Good day, everyone, and welcome to the Cogent Communications Group, Inc. Fourth Quarter and Full-Year 2012 Earnings Conference Call and Webcast. Just a reminder, today's conference is being recorded. Call will be available for replay at www.cogentco.com. Now for opening remarks and introduction, I will turn the conference over to Chief Executive Officer, Mr. Dave Schaeffer. Mr. Schaeffer, please go ahead.

Okay. Thank you, and good morning to everyone. Welcome to our Fourth Quarter 2012 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 extremely pleased with our results for the quarter and for the full year, and we are optimistic about the strength of our business and outlook for 2013. During the quarter, we experienced sequential revenue growth, gross margin expansion, EBITDA margin expansion, significant cash flow and significant growth in traffic on our network. Additionally, we returned $5 million to our shareholders on our increased quarterly dividend of $0.11 per share per quarter, which we paid in December. We continue to remain confident about the cash-generating capabilities of our business. As a result, and as indicated in our press release, we have announced an additional 10% increase on our quarterly dividend to $0.12 per share per quarter to be paid on March 15, 2013 to holders of record as of March 4, 2013. We generated positive cash of $15.2 million for the fourth quarter and $9.1 million for the year, including our dividend payments and all of our investing and financing activities. Excluding our dividend payments and amount spent on stock buybacks, we generated positive free cash flow of $19.9 million for the year. Our revenue grew for the third quarter by 3.7%, and on a constant currency basis, our sequential revenue growth from the third quarter was 2.9%. Our revenue growth for the full year from 2011 grew by 3.8% and on a constant currency basis, our full year revenue growth year-over-year was 5.5%. During the quarter, our traffic on our network growth rate accelerated and grew 27% sequentially from the third quarter, and our traffic grew 36% from the fourth quarter…

Thank you, Dave, and good morning, everyone. This Fourth Quarter 2012 and Full Year 2012 Earnings Report in this earnings conference call discuss Cogent's business outlook and contain forward-looking statements within the meaning of Section 27A and 21E of the Securities Act. 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. You should also be aware that Cogent's expectations do not reflect potential impact of mergers, acquisitions, other business combinations or financing transactions that may be completed after today. Cogent undertakes no obligation to release publicly any revision to any forward-looking statement made today or otherwise update or supplement statements made on this call. Also, during this call, we use any non-GAAP financial measures. You will find these reconciled to the GAAP measurement in our earnings release and on our website at cogentco.com. Now I'd like to turn the call back over to Dave.

Okay. Thanks, Tad. Hopefully, you've had a chance to review some of our earnings, our press release data. As within previous quarters, our press release includes a number of historical quarterly metrics. These metrics will be added to our website. We hope that you find the consistent presentation of these metrics informative and helpful in understanding our financial results and trends from operations. We experienced strong revenue performance, both for the quarter and for the year. Our fourth quarter 2012 revenue was $82.6 million, an increase of 3.7% from our third quarter revenue and an increase of 4.4% from Q4 of 2011 revenues. On a constant currency basis, our quarterly revenues increased sequentially by 2.9% from our third quarter, and on a constant currency basis, our revenues increased 5.1% from the fourth quarter of 2011. Our 2012 revenue was $317 million, an increase of 3.8% from full year 2011. On a constant currency basis, our full year 2012 revenue increase was 5.5% from 2011. We evaluate all our revenues based on product class: on-net, off-net and non-core, which Tad will cover in greater detail. We also evaluate our revenues by customer type. We classify our customers into 2 primary groups: NetCentric customers and corporate customers. Our NetCentric customers buy large amounts of bandwidth from us traditionally in a carrier-neutral data centers. Our corporate customers buy smaller amounts of bandwidth on generally a fixed connection in multi-tenant office buildings. Revenues from our corporate customers grew by 1.4% from the third quarter. These corporate customers represent 49.5% of our total customer connections at the end of 2012, and represent 51.2% of our fourth quarter 2012 revenues. Our corporate customer connections grew by 1.7% sequentially in the quarter. Revenues from our corporate customers grew 9.2% from 2011 and our corporate customer connections for…

Thank you, Dave, and again, good morning to everyone. I'd also like to thank and congratulate our entire team for the results of their hard work of their efforts this quarter and for the full year of 2012. I'll begin my discussion be providing additional details on our revenue by product class. Our on-net revenue was $60.4 million for the fourth quarter, which was an increase of 3.9% from the third quarter. Our on-net revenue for 2012 was $232.6 million, and that was a decrease of 0.2% from 2011 and an increase of about 2% on a constant currency basis. Similar to prior quarters, about 85% of our new sales for the quarter were for our on-net services. On-net customer connections increased by 3.6% for the quarter, and increased by 17.1% for the year. We ended the year with about 29,800 on-net customer connections on our network and our 1,867 on-net buildings. The revenue from our off-net business was $21.6 million for the fourth quarter, which was an increase of 3.5% from the third quarter as we continue to add off-net connections as well to our network. Our off-net revenue was $81.9 million for 2012, that was an increase of 17.6% from 2011. Off-net customer connections increased by 4.9% for the quarter and by 14% for the year. And we ended the year serving about 4,500 off-net customer connections in approximately 4,000 off-net buildings. Our non-core revenues were $600,000 for the third and fourth quarter, and represent less than 1% of our quarterly and annual revenues at this point and about 470 customer connections. On ARPU, both our on-net and off-net ARPUs were essentially flat for the quarter. Our on-net ARPU for corporate and NetCentric customers combined were $688 for the third quarter and $686 for the fourth quarter. Our…

Okay. Thanks, Tad. We're going to spend a moment talking about sales rep productivity and activity. We began the fourth quarter of 2012 with 273 quota-bearing reps and ended the quarter with 254 sales reps. We hired 46 reps in the quarter and 65 reps left the company during that quarter. Our monthly rep turnover rate was approximately 8% in the fourth quarter. This is slightly above our historical long-term average turnover rate in the sales force of about 7% per month. We began the fourth quarter of 2012 with 263 full time equivalent reps, these are reps that are ramped on to their full quota and ended the quarter with 251 full time equivalent sales reps. Productivity on a full time equivalent rep basis for the quarter was 5.2 units per rep per month of installed business. This performance was significantly better than our long-term historical average for sales rep productivity of approximately 4.2 units per rep per month. As a reminder, our rep productivity numbers are based not upon contract signing, but rather only upon the completion of installation and the commencement of billing. Now for a moment about the scale and scope of our network. The size and reach of our network continues to grow. We added 35 buildings to our network in the fourth quarter, and now have over 1,865 buildings directly connected to the fiber on our network. Our network now consists of 26,300 metro fiber miles and over 56,600 route miles of intercity fiber. The Cogent network is one of the most interconnected networks in the world, and today, we are directly connected to over 4,360 other networks that comprise the Internet. Approximately 40 of these interconnections are settlement-free peers. The remaining 4,320 networks are paying Cogent customers. We are utilizing approximately 20% of…

[Operator Instructions] We'll take our first question today from Michael Rollins with Citi.

A couple of things. First, in terms of the revenue growth guidance that you gave, the 10% to 20% range, if I look at some of the disclosures that you made in the back of the press release, it talks about constant currency normalized for the excessive churn. 2012 was around 11%. Constant currency growth sequentially over the last 2 quarters, I think, was plus or minus, about 2.9%. Can you talk about whether the base case for 2013 might be in that kind of same range as it's been in over the last few quarters and for the year? And what are the things that could get you significantly higher within that range towards the middle to higher end of that growth guidance?

Thanks for the question, Mike. So as you know, our growth tends to fluctuate slightly quarter-to-quarter. We have -- we feel very confident about the range that we have laid out. That is 10% to 20%. And that is allowing for volatility in foreign exchange. You are correct that on a constant-currency basis, our annualized growth rate has been in the 12% range. We are encouraged by a number of factors. One, rep productivity continues to be increasing. So as we look at 2012, each quarter throughout the year, the productivity per sales rep increased sequentially each quarter, and the average historical over the past 32 quarters has been trending up. We ended the quarter with I think the second highest rep productivity quarter in the company's history. Secondly, on the NetCentric portion of our business, we're seeing 2 positive trends. One, a slight moderation in the rate of price decline. Again, we feel the long-term trend is for the price per megabit to decline in perpetuity, but we have seen that the rate of the price decline moderate to some extent. And secondly, we've seen an acceleration in our rate of traffic growth. That traffic growth number is being driven by a number of new applications, predominantly video over the Internet. So I think the combination of increased rate of traffic growth, coupled with a lower rate of price decline per bit, means acceleration in the NetCentric portion of our business. To the corporate portion of our business, rep productivity is critical. We also see an increasing number of customers depending on cloud-based or remote applications that they are hosting away from their wiring process or internal server locations. You've seen our average number of corporate customer connections per building for our on-net buildings increase about -- to 10.5 from 10.4 the previous quarter. We also see the proliferation of off-net Ethernet continuing to be pushed out by the major incumbent carriers, allowing us to continue to transition people away from TDM-based services to Ethernet-based services. For those reasons, you've seen our ARPUs effectively stabilize and our corporate business, both on-net and off-net, and again with higher rep productivity, we feel pretty comfortable in our range and obviously optimistic that we can get to the higher end, not the lower end of the range.

We'll take our next question from Barry McCarver with Stephens.

I guess, first off, David, in your opening remarks, you mentioned your targeted commercial programs and effect that they had on pricing and revenue. Could you talk about how we should look, think about that for the outlook for 2013, how the timing of the continuation of those programs? And what we could expect from those?

Yes, sure, Barry. So let me take programs targeted to each of our 2 customer bases. On the NetCentric side, we offer prices as low as $0.50 per megabit. That's obviously substantially below the $1.80 that was the average new sale in the quarter. Those were typically for fixed volumes and targeted to specific potential customers. We expect to continue to do that throughout the year and continue to be aggressive in gaining market share, having gained over an additional 100 additional customers via this type of targeted program, again in the most recent quarter. So I think we will continue to be aggressive. We have tremendous operating leverage as demonstrated by both the gross margin and the EBITDA margin expansion that we demonstrated in the quarter, in light of the declining price per megabit. And the result is increased rep productivity and lower cost of revenue acquisition on the NetCentric side. Switching to the corporate business, it's maybe a little less exciting and maybe not quite as easy to characterize in a price point. We have always offered various promotional activity. These activities are designed to create a sense of urgency on the part of our corporate customers, and they can range from things such as waving install cost to offering multi-year discounts on shorter-term contracts, sometimes waiving 1 free month or perhaps giving a free month for each year of service. That net result of these types of promotions has generally been about a 7% or an 8% discount to our list price. That has not changed over time and has really been part of our sales strategy for the past 10 years. So while our nominal list price for our corporate customers has not changed in 11 years or 12 years of selling, it is true that we've always offered discounts, again about the same rate. So the effective price per connection has also been very flat for those customers. So promotions are part of sales. We're a sales-driven organization. But I think because we are utilizing such a low portion of our network capacity, only about 20%, and we continue to have a substantial amount of addressable market we're not serving, only serving 10.5 customers per building out of 51; we think that this kind of targeted promotional activity is the right way to continue to grow our revenues.

Okay, great. And then on your guidance for margin expansion, if I recall historically, you've given guidance for, as you did this time, 10% to 20% revenue growth and at least 200 basis points of margin improvement. And it sounds like, this time around, for 2013, you're targeting right at 200 basis points. Am I hearing that correctly or am I reading too much into that? What caused the change in the guidance there?

I think you're probably parsing our words a little too precisely. I am comfortable with the "at least" term, approximately huge just because there's volatility. I know that. Obviously, in the past couple of quarters, we demonstrated 530 basis points of EBITDA margin expansion. And we want to just caution investors that they should expect at least that 200 basis points, but there will be some lumpiness in quarters. And as Tad mentioned, SG&A is going to go up in Q1. It does every year. This year will be no exception. Last year was particularly anomalous because of the loss of Megaupload. But absent that, we continue to see without a $1.5 million of increase in our SG&A cost in the first quarter over the fourth quarter, and then that begins to drop off. A lot of this is payroll reset benefit increases that hit us abnormally in the first quarter, coupled with some of our audit and sales meeting expenses. So again, I don't want people to read into any reduction in our guidance. We're comfortable with the term at least 200 points.

Okay, great. That's what I thought. And then just real quick. Lastly, you mentioned that there was some CapEx in the year related to the storm. I assume you're talking about Hurricane Sandy. Could you just share with us what that CapEx was? I'm just trying to get an idea of what CapEx would've looked like for 2012 without that?

Yes. We had said that we expect the CapEx to come down by about 10% year-over-year. It only came down by 3.3%. And as we indicated, there were kind of 2 reasons for that. One was the increased expansion activity. We said we took advantage of 2 opportunities. The second was we did bear some extraordinary expenditures related to the storm. Our data center in lower Manhattan was on generator, and then ultimately needed the switchgear in the building replaced. We also experienced a number of multi-tenant office buildings in lower Manhattan, where the basements of the buildings were flooded and we had to replace equipment. We have submitted claims to our insurance carriers to look for these reimbursements. Those claims are in the kind of $1.5 million range. Most of that was capital expenditures.

We'll take our next question from James Breen with William Blair. James D. Breen - William Blair & Company L.L.C., Research Division: One, on the churn side, you came up a little bit there in the quarter. Can you talk about what your thoughts are around that? And do you expect it to come down over the next couple of quarters? And then secondly, can you give some color on sort of building expansion from where you are now? Is there a certain number you have, you had in terms of getting to over the course of the next year? Or are you sort of done building out and sort of what your thoughts are there?

Sure. Let me start with the churn question. Jim, thanks for the question. We did lose some customers as it relate to super storm Sandy. Some people just have to vacate their buildings. Some of them relocated to other buildings. I think that was probably the biggest factor in the slight uptick in churn. Remember, our largest concentration of on-net corporate customers is in New York City, and we do have a large number of financial services companies that do operate in lower Manhattan. Some of those customers are still operating in temporary space in midtown or other parts of the greater New York area. Also, on the off-net side, we did experience a little longer window to get installs done, but churn was flat. It didn't really change at all. So it didn't really impact that part of the business. We do expect kind of continued normal trends in both on-net and off-net churn, and we see no material changes. And you kind of see that represented in our revenue growth numbers. In terms of your building question, we have 1,303 corporate buildings connected to our network, 564 data center buildings globally connected to our network. And in those data center buildings, we have about 650 unique data centers because some buildings house more than 1 individual data center within the building. So a building like 1 Wilshire or 60 Hudson. Our additional buildings are in both camps -- of the 125 or so buildings we added last year, about half were corporate, about half were data centers. We expect to continue to add data centers pretty much perpetually, as far as the eye could see, at the rate of 60 to 70 a year, and we react to new data centers being constructed. There's been a number of new centers announced. Virtually all of those centers in our 36-country footprint, we are connecting to. On the corporate side, we do see a moderation in the number of new corporate buildings that we add. We have about 75 or so additional buildings we expect to add over the next couple of years, corporate buildings. So we do think the number of corporate buildings added will decline from the 60 or so a year, down to 35 or 40, and then eventually taper off to just a handful of buildings going forward. And that should help us achieve our long-term goal of continuing to bring down our capital expenditures, not only as a percentage of revenues, but absolutely in an absolute term we expect our CapEx to continue to decline.

We'll take our next question from Frank Louthan with Raymond James. Frank G. Louthan - Raymond James & Associates, Inc., Research Division: So when you look out, you mentioned adding a little bit of capacity here and there in your network. Any M&A opportunities you see out there, any assets that are for sale that you think are attractive? And what are your thoughts there? And then can you give us a little bit more color on your philosophy of returning cash? You had the buyback here and the stock's obviously done very well, which has been a nice problem to have, having bought back as much and I appreciate the dividend increase. What is your thinking there on the dividend? I mean, in the past, you wanted to return the cash you got to shareholders. Is there any -- are there any thoughts of a special dividend or raising the dividend for a period of time to return that cash? And would you ever lower the dividend or do you see that as sort of a permanent base line as you're giving that out? Just kind of in light with some recent events, would love to hear your philosophy there.

Thanks for the questions, Frank. Let me start with the M&A question. We look at numerous opportunities on a global basis. We have a very strict filter we run those opportunities through. And they really fall into 2 major categories, either operating businesses that we can buy and bolt on to our network that would generate incremental free cash flow per share, and would justify a return on capital greater than our cost of capital. It is true that we have nearly $250 million of cash on our balance sheet. We do believe we have additional borrowing capacity if we found the right opportunity, and we have not been able to find opportunities that actually will, on an operating basis, generate returns on capital greater than capital that we have a core competency in running and bringing on to our networks. So for that reason, we passed on those opportunities. And we literally looked at hundreds of opportunities since we've done our last acquisition in '04 and have not found ones that fit that criteria. The second criteria would be physical asset that would be additive to our business, either network assets or we would even look at customer bases. We continue to look at opportunities but have not found anything that we felt was better to buy the whole business than just buying select assets from other businesses. So it's not that we're disinterested in M&A. It's just we have not seen things that we felt were priced appropriately. With regard to the return of capital question, I believe we have demonstrated a prudent, but also credible pattern of returning cash to shareholders. Remember, this buyback is actually the fourth buyback program in the company's history, and we've spent about $180 million in buybacks over the years, with an…

We'll take our next question from Greg Mesniaeff with Maxim.

I have a question on some of the commentary you made about some moderation of growth in terms of your network expansion. I'm wondering, when you look at your building additions, and the cost associated with adding new buildings, what kind of trends are you seeing in terms of the cost of adding a new building, whether there would be a data center or a corporate location, if you can just give us some color on that? And also, as sort of part of that answer, some of your existing leases come up for renewal on some of the buildings you have, what are you seeing in terms of the trends there?

Thanks for the questions, Greg. So first of all, most of our capital to add a new building is actually outside plant work. It's digging in the streets, backhoe, conduit work, repaving, and those costs are not really impacted by technology and generally increase a kind of CPI. Typically, it costs Cogent about $120,000 to add a building to its network. For the corporate building, split is roughly 2/3 outside plant, about 1/3 inside plant, meaning the riser work that we need do in the building. The actual active equipment to add the building is de minimis. We either use equipment we have in inventory or even new equipment continues to fall in cost. So it's not a question of the optronics or the routing in the building, it's really a question about that outside plant work. For a new data center, about 85% of the CapEx, which is also about $120,000, is outside plant, and only about 15% is inside plant within the data center. Data centers are necessary for us to be able to reach the entire addressable market for NetCentric services. So we're going to continue to expand into virtually every data center that has scale and is truly carrier-neutral. And based on historic expansion by the data center operators, we've been getting there. And then in terms of the corporate buildings, we had a tough filter. We're really selective about the buildings we go into. I know we're in your building, the Chrysler building in New York, and that's the example of a building where there's lots of tenants and 1.5 million square feet. There's a lot of smaller buildings, but we're just not comfortable in chasing those smaller buildings. So for that reason, we think we're going to continue to see fewer corporate buildings added. But remember, our corporate on-net business, which is about 1/3 of our total business, will continue to grow, primarily driven by increased penetration. We only have 10.5 out of 51 opportunities. So again, when I walk into your building and see the marquee in the lobby, there's a lot of people that are in that building that are not yet Cogent customers. Tad's going to take the one on the building agreement renewals.

Yes. Costs to operating a building really are true lease expense for space, and then we also pay for multi-tenant office buildings we pay, now what we call a license agreement. See, it does include the rights for some space, but it's generally the rights to provide service and to connect our network into the building. With respect to trends in those costs, some of those agreements, in particular the ones for space may have a CPI or a fixed increase. But as we've seen those agreements renew, we're tending to see a lower rate for renewals, in particular for the license agreements to provide service into the building.

We'll take our next question from Tom Seitz with Jefferies. Thomas O. Seitz - Jefferies & Company, Inc., Research Division: I've got 2. Dave, you noted that the rate of NetCentric price decline has moderated a bit. Can you just talk about the competitive environment in NetCentric from your standpoint generally? From your view, has Level 3 global integration made any progress? Are you seeing Zayo more? Any insight there would be great. And then secondly, just so we have this modeled properly, can you help us with your expectations on off-net? Prior to this call, I had thought you had said that in the past, as the conversion of Ethernet saturates, off-net was going to slow a bit. What's your updated expectation on when we might see that again? I just want to make sure that we get this modeled properly.

Thanks for the questions, Tom. So let's start with NetCentric, which is almost exclusively an on-net business. There are fewer competitors in the market, and we continue to see our rate of price decline being lower than the rate of price decline among our competitors. But as prices for those competitors try to attempt to get closer to our price point, many of those competitors drop out of the market. With regard to specifically, the Level 3 and Global Crossing, they tend to be focused on other products, high-speed IP. It's only about 9% of their revenues by their disclosures. And for a variety of reasons, they have not been able or willing to try to compete directly with Cogent. With regard to other competitors, a lot of the incumbents have an IP Transit business inside of their business, but they deemphasize it. Zayo is both a great supplier to Cogent, as well as, well, a competitor on cases. But IP Transit is not their primary business. In fact, it's an extremely small business. It didn't exist inside of Zayo prior to the acquisition of AboveNet. And even with AboveNet, it was a very small portion of AboveNet and even a smaller portion of Zayo's business. So we're encouraged by their fiber construction. We continue to be an acquirer of fiber from Zayo. We're one of their larger customers, I believe. They do buy some IP services from us. They are a great customer of ours. And we see that as a good symbiotic relationship. So the takeaway on the NetCentric side is that I think competition is moderating a little. Prices will go down, but maybe not quite as deep as they have in the past. Now to your off-net question, I'm going to take part of it, I'm going to give Tad a part of it as well to touch on, but a couple of things have happened. One, we had an installed base of TDM business that we have, I think, pretty effectively cycled through and converted almost to all that we could convert to Ethernet. Secondly, virtually all of the new off-net sales are Ethernet-based, and we've been somewhat limited by the Ethernet footprints of AT&T and Verizon as the 2 primary incumbent providers. Over 90% of our off-net circuit come from the incumbent. We do buy from alternate providers when available, but there's not a lot of footprint. I do believe that as both of those companies have articulated a more aggressive Ethernet buildout, it will provide a bigger addressable market for us, for IP Ethernet services. And I think we may see a slight reacceleration in that growth rate. So what we were expecting to moderate may not moderate, not because we're not upselling the existing customers it's now, we just have more locations where we can actually sell service. And I'll let Tad also just touch on some of the actual numbers.

Sure. Our off-net business increased about 17%, 17.5% from 2011 to 2012. Underneath that, we're seeing the ARPU, which historically has been increasing quarter-over-quarter, kind of stabilized at where it is now. That's about what we're expecting to occur over the next several years. And were expecting the growth rate for that business to be relatively consistent to what it was between 2011 and 2012 as a result.

We'll go next to Donna Jaegers with D.A. Davidson. Donna Jaegers - D.A. Davidson & Co., Research Division: On -- you had some verbiage in your comments about increased demand of video that you're seeing from NetCentric customers. There's been a lot of speculation about one customer in particular. Can you give us any other color on what you're either doing for that one customer or just what you're doing in the video connectivity market?

Sure. So as we've commented on previous calls, our NetCentric business is driven predominantly by residential users. Even though Cogent itself has no residential consumer business, either our access network customers are delivering service to end-user customers, or our content companies are basically targeting those end-users predominantly. Today, between 75% and 80% of the total bit load on the Cogent network, and I think it's indicative of the global Internet, is video. In addition to that, virtually all of the growth is coming from video. Many large video producers and content distributors had historically outsourced that function to a CDN or a hosting company. Increasingly, as their unit volumes increase, they are in-sourcing, so they are building effectively their own CDNs. In order to deliver service, they need space, they need power, they need computing power, they need storage, and where it's important to Cogent, they need bandwidth. We have large and small customers that use us, some of which have been in the press. Our good friend, Mega, is back, and he's publicly announced he's exclusively using Cogent for his bandwidth. Some of our other large customers like Netflix have also commented that they're reliant on our bandwidth to help deliver their services. Our view is at everyday, we have to earn that customer's business and deliver the best value with the highest quality in the market. And I think you're increasingly seeing Cogent take market share. The fact that our unit volume increased sequentially, 27%, I mean that's probably about as strong as it's ever been in a one quarter increase. Annualized, that's like 120% growth rate. Clearly, the year-over-year comps were challenging because of the hiatus of Megaupload, they were rebuilding a new business model. But yes, we are, I think, continuing to see all of our customers generate more video traffic. And as video consumes a lot of bandwidth, they will become more price sensitive and therefore, use more Cogent. So that's part of the reason why we're encouraged about our growth trends. Donna Jaegers - D.A. Davidson & Co., Research Division: Okay. And then one sort of a longer-term question. A lot of those NetCentric customers, once they know the density of traffic in their network, they will try to buy waves on the densest part. You guys sell IP Transit, so not waves. Have you thought, as we look at this move to 100-gigabit optronics, have you thought about getting into the wavelength business?

We can clearly sell waves. We've always had that capability. Today, we operate 65% or so average 10-gig waves per route. We have capability of going to 160. We do operate some 40 gig. We have some limited 100 gig in the network for -- mostly testing purposes. We don't think it's quite yet ready for prime time, but will be shortly. We have chosen not to be in the wavelength business by design. We think transit is a more cost-effective way to move the bits. So I would actually disagree with your premise. I actually think more people are switching to Transit, away from wavelengths, that is the growthier part of the market. That's why we're growing versus those companies that have a pretty much higher percentage of their revenues coming from wavelengths through a flat or declining revenues. And transit is just more efficient. It's more flexible and it allows us to provide the combination of transport, routing and interconnection as a bundled service. So I actually think the long-term trend is a continued growth in IP transit and a moderation in the number of customers who buy wavelengths to try to recreate a less-efficient subscale network.

We'll take our next question from Ana Goshko with Bank of America Merrill Lynch.

This is going kind of long, so I'll try to be quick. I just wanted to see what you're thinking about in terms of plans for addressing the 2 main pieces of debt in your cap structure. One on your bond, the 8 and 3s, while you put that in place on a secured basis a couple of years ago, when it seemed like it was a good deal given your performance as reflected in the rating upgrade, and just the historic low rates that we're seeing in the market. It is now an expensive piece of debt relative to where you could probably issue. I know you're call-constrained for a while there, but wondering if you've run any math on the MTOB [ph] and potentially taking that out early and refi-ing at a lower rate. And then the second one would be, while it's dependent I think largely on your stock price, there is a discrete output on your convert next year? And should you face that put, wondering how you're thinking of addressing that as cash on hand or what you think to refi that, and at what level?

Thanks, Ana, and let me take the convert first. That's somewhat out of our control in the sense that you are correct, there is a put. And ultimately, the first choice belongs to the convert holders. It is dependent on stock price, and listen, it'd be a great problem to have, to have those people not want to put it, and us decide to have to call it. We are fully expecting at this point in time that we will take out the convert. Now that may change in the sense that the convert holders may choose not to put it to us. And we would most likely do that with cash on our balance sheet. We see no need to issue a new convert or either a senior, junior piece of paper to replace the convert that's outstanding today. Listen, we listen to all types of financing activity just like we look at all types of M&A. I'm not going to preclude ever issuing another convert again, but I think at this point in time, our expectation is we'll just retire that debt off of our balance sheet from cash on the balance sheet, as well as cash we're continuing to generate. With regard to the senior secured debt, you are correct. It is expensive. We were a first-time issuer, would appear to be a good deal, it's not as good a deal. There is call protection, so it's kind a moot point at this point. I think once we reach the end of that call protection period, we would evaluate the market conditions and run that arithmetic. But I think it's just pretty mature at this point since we've got almost 2 years of call protection still built in for us to think about that because I can't predict where rates will be 2 weeks from now, or let alone 2 years from now. But I think we've demonstrated to all of our stakeholders, the debt holders and the equity, we're very prudent about our use of cash. And listen, I understand the negative carry of the cash on our balance sheet, and I think about that all the time. But at the same time, we're going to be very judicious about putting capital to work. We'll look in M&A. We'll look in places organically to deploy it. We'll look at a methodical way to return it to shareholders. I think these are all legitimate, and they all change from time to time. But I think it's fair to summarize that you should not expect any more debt activity from us in the near term.

And with no other questions in queue, gentleman, I'll turn it back to you for closing remarks.

I just want to thank everyone. I always say we go a little longer than I'd expected, but these were great questions. We feel very good about the quarter. I want to thank the whole Cogent team. And again, I appreciate everyone's support. Thanks a lot. Take care. Bye, bye.

Ladies and gentlemen, thank you for your participation. This does conclude today's conference. Have a great rest of your day.