Cogent Communications Holdings Inc (CCOI) 2018 Q2 法說會逐字稿

Good morning, and welcome to Cogent Communications Holdings Second Quarter 2018 Earnings Conference Call.

As a reminder, this conference call is being recorded, and it will be available for replay at www.cogentco.com.

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 second quarter 2018 earnings conference call.

I'm Dave Schaeffer, Cogent's Chief Executive Officer; and with me on this morning's call is Tad Weed, our Chief Financial Officer.

We're pleased by our margin results for the quarter and continue to be optimistic in the underlying strength of our business and the outlook for 2018 and beyond.

Our EBITDA margin for the quarter increased by 200 basis points to 35.5% from the second quarter of 2017, representing the highest EBITDA margin percentage in our 19-year history.

Our EBITDA for the quarter increased year-over-year by $5.8 million, which is an increase of 14.5% from Q2 of 2017.

Our gross margin for the quarter continued its expansion from the last quarter and increased sequentially by 60 basis points to 58.1% and an increase by 70 basis points from Q2 of 2017.

On a constant currency basis, we achieved sequential quarterly revenue growth of 1.1% and year-over-year quarterly revenue growth of 6.3%.

Our quarterly sales productivity was 5.7 units per full-time equivalent rep per month, a productivity rate that again was significantly above our long-term average of 5.1 units per full-time equivalent rep per month.

For the quarter, traffic on our network grew at an accelerated rate, and we achieved sequential quarterly traffic growth of 10%, and year-over-year, our traffic growth accelerated to 44%.

During the quarter, we returned approximately $23.8 million to our shareholders through our regular quarterly dividend.

At quarter end, we had a total of $41.5 million available in our stock buyback authorization program, which the board has authorized to continue through the end of 2018.

We did not purchase any stock in the quarter.

Our gross leverage improved from -- to 4.22x EBITDA from 4.33 in the previous quarter, and our net leverage ratio improved to 2.93x EBITDA from 2.94 last quarter.

Our consolidated leverage ratio, as defined under our indenture, improved to 4.16 from 4.26 last quarter.

This ratio is below the 4.25 threshold that is included in our indentures and will allow us to utilize the accumulated builder basket.

Our accumulated builder basket was $169 million at the end of the quarter at the operating company level, Cogent Group, which, when combined with the $17 million of cash held at Cogent Holdings, provides us a total of $186 million of cash that is unrestricted and available for dividends and buybacks.

We expect to transfer the majority of this builder basket from our operating company to our holding company during the quarter.

We continue to remain confident in the growth potential of our business and the cash generating capabilities of our business.

As a result and as indicated in our press release, we've announced yet another sequential $0.02 increase on our regular quarterly dividend, increasing that dividend from $0.52 a share per quarter to $0.54 per share per quarter.

This represents our 24th consecutive quarter in which we have sequentially grown our regular quarterly dividend.

Throughout this discussion, we'll highlight a number of operational statistics, I will review in greater detail certain operational highlights, and Tad will provide some additional details on our financial performance.

And we will then open the floor for questions and answers.

Now I'd like to turn it back over to Tad to read our safe harbor language.

Thank you, Dave, and 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.

Please refer to our SEC filings for more information on the factors that could cause actual results to differ.

Cogent undertakes no obligation to update or revise forward-looking statements.

If we use any non-GAAP financial measures during this call, you will find these reconciled to the GAAP measurement in our earnings release, which is posted on our website at cogentco.com.

Now I'd like to turn the call back over to Dave.

Thanks, Tad.

Hopefully, you've had a chance to review our earnings press release.

Our press release includes a number of historical metrics.

Now with regard to some of our long-term guidance.

Our targeted guidance for long-term full year growth remains in a range of 10% to 20%.

Our long-term EBITDA annual margin expansion targeted guidance is an annual improvement of approximately 200 basis points.

Based on our results for the first half of the year, we expect our full year revenue growth to be slightly below that targeted range.

However, we do expect our annual EBITDA margin expansion to improve by more than 200 basis points for the full year.

Our revenue and EBITDA guidances are intended to be long-term goals and are not intended to be used as specific quarterly guidance.

Now I'd like to have Tad give you some additional details on our operational results for the quarter.

Thanks, Dave, and, again, good morning to everyone.

I'd also like to thank and congratulate our entire Cogent team for their results for the quarter and continued hard work and efforts during another very busy and productive quarter for Cogent.

Some comments on corporate and net-centric revenue and customer connections.

As a reminder, we analyze our revenues based upon network type, which is on-net, off-net and noncore.

And we also analyze our revenues based upon customer type, and we classify all of our customers into 2 types, net-centric customers and corporate customers.

Our net-centric customers buy large amounts of bandwidth from us in carrier-neutral data centers, and our corporate customers buy bandwidth from us in large, multi-tenant office buildings.

Revenue and customer connections by customer type.

Revenue from our corporate customers grew sequentially by 2.7% to $83.3 million and grew year-over-year by 11.9%.

We had 42,673 corporate customer connections on our network at quarter end.

Quarterly revenue from our net-centric customers declined sequentially by 3.4% and grew year-over-year by 1.4%.

We had 33,520 net-centric customer connections on our network at quarter end, which declined from last quarter due to significant circuit grooming, consolidating multiple 10-Gb circuits to fewer 100-Gb circuits at the same location from some of our larger net-centric customers.

Our net-centric revenue growth does experience significantly more volatility than our corporate revenues

(technical difficulty)

impacted by the impact of foreign exchange and certain other seasonal factors.

Revenue and customer connections by network type.

Our on-net revenue was $93 million for the quarter, which was a sequential increase of 0.7% and a year-over-year increase of 8.7%.

Our on-net customer connections increased by 3.2% sequentially and increased by 14.1% year-over-year.

We ended the quarter with over 65,400 on-net customer connections on our network and our 2,599 total on-net multi-tenant office buildings and carrier-neutral data center buildings.

Off-net revenue was $36.1 million for the quarter, which was the same off-net revenue as last quarter and increased year-over-year by 6.3%.

Our off-net customer connections increased sequentially by 2.3% and by 12.3% year-over-year.

We ended the quarter serving over 10,400 off-net customer connections and over 6,500 off-net buildings, and these buildings are primarily in North America.

Some comments on pricing, pricing per megabit.

Consistent with historical trends, the average price per megabit of our installed base of customers decreased for the quarter.

However, the average price per megabit for our new customer contracts increased for the quarter primarily due to customer mix.

However, we do expect, on a long-term basis, that the price per megabit for new contracts will decline.

The average price per megabit for our installed base declined sequentially by 3.6% to $0.87 and declined by 25.4%, consistent with historical trends from the second quarter of last year.

The average price per megabit for our new customer contracts for the quarter did increase sequentially by 8.2% to $0.44 and declined by 18.5% from our new customer contracts that were sold in the second quarter of 2017.

If you look at the price per megabit for new customer contracts on a sequential 6-month basis, the average price per megabit for our new customer contracts for the first 6 months of 2018 decreased by 12.4% to $0.42 from $0.48 for our new customer contracts that were sold in the second half of 2017.

Some comments on ARPU.

Our on-net ARPU and our off-net ARPU both decreased sequentially for the quarter.

Our on-net ARPU, which includes both corporate and net-centric customers, was $482 for the quarter, which was a decrease of 2.5% from last quarter.

Our off-net ARPU, which is comprised predominantly of corporate customers, was $1,162 for the quarter, which was a decrease of 2.6% from last quarter.

Some comments on churn.

Our on-net and off-net unit churn rates were relatively stable during the quarter.

Our on-net unit churn rate was 1% for the quarter, which was the same rate as last quarter.

Our off-net churn rate was 1.2% for this quarter, a slight increase from 1.1% last quarter.

We offer discounts related to contract terms, all of our corporate and net-centric customers, and we also offer volume discounts to our net-centric customers.

During the quarter, net-centric customers took advantage of our volume and contract term discounts and entered into long-term contracts for over 2,200 customer connections, which increased their revenue commitment to Cogent by over $22.1 million.

Some comments on EBITDA and EBITDA as adjusted.

Our EBITDA and our EBITDA as adjusted are reconciled to our cash flow from operations in all of our press releases.

Seasonal factors that typically impact our SG&A expenses and, consequently, our EBITDA and EBITDA as adjusted include the resetting of payroll taxes in the United States at the beginning of each year, an annual cost of living or CPI increases, the timing and level of our audit and tax services, net neutrality fees, and the timing and amount of our gains on equipment transactions and also benefit plan annual cost increases.

Our SG&A expense decreased sequentially by $0.7 million or by 2.3%.

And this occurred despite a reduction in the benefit of the new U.S. GAAP requirement that we capitalize and amortize certain of our commissions as customer contract costs.

Previously, we had expensed these costs.

The SG&A impact on our commission expense was a benefit of $1 million last quarter and was a benefit of only $300,000 this quarter.

The new U.S. GAAP revenue pronouncement that we adopted in the first quarter, began this quarter had a relatively immaterial impact in our reported revenues.

The impact was a negative $40,000 last quarter and was a negative $100,000 this quarter.

Our quarterly EBITDA increased by $1.8 million or by 4.1% sequentially to $45.9 million and increased year-over-year by $5.8 million or by 14.5%.

Our quarterly EBITDA margin increased by 120 basis points sequentially to 35.5% and increased year-over-year by 200 basis points.

And as Dave mentioned, our EBITDA margin percentage of 35.5% this quarter represented the largest EBITDA margin percentage in the company's 19-year history.

Our EBITDA as adjusted, which is EBITDA and including gains on our equipment transactions, was relatively similar.

Our quarterly EBITDA as adjusted increased by $2.1 million or 4.7% sequentially to $46.3 million and increased year-over-year by $5.1 million or by 12.5%.

Our EBITDA as adjusted margin increased sequentially by 150 basis points to 35.8% and increased by 100 basis point -- 50 basis points year-over-year.

Our equipment gains have recently been declining and were only $0.4 million for the quarter compared to $1 million for the second quarter of last year and only $100,000 last quarter.

Some comments on earnings per share.

Our basic income per share was $0.15 for the quarter compared to $0.15 last quarter and $0.10 for the second quarter of last year.

Our diluted income per share was $0.14 for the quarter and $0.15 last quarter and $0.10 for the second quarter of last year.

Some comments on foreign exchange.

Our revenues reported in U.S. dollars and earned outside of the United States was about 23% of our total revenues.

About 17% of our revenues this quarter were based in Europe, and about 6% of our revenues were related to our Canadian, Mexican and Asian operations.

Continued volatility and foreign currency exchange rates can materially impact our quarterly revenue results and our overall financial results.

The foreign exchange impact on our reported quarterly sequential revenue was a negative $800,000, and the year-over-year foreign exchange impact on our reported quarterly revenue was a positive $1.9 million.

Our quarterly revenue growth rates on a constant currency basis were 1.1% sequentially and 6.3% year-over-year.

The average euro to U.S. dollar rate so far this quarter is $1.17, and the average Canadian dollar exchange rate is $0.76.

Should these average foreign exchange rates remain at the current average levels for the remainder of the third quarter of 2018, our estimate is that the foreign exchange conversion impact on our sequential quarterly revenues for the third quarter would be a negative $0.5 million and that the year-over-year FX conversion impact on quarterly revenues will also be negative and be $0.3 million.

Customer concentration.

We believe that our revenue and customer base is not highly concentrated.

Our top 25 customers represented less than 6% of our revenues this quarter.

On CapEx.

Our capital expenditures for the quarter declined sequentially by 19.6%.

Our capital expenditures were $12 million this quarter, the same amount as the second quarter of last year but a decline from the $14.9 million we incurred last quarter.

Capital lease and capital lease payments.

Our capital lease IRU obligations are for long-term dark fiber leases and typically have initial terms of 15 to 20 years and sometimes longer.

And often, these leases include multiple renewal options after the initial term.

Our capital lease IRU fiber obligations totaled $159.9 million at quarter end.

And at quarter end, we had IRU contracts with a total of 232 different dark fiber suppliers.

Our capital lease principal payments under these IRU agreements was $3.8 million for the quarter.

And if you combine capital lease principal payments with CapEx, there also was a decline sequentially and that total was $15.7 million this quarter compared to $7.2 million last quarter and $14.2 million for the second quarter of last year.

Comments on cash and operating cash flow.

At quarter end, our cash and cash equivalents totaled $224.3 million.

And for the quarter, our cash decreased by $11.7 million as we returned $29.1 million of capital to our stakeholders.

During the quarter, we paid $23.8 million for our regular quarterly dividend payment, and $5.3 million was spent on the semiannual interest payment on our debt.

Comments on debt ratios.

Our total gross debt, which includes capital lease IRU obligations, was $736.5 million at quarter end, and our net debt was $512.2 million.

Our total gross debt to trailing last 12 months EBITDA as adjusted ratio, as Dave mentioned, improved 4.22 at quarter end from 4.33 last quarter; and our net debt ratio improved to 2.93 from 2.94.

Our consolidated leverage ratio, which -- as defined by our indenture agreements, again was 4.16 and below the 4.25 ratio needed to meet the covenant test in our indentures to access our $169 million of accumulated builder basket.

Our consolidated secured leverage ratio, as defined by our indenture agreements, was 3.09 at quarter end.

Some comments on bad debt and day sales outstanding on accounts receivable.

Our bad debt expense was only 0.6% of our revenues for the quarter.

And our days sales outstanding, or DSO, for worldwide accounts receivable was again 23 days, the same DSO as we incurred last quarter.

And as always, I want to thank and recognize our worldwide billing and collections team for continuing to do a fantastic job on collections and customer service.

I would like to turn the call back over to Dave.

Thanks, again, Tad.

Now for a couple of comments on our network, scale, scope and expansion.

The size and scale of our network continues to grow.

We have over 927 million square feet of multi-tenant office space, on-net, in North America.

Our network consists of over 31,900 metro fiber miles and over 57,400 intercity route miles of fiber.

Cogent remains the most interconnected network in the world, where we are directly connected with over 6,360 networks.

Less than 30 of these networks are settlement-free peers.

The remaining, over 6,330 networks, are paying Cogent transit customers.

We are currently utilizing 27% of the lit capacity in our network.

We routinely augment capacity in sections of our network to maintain these low utilization rates.

For the quarter, we achieved sequential quarterly traffic growth of 10% in what is traditionally a slow seasonal period for traffic growth.

And we saw a significant improvement in our year-over-year quarterly traffic growth to over 44%.

We operate 52 Cogent-controlled data centers with 587,000 square feet of space, and we're operating those facilities at 32% utilization.

Our sales force turnover rate in the quarter was 4.8% per month, again better than our long-term average turnover rate of 5.7% per month and I think a testament to the training and retention programs that we have put in place.

We ended the quarter with 438 reps selling our services.

Cogent remains the low-cost provider of Internet access and transit services.

Our value proposition to our customers remains unparalleled in the industry.

Our business remains entirely focused on the Internet and IP connectivity and data colocation services.

Our services provide a necessary utility to our customers.

Beginning at the start of Q2 on April 1, we began selling our SD-WAN services.

We do not expect a material contribution from these services for the next several quarters.

We expect our annualized constant currency long-term revenue growth to be consistent with our annualized guidance of 10% to 20% and our long-term EBITDA margin expansion rates to remain approximately 200 basis points per year for the next several years.

Our Board of Directors has approved yet another increase on our regular quarterly dividend of $0.02 per share sequentially, bringing our quarterly dividend to $0.54 per share.

Our dividend increases demonstrate our continued optimism regarding the cash flow capabilities of our business as well as the reduction and leverage that Tad outlined.

We remain opportunistic about the timing of the purchase of our common stock at quarter end.

We do have $41.5 million remaining under our current authorization program, which remains in place through year-end.

We are committed to returning an increasing amount of capital to our shareholders on a regular basis, utilizing all effective mechanisms to return that capital.

With that, I'd like to now open the floor for questions.

(Operator Instructions) And our first question comes from Philip Cusick from JPMorgan.

Dave, can you update us on net-centric trends?

It seems like that was accelerating nicely earlier in the year but slowed a bit this quarter?

Yes.

So part of that slowdown was obviously foreign exchange, and Tad commented on that.

Secondly, last quarter, we had a very significant reduction in the price per megabit of new sales, steeper than it had been traditionally.

And we actually commented on that in the last earnings call, explaining that most of the growth in traffic and the acceleration in traffic growth has been coming from our largest customers, which do get the very lowest price points.

Because of those lower price points achieved throughout Q1, we saw the full effect of that in Q2.

And that was the reason for the sequential rate of decline, the combination of FX and the impact of those much lower prices sold in Q1.

Now in Q2, we actually saw a bit of a reversal of that trend, where most of the new contracts sold were actually to smaller customers.

For that reason, we actually exhibited a sequential increase in the price of new contracts sold of 8.5%.

This is not a long-term bottoming out of price declines and an increase in pricing but rather a mix shift.

We expect both new contracts and the installed base to decline at roughly the same rate over a protracted number of years, and that's roughly about 22% or 23%.

There could be some volatility.

We also are obviously encouraged by the increased rate of traffic growth.

Normally, Q1 to Q2 tends to be a relatively flat quarter.

This year, we experienced a 10% sequential growth rate albeit from large customers with low pricing.

And on a year-over-year basis, we saw a material improvement in our growth rate to 44% from the 37% last quarter.

So while there is short-term volatility in the 33% of our business that is net-centric, over the long run, we remain encouraged that, that business will return to its long-term average growth rate, which is slightly below that of the corporate business but is still approximately a 10% growing business or 2.2% sequentially on average.

And we think we will return to that growth rate.

Okay.

And one completely different.

You rarely disclose your sales force size, but can you remind us what that is as a mix of total employees?

And is that a ratio you see as ideal or target for the company?

Yes, Phil.

So our sales force was 438 quota-bearing reps.

The sales manager -- sales operation teams bring that total to about 575 individuals.

Roughly 42% of our employee base was quota-bearing.

48% of the entire base was in the sales organization.

We're about 930 employees in total.

We expect to grow the sales force at between 7% and 10% per year for the next several years, while we expect operational headcount growth to be slower at probably 2% to 3%.

So the mix will increasingly become more sales-centric.

Because of the efficiencies and running our business and the standardization of our products and the systems that we've deployed, we can sustain 44% traffic growth, 20% growth in unit number of connections, and do that with an increase in operational and overhead employees of only about 2% to 3% per year.

The sales force, however, is the engine that will drive accelerating revenue growth, and investing in that sales force has been and continues to be our major focus.

And our next question comes from Scott Goldman from Jefferies.

I guess one quick follow-up on that last one.

Given the commentary around traffic growth in this quarter and the fact that the pricing on new orders was up sequentially in 2Q for the reasons you just mentioned, I'd assume the phenomenon there is then the 3Q should maybe be outside growth on the upside there than before, perhaps settling down to more normalized levels in 4Q.

I just want to make sure that was the case.

And then 2 other quick ones, if I could.

You mentioned SD-WAN and obviously very early days there.

Just wondering what the client interest has been, whether the use cases are to cannibalize MPLS or sort of make this additive in their networks and why they're choosing Cogent for that service.

And then lastly, if you could just comment, a strong number of building adds in the quarter both from the multi-tenant office buildings as well as data centers.

Wondering what you're seeing there in terms of the growth of the buildings or if you've changed your criteria at all.

I know a lot of questions in there, but thanks, Dave.

Oh, that's okay.

Those were all good questions, Scott, and thanks for them.

So first of all, on growth, we do not give specific quarterly targets.

But the continued acceleration in traffic growth and the mix in new sales coming from smaller customers does lead us to believe that our sequential net-centric growth will improve in the next several quarters.

Remember, net-centric revenue is predominantly driven by unit volume.

And typically, sequentially, the summer months tend to be a bit slower, but we've seen a bit of a change in that trend in Q2 and also in the first part of Q3.

So we do expect an improvement in our both sequential and year-over-year net-centric growth going into third quarter and then continuing, I think, through the remainder of the year.

And you could see both the number of networks that we're attaching to as well as the -- just growth in traffic.

Many of the negative impacts of port congestion have now completely resided, and we're now in a position where customers feel comfortable enrolling new applications and services out and driving growth.

Now to the SD-WAN question, we have been marketing our services now for about 4 months.

We have gotten very strong customer interest, as I mentioned, against the base of nearly 80,000 customer connections.

I think it will not be material for the next several quarters, but I think the #1 feature that customers seem to focus on is the security capabilities of the SD-WAN service and some of the superiority over even MPLS, which is both more expensive and more difficult to administer.

So I see SD-WAN continuing to cannibalize at an accelerating rate that market.

To remind investors, Cogent does not have a MPLS base to cannibalize.

We do sell VPLS, and those services remain strong as well.

In fact, VPLS services, as a percentage of total revenues, slightly increased in the quarter from 16% to 17% of total revenues.

And customers deploying SD-WAN could either elect to do it over a dedicated Internet port or they can elect to do it over a VPLS port and get the benefit of both encryption and encapsulization for 2 layers of security.

And we've actually seen, I think, many of the customers that have focused on our SD-WAN service really prefer that additive layer of security and kind of a dual approach of encapsulization and encryption.

Now to the building adds question, and it's a two-part answer.

Let's start with the data centers.

We continue to see a huge influx of capital and new data centers being built around the world.

We operate in 43 countries and 199 markets, and both existing operators and new operators continue to add to their footprint.

We also have seen several situations where previously carrier-controlled data centers have been spun out and are now carrier-neutral because they're in the hands of new operators.

Former CenturyLink facilities, the pending AT&T facilities, some of the Verizon facilities all fit that model.

So we're really seeing 2 different things

(technical difficulty)

Thank you.

It's Dave again, and I apologize for that technical interruption.

I'm not 100% sure what happened.

Let me finish answering Scott's question.

With corporate buildings, the big uptick was actually our decision earlier this year to expand in New York from Manhattan, to expand our footprint in Brooklyn.

As that market has revitalized, we actually added, through the course of the year, about 15 buildings in that market.

That was a bit of an anomaly, but we do expect the pace at which we're adding buildings to remain at about 120 to 130 per year, consistent with our historical trends.

And our next question comes from Colby Synesael from Cowen.

Great.

It's actually Greg Williams sitting in for Colby.

Dave, you provided some nice color on the net-centric acceleration in the second half of the year with product mix, et cetera, but you're still going to miss your overall revenue long-term target.

So with the top line growth moderating on a constant currency basis, are you still thinking it makes sense to message the 10% to 20% growth rate?

And why?

And the second question I have is just on the off-net revenue.

We saw a notable slowdown in the second quarter in a row.

Is there anything to call out?

And should we assume this is the new normal going forward?

Okay.

Let me take the general guidance question.

The general guidance question is designed to be long-term guidance and again not specific to the quarter.

Our corporate business continues to increase as a percentage of our revenues.

And today, is roughly 2/3 of revenues and is growing at 11.9%, well within our range, and 2.7% sequentially.

The net-centric business can be extremely volatile, and it is 1/3 of our revenues.

We will probably be below our range in part because of foreign currency, in part even on a constant currency because of the slower growth in the net-centric business.

But we do think that business will return to its long-term average growth rate, as I mentioned in my answer to Phil's question, that is 2.2% sequentially or just under 10% on a year-over-year basis, getting the entire business into the 10% to 20% range.

To get to the higher end of that range, the net-centric business would have to have a period of outsized growth, which is actually done on a number of occasions as new applications come online.

So because this is not quarterly guidance, we feel that the guidance we are giving remains appropriate.

Now to your off-net question, we expect off-net and on-net corporate to grow at about the same rate.

Our off-net business is a derivative of on-net corporate sales.

That business is growing at about 12%.

In the last few quarters, we have been able to negotiate significantly lower tail prices for those off-net circuits.

We pass those savings on to our customers, and that is why the ARPU declines were slightly larger.

Our margins actually remain constant on that product and allowed our entire margins to expand.

But over the long run, we expect corporate on-net and corporate off-net to both grow at roughly about 12% year-over-year.

And our next question comes from Matthew Niknam from Deutsche Bank.

Just 2 if I could.

One, on the pacing of shareholder returns, I think with the leverage ratio -- the growth leverage ratio now dipping below 4.25, Dave, how should we think about the pacing of dividend growth and/or buybacks from here and how that may change?

And then secondly, just one clarification.

I think you'd alluded to 200 basis points or more of EBITDA margin expansion.

I'm wondering if that's reported EBITDA or if that's adjusted EBITDA inclusive of the equipment sales.

And what gives you the confidence in getting there as growth decelerates?

Sure.

So let me take the return of capital.

Clearly, we're pleased that we are delevering and continuing to increase our dividend concurrently.

We are now able to access the $169 million builder basket, which has been trapped at the operating company level.

So we will move that to holding.

That gives us a great deal more flexibility and $186 million of available capital to return as well as we have borrowing power both at the operating company and holding company level if we so choose to utilize it.

I think the board, every quarter, will continue to evaluate its dividend policy and decide whether or not the pacing of the increase is correct or needs to be accelerated or decelerated.

We view buybacks as a tool to help us capture market volatility.

And because of decreased aggregate market volatility and Cogent's decreased volatility relative to the rest of the market, we have been a little more reluctant to use buybacks.

Buybacks should always be used when the share price is below the intrinsic value of the business.

Cogent firmly believes that its share price is below its intrinsic value but we have not been able to capture the volatility.

I could be wrong, but I would suspect volatility will come back to the market and we'll be able to capture some of that to supplement our dividend policy.

But we are committed to maintaining a prudent leverage ratio and accelerating returns of capital as our cash flow growth continues.

Now to the margin expansion, I'll take the adjusted number first.

The answer is we'll probably be at or slightly below 200 there because we're not going to have the equipment sales at the same pace this year as we had last year.

And as we've again said to investors, that's somewhat episodic and volatile.

With regard to the underlying EBITDA not including equipment sales, the improvements that we saw, both sequentially and for the first half of this year, give us a great deal of confidence that even though our revenue growth rate is slightly below our targeted range, we will actually exceed on EBITDA margin expansion and be above this year for the remainder of the year, above 200 basis points.

But again, over the long term, we expect that 200 basis points to continue over a multiyear period at approximately 200.

Could it be a 190 or 220?

Sure, there can be some slight variability in those out years, but we feel highly confident in that margin expansion.

And a lot of that's coming from operating leverage.

You saw that we've been able to achieve lower cost of goods sold as well as SG&A efficiencies.

And I think those trends will continue for the remainder of this year and for the next several years.

And our next question comes from Walter Piecyk from BTIG.

Dave, with the $170 million EBITDA growth from the margin expansion, you only have $40 million on the share repurchase.

So that still gives you plenty of incremental cash to accelerate the dividend growth faster.

So I'm just curious why that didn't happen this quarter or why it won't happen in future quarters.

What's the argument against that from the board?

Sure.

So first of all, this is the first quarter that, that money was available for us to move to holdings, and we are in the process of moving that money literally as we speak.

Some of it is in Europe.

Some of those transfers have already occurred.

They -- we did have to report to be able to do that.

And I think the board will look at the dividend each and every quarter.

I don't want to mislead investors to say that there's a guarantee that will increase the pacing of the dividend growth, but we are committed to maintaining a leverage ratio of between 2.5 and 3.5x EBITDA on a net basis and we are delevering.

So arithmetically, the current rate of dividend increase is not sufficient to allow us over the next year or 2 to stay in that ratio.

We shouldn't fall below.

So we will either have to increase the pacing or accelerate the buyback program.

And again, that's the topic that the board reviews every quarter, but they wanted to get the monies transferred this quarter before they did anything.

Can you make a commitment to shareholders that, that decision is going to be made by the time next quarter is reported?

Because certainly, again, based on the cash and based on the share repurchase and based on your EBITDA growth, I mean, based on what you just said as far as that commitment to give it back to shareholders, there should be a decision that occurs by next quarter.

Sure.

The commitment I will make is that we will give the money back to shareholders and not operate an inefficient balance sheet.

I cannot -- I am 1 vote out of 6 on my board, and I cannot speak for the entire board.

They have to approve that authorization.

But they have reaffirmed their commitment to maintaining the leverage ratio that we have.

And as a result of that, we have to do 1 of the 2 mechanisms, which is either increase the dividend or buy back stock.

And which of those we're going to use has still not been totally determined.

So I'm not willing to put a stake in the ground to the mechanism.

What I am willing to commit to is we will use the balance sheet efficiently and return more capital.

Got it.

Just a question for Tad -- let's get him involved.

He, in his comments, talked about -- I think it was like -- what did you say, trimming of circuits in terms of what's going on in the net-centric, going from X to Y in terms of people basically buying larger circuits.

And then later on the call, there was some discussion about, yes, you're expecting usage growth to go up in net-centric.

What does that mean for revenue growth?

Are you expecting sequential revenue growth?

Like how does that all that kind of put together when you look at the circuit?

I forget the term you used -- circuit grooming, I guess, was the term and the mix shift and all this other [junk] -- does that translate to sequential revenue growth, do you think, in the third quarter for net-centric?

Sure.

The comment was to explain the decline sequentially -- quarterly sequentially in the units -- net-centric units.

Right.

So does that continue into Q3?

That's -- I guess that's the question.

We don't have that forecast to continue.

But from our revenue forecasting, it's really not significant because we still have the customer.

They're buying more traffic.

Now the price may decline and the price to decline for getting larger circuits, but the overall revenue forecast, frankly the way we model, is not based on corporate and net-centric.

We model based upon off-net and on-net because of the different margin characteristics of those lines of revenue.

And let me follow up on that, Walt.

So what is happening is customers are looking to minimize their cross-connect cost.

If a customer had 3 or 4 10-Gb circuits in a particular carrier-neutral facility, it may be utilizing 25 Gb of, say, the 40 Gb of circuit capacity that they have.

They have concluded it's in their interest to reduce those 4 cross-connects to 1, upgrade to a single 100-Gb port so they have 2.5x as much capacity, capability.

And as the traffic growth on that port increases, obviously -- the 44% per year is the average, we get more revenue offset by decline.

Now the mix shift piece of that is in a quarter -- are most of the sales going to the biggest customers or the smallest customers?

Some quarters, it's the very large FAANGs that are driving the biggest growth or big access networks.

The next quarter, you could just have a [mix shift] player.

It's more smaller dot-coms or smaller hosting companies that are driving the bulk of orders.

There has just been volatility over the past.

This is not a new phenomenon, but what we're trying to counsel investors on is when they see the price per megabit of new sales go up, don't think...

I understand that because people are focused on that metric, and I don't really care about that metric myself.

What I'm focused on is understanding if the usage growth is going to translate to sequential revenue growth in the net-centric business.

The answer is yes to that.

That's what I was saying -- the improved growth in that business over the remainder of the year.

Got it.

Why do you even -- well, we can talk about later about why you even report that other metric.

And our next question comes from Nick Del Deo from MoffettNathanson.

Going back to SD-WAN, if we look out over the next several years, how material an impact, do you think adding SD-WAN to the portfolio will have on the financial performance of the business?

Is this going to help accelerate corporate growth?

Or should we think of it as helping to sustain growth kind of in the current zone for a longer period of time than you otherwise would have been able to?

I think it will actually accelerate growth and also allow growth to continue for a longer period of time.

Ultimately, our growth in corporate is limited by the addressable market.

As we think about corporate DIA services, it is approximately a $9 billion addressable market in North America.

We have roughly 10% of that market, on-net, with Cogent facilities, and another 45% of that market is available to us with off-net services based on fiber that we can buy from 1 of 90 different suppliers.

We continue to increase our penetration in our corporate buildings.

Our corporate penetration is over 20 circuits per building sold to over 13 unique businesses.

Roughly 50% of our corporate customers buy 2 services from Cogent.

They buy a Internet service and a VPN service.

The VPN market is, by bit volume, much smaller than the business DIA market.

It's actually about 1/3 of the bit volume, but yet it is 5x as large in dollar services, equating to a 15x premium per bit.

SD-WAN and VPLS are both technologies that will migrate traffic away from those expensive MPLS circuits to a VPN utilizing the Internet either with VPLS, SD-WAN or actually both.

We think that the total addressable market for Cogent's corporate services will effectively increase by 30%.

Because of the price reduction per bit converging to Internet pricing, we think our total addressable market for corporate services will be about $12 billion, of which we will be able to serve about 10% of it on-net and about 1/2 of it off-net.

That will drive both a slightly faster growth rate, but also allow us to continue growing at the roughly 12% growth rate for a longer period of time.

Probably a decade from now, 1 of 2 things will happen.

Our corporate growth rate will slow because we have saturated our footprint or we elect to add more buildings.

Because we have remained so disciplined around our return on capital, we do not anticipate a spike up in the number of buildings.

Witness the fact that our average buildings added this year are actually slightly larger than the average of the embedded base, and our average building is roughly 50x the size of the average corporate building in America.

It is this disciplined cream skimming strategy that I think allows us to deliver these superior returns on capital to other telecom companies.

Okay.

That's helpful.

And then maybe one follow-up to Walt's question before.

How far are we into that conversion cycle from 10 Gb to 100 Gb?

I understand that it helps to save your customers some money on cross-connects.

But are there any material savings that you can wring out of it at this point?

If I'm not mistaken, the customer typically pays for the cross-connect when they connect to you.

Well, that is correct, Nick, and there is no real savings to Cogent.

But obviously, if we can reduce monies that the customer is spending to third parties, i.e.

the data center operator, that gives them more money they could theoretically spend with us.

In terms of our progress through that, I think we're still in a relatively early stage.

If you go back in kind of the 2006, '07 time frame, we went through a similar grooming cycle as net-centric customers were migrating away from 10 -- 1 Gb interfaces to groom multiple 1s onto 10s.

We're going through a similar cycle now, and I think we're probably only maybe 15% or 20% of the way through that.

This has really been the larger customers, the early adopters, but I think it will continue.

And there's no real impact either on our cost or revenues.

It does allow customers to have a little more headroom on their ports and, therefore, hopefully maybe grow a little bit faster.

But I think it's just kind of a different name on the same product.

And as Tad kind of indicated, it's units of bits that drive, not connections, on the net-centric side.

And our next question comes from Tim Horan from Oppenheimer.

Can you give a little bit more color on volume growth broadly speaking?

Is -- the net-centric customers, do you think you -- are they at a run rate at this point where you're capturing most of their volumes?

Or is this still kind of pent-up demand that you can kind of capture?

Yes, sure.

So we have rebounded from traffic growth of about 20% at the height of port congestion now to 44%.

We're still slightly below our long-term average, which is 49%.

So I do think there is some continued acceleration over the next several quarters in that year-over-year traffic growth rate.

To the second part of the question, which to me is maybe even the more important one, which is, are we winning a larger share of the customer's wallet?

I think the answer is yes to that.

We continue to see less emphasis on transit by our competitors, less choices.

Now partially offsetting that positive is the fact that a large amount of growth over the past year has come from the very biggest customers which get the very lowest pricing.

We don't determine where the growth is coming from.

We really follow that growth and make sure that we get as much of their share.

So if consumers are using the larger websites more and more and kind of shunning some of the smaller sites, that will push more traffic to the big guys.

And therefore, they will get the best pricing.

If, however, we continue to see a new wave of smaller companies starting to grow, which has been the historic pattern of the Internet, that can actually be a positive for revenue growth because most of the growth is coming from smaller businesses that don't get quite as lower price per megabit.

But I think across the board for both types of customers, we continue to grow much faster than the market.

And the hyperscale Internet companies are spending like an unprecedented amount of CapEx right now.

It looks to be in the $90 billion range this year.

And if you go back 5 years ago.

It was kind of $20 billion, $30 billion.

Are you taking advantage of that in any way?

Is that helping you guys out?

Or is it maybe just changing the dynamics of the Internet itself?

So it is absolutely a positive to Cogent, 3 ways.

First of all, we sell to every one of those major hyperscale providers.

And even though they may get lower prices per megabit because of their buying power, they give a disproportionate amount of that growth to Cogent.

And again, it's because the value we deliver to them relative to our competitors.

Secondly, it actually helps our corporate business because as hyperscale becomes more prevalent, more and more on-premise computing and on-premise software is migrating to a hosted environment and a hyperscale environment, which then becomes a positive for Cogent and helps drive customers who previously had maybe not chosen to buy Cogent in a corporate building to now needing and wanting our service.

It's why our growth and penetration remains about 1.8 connections per building per year even for buildings that have been on-net for 10 or 12 years.

And oftentimes, I get the question, "Well, if the customer hasn't bought from you yet, why would they buy in the future?" And the answer is their business is changing, and this hyperscale phenomena is the key driver of that.

And the third level of benefit to us is actually on the residential consumer.

As those hyperscale models continue to drive down the cost of developing new applications on the Internet, end-user consumers spend more time online with more bandwidth-intensive applications and because we sell to over 6,300 networks around the world that distribute our bandwidth, if their users, which are primarily residential, are using more bits, that's a good thing for Cogent.

So all in all, we see these as positive trends.

Great color.

Just last, you talked in the past about broadcast quality video reaccelerating the Internet growth possibly, and I think we've seen some multiplayer gaming that's very, very high quality also.

Do you think the overall Internet growth is improving?

I think you'd have probably better color on that than anyone at this point.

I think the answer is yes.

So the damage that was done to the entire ecosystem due to port congestion several years ago was barely significant.

Now service providers are willing and able to deliver to their customers the products that they have sold them and have adequate connectivity.

I think in many ways, for example, the litigation in New York against Time Warner, Spectrum/Charter is a strong signal that while the federal government has relaxed net neutrality rules, local regulators are going to protect consumers.

And because those consequences are so dire, I think every ISP now is trying to deliver a quality product to its customers.

As a result of that, we continue to see accelerating migration of video away from linear delivery to over-the-top.

And then secondly, we're seeing the resolution qualities continue to improve.

And I think we're actually at the early stages of seeing integration of what had more been a vertical for gaming and a vertical for video being completely not conjoined or different verticals now [sorting] to morph into a world where the high-resolution game and the high-resolution video may become more intertwined with one another.

And I think we are in a period of accelerating rate of over-the-top video.

And our next question comes from Frank Louthan from Raymond James.

This is Alex Sklar on for Frank.

Speaking of the net-centric, on the term and volume discounts, I think you said this quarter represented 2,200 units.

So over the past 12 months now, you've essentially given volume discount to about 30% of your net-centric base.

I'm curious, are you ever going to reach a point where the pace of those slow down or, based on the average term outstanding, we can kind of always expect about 1/3 of your base going to take advantage of those every year.

So I think the answer is you should expect it to continue about as is.

So if we look in our base, the majority of our customers sign 3-year contracts.

Actually, our average contract length is about 27 months.

Our net-centric customers, on average, grow at about the pace of the Internet, so about 25%.

As their volumes grow, they try to use that volume growth to come back to Cogent and renegotiate mid-contract provided that they will increase their monthly spend with us, meaning the growth and traffic exceeds the reduction in price per megabit.

We're happy to adjust those contracts, typically reentering into a new 3-year contract with the expectation that maybe 1/3 or 1/2 of the base will come back and again experience similar-type growth and renegotiate before the end of the contract.

So it is in our interest to maximize revenue from the customer.

It's in our interest for the customer to grow their traffic faster than the market in general.

On average, we -- our customers grow about like the market, and our growth comes from winning new customers and winning wallet share, to Walt's question, among customers.

And that volume discount is another important tool and that ability to win wallet share from customers.

I might just add one thing.

For a customer to change their contract in term, they must increase their total contract value to Cogent.

So there could be a price decline, but the total contract value must increase.

Got it.

And actually, 2 follow-ups on the accounting for Tad.

On the capital as commission benefit, can you just clarify that $300,000 was a sequential impact?

Or is that year-over-year?

And when do you actually expect that to normalize?

And then I know you only have a small number of operating leases, but any color on how you're thinking about the lease accounting change for next year or how it might impact the income statement?

Sure.

The $300,000 is the impact for the quarter.

The impact during the first quarter was $1 million.

So if you look at sequential SG&A for the first quarter, it declined from the commission benefit by $1 million, and kind of the difference between the amount paid in cash and the amount recorded was $300,000 for the quarter.

That's the way to think about that because of the capitalization.

It is difficult to model going forward because you don't know necessarily which orders will cancel, the amortization rate, et cetera.

I would not expect it to be materially different from what it was this quarter going forward.

With respect to the new lease pronouncement, it has absolutely no impact on the income statement.

It's entirely a balance sheet exercise.

So what we will be required to do on January 1, 2019, is to record a new asset called operating lease asset, if you will, and a new liability -- operating lease liability, and those will be put on the balance sheet.

As payments are made and as we account for our operating leases, it will be the normal P&L charge.

That asset and that liability topside, if you will, will be reduced each month and each quarter based upon the amortization tables with nothing hitting the P&L.

This is completely a topside exercise.

And I am showing no further questions at this time.

I would now like to turn the call back to Mr. Dave Schaeffer, Chairman and Chief Executive Officer, for any further remarks.

Well, first of all, thank you all very much.

I apologize for the technical hiccup.

I'm not sure what happened, but I'm glad we were able to answer all the questions.

Just in summary, I think we feel very strong about our business.

Our operating leverage, which I know has been questioned at times, continues to, I think, shine through.

And the efficiency of our model, whether it be in headcount or costs, continues to outperform others.

And our growth remains very strong in 2/3 of our business on a very consistent basis and volatile in another portion of our business, but continuing to improve as the traffic growth accelerates, as the net-centric customer base continues to widen and as prices remain at roughly the same rate of decline that they've historically declined.

We anticipate improvement in that business throughout the year and, therefore, our ability to generate more cash, and it's exactly why we have been so confident in increasing our dividend on a going-forward basis.

So again, thanks, everyone and we look forward to talking soon.

Take care.

Bye-bye.

Ladies and gentlemen, thank you for your participation in today's conference.

This concludes today's program.

You may all disconnect.

Everyone, have a great day.