All right. Good morning. My name is Frank Louthan. I'm Senior Wireline Analyst here at Raymond James, and great to be back in person. Glad everybody's here. Today, once again, we have Dave Schaeffer, Founder and CEO of Cogent Communications, here with us, once again at the conference. We'll go through a few questions, and then if you all folks have some in the audience, feel free to jump out. Dave, why don't we start kind of with the local news here. You've taken on sovereign governments in the past, but you over the weekend came out and you have made some changes and pulled some of the network from Russia over the conflict.
Talk to us a little bit about, you know, what exactly you're doing to, you know, with your Internet access into Russia?
Sure. First of all, I'd like to thank all the investors for their time, thank Raymond James for welcoming us back to a great venue, and Frank for his efforts in covering the company. Cogent is a provider of global Internet access. We have two market segments. Roughly 60% of our revenues come from selling to corporate customers unimpacted by events in Ukraine, for the most part. 40% of our revenues, but 97% of our traffic comes from selling bulk Internet connectivity in about 1,415 carrier neutral data centers in 50 countries around the world. We sell to over 100 incumbent telecom providers, the PTTs of those countries. We have physical network in the Ukraine that actually goes through Kyiv all the way to Kharkiv. We have salespeople in the country.
We serve also a number of major Russian service providers from Rostelecom, the incumbent, MegaFon, the number two provider. Most of the mobile operators and cable operators buy connectivity from us. We made the decision last week to terminate services to those customers based in Russia. Our belief was that the network that we operate could allow those companies or the government of Russia to use our facilities for cyber warfare. Cogent carries about 25% of the world's Internet traffic. These are very large connections and could be used in an offensive manner. They can also be used to disseminate propaganda or disinformation. Now, it was a difficult decision because we are concerned about an open Internet and freedom for content to flow uncensored to end users.
We felt that many of the other smaller providers who covered the market would still allow those Russian citizens to get access to flat files, text messages, lower bandwidth forms of content through other providers. Almost no, incumbent operator or major access network uses just one provider. It was a difficult decision, but we think it was the correct one.
What percentage of the Internet in Russia do you think is affected by your decision?
By independent studies, we are the second largest carrier in Russia. The number one carrier is the former incumbent from Sweden, Telia, which has spun off its network as a standalone backbone operator. We are number two in that market. You know, difficult to say by percentage, Frank, but you know, my guess is probably 25%-30%.
Okay. What is your revenue exposure here to that you're affecting?
It's fairly minimal. First of all, the service that we sell on a wholesale basis is a metered service, so we get paid for the service as the customer uses it. Secondly, the average price per megabit for the past 22 years since Cogent has been in business has fallen at a compounded rate of 30% per year. Roughly 25% of our revenues originate outside of the U.S., and the Ukraine and Russia together represent only a couple % of that 25%. Our revenue exposure is far sub 1%.
Got it. Okay. All right. Moving on to the rest of your business here. What are some of the major hurdles, you know, looking at the corporate business, you know, return to office, kind of keeps on again, off again, and so forth, and some of the, you know, the COVID pandemic made it difficult for your sales process. How is that rebounding?
The pandemic has taken longer and been far more challenging than I would have expected two years ago when we initially sent our workforce home and we saw many of our customers shutter their offices. Cogent's ISP business, where we sell directly to end users, is focused on the central business districts of major North American cities. We have roughly 1,820 skyscrapers connected to our network, 990 million sq ft. Just under 1 billion sq ft. The average building's about 41 stories tall, has 51 businesses in it and is about 550,000 sq ft. Our sales value proposition is to go to those buildings and provide a better internet connection.
We have a physical network that encompasses 61,000 mi of inner city fiber, another 17,000 mi of metropolitan fiber configured in 216 markets and about 1,030 physical rings. From those rings, we connect into those large buildings, put equipment in the basement, and build an infrastructure up the riser to sell internet service. If those businesses are not in those buildings, it is challenging for us. At the beginning of the pandemic, we actually saw a surge in that business because many of our customers used their firewalls to concentrate ad hoc virtual private networks for their employees when they work from home. That was a boom.
We saw many of our customers upgrade those connections from 100 Mbps, which was Cogent standard product, to a newer product, which has now become our de facto standard, which is a 1 Gbps connection. However, when those employees were sent home, businesses ceased to operate remote offices and also ceased to think about upgrading their networks. Cogent sells two primary products to its corporate customers. We sell Internet access, either at that primary location that I've described or a branch office, and we also sell a VPN service between those locations. The decision process for those sales has greatly slowed down. It has meant that our corporate business, which for 62 quarters as a public company, had grown for 53 of those quarters.
Prior to the pandemic, we only had two negative quarters of corporate growth, Q4 of 2008, Q1 of 2009 in the depth of the financial crisis. Fast-forward to 2020, we've now had seven sequential consecutive quarters of negative growth. In talking to IT departments, they feel a bit uncertain about what their management intends. They were told to come back to the office and prepare their networks last summer, then they saw that suspended. It was gonna be last fall, then it was gonna be January. I think this time it does feel different, and we are seeing companies starting to repopulate those offices. It is actually a non-homogeneous answer in that if you look at the South and Southwest, I think most businesses are back to operating close to normal.
Whereas if you look at the Northeast or Northwest, I think many businesses are still only in the office one or two days a week, and maybe with just a fraction of their employee base.
Okay, great. Talk to us a little about the sales hiring that you pursued. You know, what's changed? You laid out some goals for hitting, you know, some sales hiring goals for the year. What's kinda changed, given the confidence that you'll be able to pursue that and hit those numbers?
Cogent derives 99% of its revenues through a direct sales force. Different to most telecom companies who rely heavily on agents or third parties to sell, we have an outbound telesales model. Prior to the pandemic, we would churn about 5.4% of our sales force per month. That sounds very high, but anyone who's ever worked in a telemarketing job knows how difficult that job really is. When the pandemic hit, we needed to make sure that our workforce was safe. We quickly pivoted 70 sales teams that sit in 40 offices around the world, from Singapore to Stockholm, from New York to Toronto to Los Angeles to Dallas or, you know, Chicago, to work from home. We then needed to come up with a mechanism to hire new people remotely and train them. We modified our training.
It all went online, and we continued to hire vigorously during the pandemic. However, we saw our sales force turnover rate actually increase during the pandemic to over 7%. That increased turnover has resulted in the sales force shrinking from about 600 to about 500 quota-bearing reps. We continued to hire, but we realized that we needed to get the new hires back into an office so they could get the mentoring and training that they needed. We began that effort in October. We had to suspend that in December because of the Omicron variant. On Tuesday, March 1st, we returned all employees in North America back to the office.
We think that's gonna result in a reduction in turnover and an increase in sales productivity as we can actually train those people in place, and as a result of that, our revenue growth should begin to accelerate.
Walk us through how you changed sort of the training and hiring process during the pandemic and what of that's gonna be more permanent?
Pre-pandemic, we had two training centers, one in Washington, D.C., one in Pasadena, California. When a new rep was hired, they actually went to that facility for a two-week class. They received in-person instruction and testing. They then supplemented that for the next six weeks with additional online modules. They had a learning manager as well as their normal sales manager to support them, as well as their colleagues. When the pandemic hit, all of that training material went online. The same process, but you're now doing it from your kitchen table, and there's no one to interact with face-to-face real time. You can't look over at the cube next to you and see how that Leaders Club salesperson was closing business. That isolation and the inability to have face-to-face training just meant that reps were much less likely to succeed.
We actually had our greatest hiring year in the company's history in 2021. We hired 450 new sales reps in the year. That's actually up 20% from pre-pandemic levels. There's no shortage of candidates, but there's a shortage of ability for them to be successful without having those learning managers. We have 13 learning managers who travel the world to sit with teams and provide coaching and counseling in addition to standard managers. We also have two full-time management trainers, and I think what we're going to do post-pandemic is have a more hybrid model where we will be doing some in-person training supplemented by those online classes that we developed.
All right, great. Let's switch gears to the NetCentric business a little bit. You know, what do you think you can capture incrementally over the next couple of years as far as traffic there versus what we've seen in the last two years during the pandemic, with the shifts in the traffic? If we fast-forward a couple of years, what will the traffic profile look like?
While our corporate business has underperformed our NetCentric, our wholesale business has delivered the best consistent performance throughout the pandemic in the company's history. In fact, last quarter was the best year-over-year quarter at 25% revenue growth in the company's history. With people at home more, they are streaming more. They are using more Zoom. They are consuming more streaming video services. Video consumption in the developed world is slightly over 300 minutes a day. Going into the pandemic, roughly 18% of that was done via streaming. Today, about 45% of all video consumption is streaming. We've definitely seen an acceleration in a trend that was already in place. We've also benefited from two other factors. One, the streaming services themselves have broadened out. Rather than being just one large streaming provider, there are now a half a dozen.
That gives us more pricing power. Two, we have seen the traffic become much more internationalized. Cogent is particularly strong outside of the U.S., and it is more likely outside of the U.S. that we will actually get paid on both sides, meaning we're paid by the sender of the content and the receiver. Let me give you an example. If you're watching an Amazon Prime movie, Cogent's paid by Amazon to carry that movie. If you are a customer in Vietnam watching that movie, Telecom Vietnam buys its connectivity from Cogent. We're getting paid on both sides in approximately 72% of the instances. As a result, our revenue decline had been driven by lower prices. Those prices are still declining at about 23% per year, but the effective price per megabit is actually going up.
Cogent's aggregate NetCentric revenues today are growing at about 25% year-over-year. We carry about 25% of the world's traffic, and I think over the next several years, Frank, you should expect to see the NetCentric growth revert back to long-term average growth rates of about 9% per year, with traffic growing in the mid-20s% and prices declining in the low 20s%.
One of the shifts that had happened in that NetCentric business was a shift to longer term contracts, and so forth, in exchange for lower pricing. Have we largely gotten through that adjustment in that base that's helping that sort of start to flatten out and start to see that grow again? Is that another factor that's coming into play?
There are three factors that lower the price per megabit. One is, the more you consume, the lower the price. Two, the longer you commit to, the lower the price. The third is, the more you commit and the less that is on a variable basis, the lower your price. We have seen our effective price per megabit stabilize and actually go up, in part because our average contract length today is about three years. That's typically the longest contract a customer is comfortable in signing in an industry where, you know, unit deflation is about 23% a year.
Okay, great. Let's talk a little about the off-net business, you know, that suffered a bit in the pandemic. What are you seeing in the off-net business? How stable do you think that is? And then what is your sales force doing to make sure you target the right kind of off-net business so that you don't see higher churn in those circuits?
Off-net as a standalone business is not a viable business. The reason for that is the value proposition to the customer is not compelling enough. We're buying the local loop from someone else. Our cost of revenue acquisition would be too high relative to the gross margin achieved over the customer's life cycle. At Cogent, we start our relationship with an on-net corporate relationship. Our off-net business is almost exclusively corporate to serve secondary offices. We have relationships with 90 different off-net providers that can allow us fiber connectivity to slightly over 4 million buildings, representing about 45% of the square footage in the market. Roughly another just under 5 billion sq ft. We double the price of the loop that we buy.
We have concluded that it is not economically viable for Cogent to serve that building because it's too small or too far from fiber. The average off-net building we serve is only about 19,000 sq ft compared to the on-net buildings that we serve that are 550,000 feet. Those buildings have four businesses, just under four, compared to the 51. We get better pricing each year by working the competitive market between cable and telco providers and their fiber over builds. Our corporate business is 60% on-net, 40% off-net. Our total business is 75% on, 25% off. When we add a dollar of on-net revenue, it carries a 100% gross margin contribution and $0.95 of EBITDA. When we add a dollar of off-net, it's a 45% EBITDA margin contribution.
That business will grow on average in line with our corporate business. Our corporate business over the past 15.5 years has averaged a 10% top line growth rate.
All right, great. Walk us through kind of the competitive landscape of you know getting access to buildings where you provide the connectivity, and how do you think you can keep taking share from competitors?
The average building that we connect to has three providers: ourself, the incumbent, and one other alternate provider. We are typically the only provider that has pre-wired the building. We guarantee the customer that we can turn them up in 17 days. We actually do that in an average of nine days. Once we turn that customer up, their service is 3x more reliable on Cogent than our competitors. The customer gets between 30x and 60 x the actual throughput on a non-oversubscribed, non-blocked, non-metered, and protected service. Our competitors are offering an inferior product. Two very different value propositions in our two customer bases. For our wholesale customers, we go to a data center that's effectively a supermarket for bandwidth and undercut any other provider by 50% to gain market share.
In our corporate business, we price at parity to those two other competitors, but deliver a vastly superior product. Today, we have about 15% of those businesses as customers, and we believe that as businesses move more applications to the cloud, have more employees working in a hybrid work environment, the need for additional bandwidth will increase and increase the likelihood Cogent will win that business. The final point is, in the post-pandemic world, we're seeing a lot of corporations add a second point of virtual network concentration in a data center. This is actually a brand-new product for Cogent that we could sell, and an incremental opportunity for our corporate customers.
All right, great. See if we have any questions from anybody in the audience.
Yep. One question. What did Cass say was the incremental opportunity for corporate?
What's the incremental opportunity in the corporate business?
We think we will eventually get to about 50% market share based on the value proposition. We think our corporate business should be able to grow about 2.5 x its current size. Today it's, you know, roughly a $350 million business. That should end up growing to about $800 million in total scale.
Right. Any other questions? Yep.
How do you get to that 50%? Is that the % of new wins on new R2?
That's by the number of businesses located in that building that we're selling services to. Literally looking at, you know, whether it be leasing data, CoStar or LoopNet data, just literally going into the building and looking at the names on the marquee in the lobby, and we're selling to about 15% of them. We're actually selling about three circuits per customer, so we're typically selling them an internet connection, a VPN connection, and in some cases, a backup connection.
All right, let's talk about capital allocation, Dave. What are your leverage goals and can we continue to see the dividend increase at the same rate? At what point can investors think buybacks might be back in the mix?
We have returned over $1 billion to shareholders in the past 15 years. We have bought back 10.3 million shares, which represents about 22% of our float. We have a dividend and have grown that dividend now sequentially for 38 consecutive quarters. The rate of growth in that dividend is roughly 13% year- over- year. Our dividend is very tax friendly. 79% of the dividend is treated as a return of capital, and therefore tax deferred. 21% is treated as ordinary dividends. We have excess cash on the balance sheet, about $350 million. We need less than $100 million to run the business, so we have a goal of slowly disgorging that excess cash. That's a combination of methodically growing the dividend coupled with opportunistically doing buybacks.
If we saw a massive market correction, kind of à la 2009 type of volatility, we would actually look to potentially even suspend a dividend and use buybacks exclusively as a mechanism of return of capital. In an environment when, you know, the aggregate markets are at near all-time highs, you know, we feel that the dividend is a more effective way to return capital. We're somewhat agnostic to which method. What we are committed to is growing cash flow and returning.
On the balance sheet, what sort of leverage targets do you have and where do you think how that plays out in the next 12 months?
Yeah. Our leverage goals are for net leverage between 2.5x-3.5 x EBITDA. We're actually slightly above that at 3.58x right now. The board is comfortable with that as a temporary measure because of the slowdown in the corporate business due to the pandemic. As we look at all of the leading indicators, whether it be key swipes in offices, leasing activity, Cogent sales proposals being issued or actually sales funnel, it does appear that we've reached the bottom of the corporate trough, and now we're going to see consistent improvement in that business. It may not go back to positive growth in Q1, but over the next few quarters it should get back to consistent positive sequential growth.
You've got a fair amount of insight into the office world. You own some real estate and what you see from your business. What's your crystal ball say as far as kinda what the percentage of that office space and return to work kinda comes back to you relative to two years ago?
I think the pandemic accelerated a trend that was probably in place pre-pandemic, and that is the migration to more flexibility in work and a shorter work week. I think at the end of this, whether it's five years or 10 years from now, America will transition to a four day work week. But people will still need offices to congregate, to build a corporate culture, and I think offices remain a critical real estate class. I think when we look at things like hotels, we're probably over-hoteled in the U.S., we're definitely over-retailed. I think for office, in most other recessions, what we have seen is the oversupply of office was consumed through new business formation.
I think what is going to be different this time is probably close to 10% of that office inventory is coming out of the system being converted to residential. I think office occupancy rates are gonna go up in large part because there's just gonna be less office space available.
All right. Great. Dave, thank you very much. Really appreciate it. We've got a breakout session after this if anybody's interested, so thanks again.
Sure. Thanks, Frank.