returning to work in the U.S., much of the rest of the world is behind. The U.K., Singapore, Korea, Israel are markets where we’re seeing good re-openings, but there are still shutdowns and people required to stay in place in many areas. I think we’ll continue to see elevated traffic growth rates for the foreseeable future. There is also the long-term phenomenon of reduction in linear packages and adoption of multiple streaming packages. I think we’re still at the early stages of people figuring out how many concurrent packages they will have. The broadening of the streaming supply chain helps in two ways: it helps us charge more for a stream that’s being delivered, and it means receivers who are also paying us are probably receiving more bits for more hours of the day.

Your next question is from the line of Nick Del Deo with MoffettNathanson.

First, on the NetCentric front, the improvement there has been pretty notable. The drivers you call out all make sense, but they’ve been chugging along for a while, so it’s hard to tie to the change in growth. OTT adoption has been growing for some time; you’ve been pushing overseas for some time; the customer base has been broadening for some time. The pace of traffic growth this quarter even ticked down a little bit. Is there anything else that’s different now that’s having such a dramatic impact on the growth and drove such an inflection?

Okay. First of all, thank you for participating. I think these are continuations of long-term trends. The NetCentric business tends to be a bit lumpy and these trends take several quarters to fully manifest. You are right that the sequential traffic growth did tick down slightly as it normally would this time of year. But effectively we have been able to get a higher revenue per bit. You see that two ways: one, the average price of new bits sold actually slightly went up, which reflects a broader buyer universe; two, we were helped because more of the access network growth is coming outside of the United States where we are stronger. In the U.S., the three largest access networks are not Cogent transit customers in the U.S., but they do buy transit from us in international markets. That is not true of some of our European and Asian peers; they buy from us globally. So this internationalization phenomenon has a long way to go, and it is also impacted by COVID. It has been worse in continental Europe than in most of the U.S. Most of the U.S. appears to be reopening.

Nick, I would add that there is clearly more growth overseas. HBO Max, for example, which had a slow start is now beginning to pick up steam as it expands markets. Outside the United States, Cogent is often a more natural vendor given our presence in many countries and our connections to many CNDCs and access networks. As streaming goes international, we have benefited from an outsized portion versus competitors.

Okay, and one last on the corporate segment. Thinking about the mechanics of how a business goes from a potential tenant in a building to a Cogent customer, how should we think about the time it typically takes from when a business starts looking for space and signs a lease and takes the space and moves in and signs up for a DIA?

For a new tenant, that’s the best news. There is a forcing event. Typically a company will look anywhere from one month if they are moving into a spec suite that’s pre-built to as much as a year in advance if there’s an extensive build-out. Because many new leases tend to be for shorter duration and landlords are more willing to build spec suites, the window from decision to actual move-in on average is shortening. For existing tenants who are evaluating bringing employees back and are not currently Cogent customers, there are three key factors: one, some portion of the workforce will continue to work from home, requiring more bandwidth in and out; two, applications are increasingly off-prem, requiring re-evaluation of Internet access; and three, the transition from MPLS to Internet-based VPN architectures, which was happening before the pandemic, is now accelerating as companies reconfigure networks. That presents a strong opportunity for us to win business.

Your next question is from the line of James Breen with William Blair.

Thanks for taking the question. Dave, have you had any discussions on the landlord side about companies starting to downsize their real estate footprint and how that could be an opportunity for you where average tenants per building starts to go up as space consolidates? You’ve said the best opportunity for selling is when changes happen. Just give your thoughts on that. Thanks.

Yes, we have reached out to our roughly 350 institutional owners of multi-tenant office buildings. What we’ve heard from landlords are three things: one, an increased need for high-quality bandwidth as a tenant amenity to attract people to their space. Some landlords have approached us to provide service in buildings with different commercial arrangements, including covering some construction costs. Second, landlords are increasing the amount of spec space designed for smaller tenants to attract them, and landlords are spending to build these specs hoping tenants come quickly. Third, there is a decrease in the proportion of buildings used for shared fancy amenities; landlords are focusing more on leasable space. All of these trends increase our total addressable market and are positive for Cogent.

Your next question is from the line of Brennan Nispel with KeyBanc.

Hey guys, it’s Evan Young on for Brandon. You’re showing a growing trend in international NetCentric demand. Where do you see the mix between domestic and international revenue settling in the next few years? Does more international demand affect the average price per bit? Thanks.

There are two different trends. First, content production remains primarily an American phenomenon: virtually all the major streaming services are U.S.-based. They may purchase both domestic and international capacity from us as they expand. Second, adoption of streaming in less developed markets and in Europe lags the U.S., which is positive for growth. Over time international will represent a larger percentage of our NetCentric business, but it’s difficult to predict the exact mix three to four years out. The U.S. has declined from 85% of global internet traffic to 33% over 20 years; that will continue to decline over decades as network buildout and teledensity increase globally. Government-driven fiber builds in the U.K., France, Germany and other places are accelerating international fiber availability, which supports our growth.

Your next question is from the line of Frank Louthan with Raymond James.

Alright Dave, are you concerned about some of the fixed wireless claims from larger carriers, particularly for the off-net business or backup-type services? It seems like some of that is targeted at suburban areas where some of your off-net business generally lives. Just give your thoughts on that.

Anything that brings more traffic to the Internet is positive for Cogent. More access technologies, whether fixed wireless or fiber-to-the-premises, domestically or internationally, bodes well. Our off-net corporate business is an adjunct to our on-net business. There is no fixed wireless service that is on a per-unit basis anywhere nearly as reliable or cost-effective as our on-net services. We have refused to resell fixed wireless access because of the high cost per bit. Fixed wireless will work in white-space areas where fiber cannot be cost-justified, but the primary global effort is to close the digital divide with fiber, which is a necessary infrastructure for society.

And just quickly, what is your outlook for sales headcount growth for the year? Where do you expect to end up relative to where you did at the beginning of the year?

Our stated goal remains to grow the total sales force at 7% to 10% per year. We are currently up 1% year-over-year, so we are lagging. Part of that is the challenges of managing a remote sales force during the pandemic and our disciplined approach to managing underperformers. We have implemented systems to monitor rep productivity and that’s why sales force productivity is increasing. We hope for a return to our offices, at least in North America and the U.K., by the end of summer, which should allow us to reaccelerate hiring and get back on track to the 7% to 10% growth target.

Your next question is from the line of Angela Zhao with Bank of America.

Hi, thank you for taking my question. This is Angela on for Mike Funk at BofA. Earlier in the call you talked about the new sales force being focused on NetCentric margins. Could you expand on that and talk about the overall margin progression for the remainder of 2021? Also, can you talk about the projected growth in enterprise connections going forward from here? Thanks.

When we add $1 of on-net revenue, it carries a large gross margin contribution. In our corporate business, roughly 60% of sales are on-net, 40% are off-net. In NetCentric, it’s over 90% on-net. So the increase in NetCentric growth bodes well for gross margin and EBITDA margin expansion. We have guided to a 200 basis point per year multiyear expansion in EBITDA margin, with roughly half coming from gross margin expansion and half coming from SG&A leverage. We achieved this historically but most of the benefit was on gross margin; increased costs of operating in the pandemic offset some SG&A improvement. As we return to the office, those factors should normalize and we expect to continue about 200 basis points per year until EBITDA margins plateau around 50% based on the size of our addressable market. On enterprise connections, we expect to continue to increase penetration in on-net buildings and to return to an average of 1.5 connections sold per customer, with half of the customers taking a second connection either for backup or for VPN. Not all secondary offices will be shuttered, so we think off-net unit volume growth should mirror on-net long-term growth, around 11%–12% which has been a long-term average. Additionally, input costs for loops off-net have been coming down; we’ve seen a 9%–10% annual cost reduction in those loops and expect that to continue as cable and telco overbuilds put pressure on suburban infrastructure costs.

Your next question is from the line of Bora Lee with RBC Capital Markets.

You touched on aspects of this during the call. I’m trying to understand the path that the traffic takes by region. Can you talk about the network patterns you saw in the U.S. as it relates to COVID shutdowns and reopening and where Europe is on that curve, and provide any quantification you can?

Sure. A couple of points. First, 90% of our traffic is from our NetCentric business. On the corporate side, when people work from home the required bandwidth in and out of a business becomes more symmetric, whereas in normal times traffic can be more asymmetric. In the hybrid corporate model, symmetry is increasingly important. Regarding content to residential consumers: content is distributed in data centers; we have over 1,300 data centers which gives us a competitive advantage. Content wants to be closer to customers to reduce latency and needs high utilization to justify space and power costs. We then need direct connectivity to as many access networks as possible; we have that. In North America, more of our exiting traffic goes to our peers rather than our customers. We have large cable operators as customers in some markets and thousands of small regional players. Internationally, we have many PTTs, cable operators and competitive carriers buying from us, so we have a higher percentage of traffic where we’re getting paid on both sides internationally versus the U.S., and this internationalization has improved our effective price per megabit.

Great. And in terms of sales reps, you said tenure is an indicator of how efficient the sales force is. Can you provide an updated number given the changes in headcount?

Our sales force tenure has increased slightly during the pandemic, from about 28 months to around 30 months, which is where productivity peaks. Most sales force turnover occurs in the first six months of a rep’s tenure. The first month is for onboarding and training, and the next five months are when they are expected to hit KPIs to build funnels. That’s where we saw struggles during the pandemic: training reps to build adequate funnels remotely. Those reps that cannot build funnels in the first five or six months are typically managed out.

Okay, so it sounds like your sales force is reasonably well positioned to take advantage when things reopen?

I believe so. We can always improve, but we are in a position to capture market share as people come back to the office.

One last question: it looks like you launched your local peer connect service in the third quarter. Can you provide color on the service, what you’re hoping to achieve and customer interest since the launch?

We believe transit is the best way for smaller networks to connect to the public Internet. However, about 10% of the market prefers single-location exchanges. We built a global exchange where there is no distant sensitivity and no fixed port commitment. It has been well received: we have sold hundreds of ports to hundreds of networks. It does require networks entering the exchange to agree to exchange traffic with one another, whereas transit makes that obligation the responsibility of your upstream provider. The vast majority of our traffic still goes through transit because it’s easier and cheaper and provides more scalability, but there is a segment of the market that wants direct control through an exchange, and our exchange has seen very encouraging early adoption.

At this time there are no further questions. I will turn the call back over to the host for any closing remarks.

I’d like to thank everyone. I know these calls go long. We did work on trying to shorten our script, but we did have the capital markets activity that we needed to touch on. Again, I want to thank everyone. Please stay in touch and thank you for your support. Stay safe and well. Bye-bye.

Ladies and gentlemen, this concludes today’s conference call. Thank you for your participation and have a wonderful day. You may all disconnect.