Cogent Communications Holdings Inc (CCOI) 2006 Q1 法說會逐字稿

Good morning and welcome to the Cogent Communications group first quarter 2006 earnings conference call. As a reminder this conference is being recorded and 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.

Thank you very much, and thank you all for joining us today and good morning. Welcome to our first quarter 2006 earnings conference call. I am Dave Schaeffer, Cogent's Chief Executive Officer. With me today is Tad Weed, our Chief Financial Officer. We are extremely pleased with the results for the quarter and want again, personally, I'd like to have the opportunity to thank the entire Cogent team for their efforts in helping us achieve these results. Throughout our discussions we will continue to focus on the results from our on-net business which, just to remind you, is our growth business. Our on-net services are provided to customers located in buildings that are physically connected to our network, the most literal definition of on-net. We will also highlight a few operational statistics that we believe will help demonstrate our expanding scale in the market.

I will start the call today with a review of certain operational highlights. Tad will provide some additional detail on our financial performance for the quarter and walk through our expectations for second quarter 2006, as well as our expectations for physical year 2006. Following our prepared remarks we will open it up for questions and answers. I'd like to take a moment and apologize for the technical difficulties from our last call. We actually have changed vendors for the conference bridge, and hopefully we will be able to answer all of the questions of the audience. We do expect a very efficient question-and-answer session at the end of this call. Now I would like to turn it over to Tad to read our Safe Harbor language.

Thanks, Dave, and good morning, everyone, this first quarter earnings report and this call discuss Cogent's business outlook and contain forward-looking statements. The specific forward-looking statements cover Cogent's expectations for revenue, EBITDA as adjusted, our percentage of on-net revenues, and operating results, and loss per share for the second quarter 2006 and fiscal year 2006. These particular 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 the potential impact of any 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 if we use any non-GAAP financial measures as defined by the SEC and Regulation G, you will find these reconciled to the GAAP measurement either in our earnings release or on our website at cogentco.com. Now I would like to turn the call back to Dave.

Thank you very much, Tad. I'd like to start with some highlights from our fourth quarter's results. Hopefully you've all had a chance to look at our earnings release. As with previous quarters in the press release we have tried to include a number of operational and financial metrics as well as several historical quarters to give you the ability to compare and contrast our results. These metrics will also be added to our website. Hopefully you will find these metrics informative and helpful in understanding the results of our operations. Always we would like to welcome suggestions for additional metrics for helping us refine these metrics if they can better help you understand the performance of our business.

Our first quarter in 2006, our revenue was 34.4 million, at the top end of our revenue guidance range of 33.5 to 34.5 million, which we provided to you on our last earnings call. EBITDA of 3.3 million was above our guidance range, which was previously stated at 2 million to 3 million. Our gross margin improved from 39.9% to 41.0% from Q4 '05 through first quarter 2006. We're very pleased with the continued growth in our on-net business, whose operational leverage is driving the expansion and our EBITDA and gross margins. Our on-net revenues grew by 8.1% sequentially quarter over quarter from Q4 2005 to Q1 2006. In looking at our revenue in more detail, our total service revenue grew from 34.4 million in Q1 '06, from 33.2 million for Q4 '05, which represented a 3.7% growth in that aggregate revenue. Traffic on our network grew by approximately 15% sequentially from Q4 2005 to Q1 2006. We believe that this continued growth in traffic indicates our continuing gain in market share in the market and also is a validation of our business model, which allows us to continue to operate as the low-cost provider and as a price leader, providing the highest quality at the lowest price to our customers.

On-net revenue grew by $1.7 million, or approximately 8.1% sequentially Q4 '05 to Q1 2006. This growth rate was over twice the growth rate we had experienced sequentially previously from Q3 2005 to Q4 2005. Our on net revenue increased as a percentage of our total revenue from 63% at the end of Q4 2005 to approximately 66% of our revenue in Q1 2006. Our on-net customer connections grew by 13% quarter over quarter, from approximately 4700 customers on-net at year end to approximately 5300 customers at the end of first quarter. Revenues from our legacy off-net business, to remind you, this business requires us to purchase a local loop from an incumbent telephone company operator to connect to our customers, declined by 5.3%. This was an improvement from the rate of decline that we experienced Q3 of 2005 to Q4 of 2005, where that decline was 8.8%. Total connections to our network at the end of March 2006 were slightly over 10,000.

Now I'd like to transition and talk a little about the pricing environment and the average revenue per subscriber that we are experiencing. Our pricing policies remain unchanged. We are continuing to be committed to being a price leader. The pricing for our most widely sold product, that is a 100-megabit-per-second non oversubscribed Internet connection for $1000 remains unchanged at $10 a megabit. I think we are experiencing probably somewhat of a slowing from some of our competitors in the rate at which they are reducing their prices, but we remain the low price provider in the market.

Our on-net customer connection growth at 13% exceeded our revenue growth at 8% primarily due to the nature of the customers that we are connecting. As we have been adding additional sales resources, we have been focusing on customers who may purchase smaller amounts of bandwidth, particularly in our service provider segment where we are able to go to either distribution networks or content companies that were smaller than our traditional target universe as we now have sufficient sales resources to concentrate on these customers. Our on-net average revenue per user was $1525 for Q1 2006, versus 1605 for Q4 2005. A decline of approximately 5%. This was less than the decline rate that we experienced Q3 of 2005 to Q4 2005, which was about 8.7%. Again, this is related to the types of customers that we're focusing on, not any changing in our pricing strategy.

Our off-net average revenue per subscriber was $795 for Q1 2006 versus $790 for Q4 2005. Virtually flat or a slight increase. The off-net ARPU decline that we had experienced Q3 2005 to Q4 2005 was approximately 5.7%. So we're seeing some stabilization there and probably the slight uptick is related to the size connections of the customers that we have been selling on a going forward basis. Our earnings before interest, taxes, depreciation, as adjusted, were 3.3 million for Q1 2006. An increase of 34% from Q4 2005, where that number was 2.5 million. This EBITDA increase was primarily due to gross margin expansion due to the increase in our on-net business. Now I'd like to turn the call back over to Tad where he can spend a few minutes going over a few more of our financial details.

Thanks, Dave, and again, good morning to everyone, and I also would like to thank and congratulate our entire team on their hard work and performance for this quarter. On loss per share, our loss per share basic and diluted common share was $0.38 for Q1 '06, and that compared to $0.47 for Q4 '05, which was an improvement of $0.09 per share.

During the quarter, as other companies had to, we adopted statement 123R, which is the requirement to expense all stock option grants. That lowered our EPS by about $0.01 during the quarter, and the increase to our net loss was about $130,000. On foreign currency, the euro to U.S. dollar impact did diminish this quarter compared to the impact we had experienced throughout 2005, and the rate change actually positively impacted our Q4 '05 to Q1 '06 comparable revenues but the impact was less than $100,000. Our capital expenditures were 4.7 million for Q1 '06, which is within our quarterly estimate of approximately 5 million. On cash, debt, and our balance sheet, as of March 31, the end of the quarter, cash and short-term investments totaled 21.2 million, and during the quarter we used net cash of 9.3 million, which was consistent with our plan. Included in this amount is an annual capital lease payment for our European fiber in France of approximately $3.7 million, and that's consistent with the contract. No amounts were borrowed under our credit facility, which is 20 million during the quarter.

Our true debt remains at 10 million, and that's our 7.5% Allied Riser notes that are due in June 2007. Capital lease obligations were 89.4 million and 5.8 million of this amount is a current liability. And as a reminder these obligations have a remaining weighted average life of approximately 13 years. Day sales outstanding for worldwide accounts receivable was 39 days, an improvement from the end of 2005 and better than our target of 40 days. So we experienced an improvement in collections and bad debt expense. Operating cash flow was 1.6 million, a use of 1.6 million for the quarter, and that's an improvement from 2.7 million for the fourth quarter of 2005.

On guidance for 2006 we are reaffirming our 2000 guidance that was provided to you in our last call, and as a reminder this includes revenue of between 150 to 155 million. The components of that revenue on net revenue growth of 40 to 45% and off-net revenue decline of 25 to 30%, and a noncore revenue decline of 25 to 30%. Gross margins expected to be in the mid to high 40%, and EBITDA between 20 to 22 million. SG&A expenses expected to be 30 to 33% of revenues, and equity-based compensation expense, including the impact of 123R, between 12 and 13 million. Depreciation and amortization of approximately 62 to 65 million, net interest expense of 10 to 11 million, resulting in an EPS of -- loss per share of 140 to 160, and that EPS assumes 44 million shares outstanding. All of these amounts do not include the impact of acquisitions and do not include any potential financing transactions.

Guidance for Q2 2006 based upon the current run rates of our business and our expectations for sales rep productivity, the increase in that force and marketing efforts and planned network expansion is as follows. Revenue expected to be between 35 and 36 million, the components of that revenue on-net revenue growth of approximately 10%, off-net revenue decline of approximately 10%, and noncore revenue decline of approximately 10%. Gross margins expanding from the 41% to between 42 to 43%. EBITDA growing to between 3.5 to 4.5 million, Q1 '06, and that again was 3.3 million. SG&A between 30 to 33% of revenues. Equity-based compensation expense including option expensing under 123R of approximately 4 million. Depreciation and amortization of approximately 15 million, resulting in an EPS loss of $0.35 to $0.40 per share, and this also assumes 44 million shares outstanding and no acquisitions and no financing transactions. Now I'd like to turn the call back over to Dave.

Thanks, Tad. Again, thank you for your great work as well. I'd like to take a moment and talk about some of the expansion that we have been conducting. We added 13 buildings to our network in the comparable time frame in Q4 of 2005. That number was 14. As of March 31, 2006, we had 1053 buildings directly connected to our network. The break down of those buildings are approximately 830 of those buildings are multitenant office buildings where we serve our corporate customers, and about 220 of the buildings are carrier-neutral collocation facilities. We have today over 35 buildings that are in the process of being connected to our network. We again firmly believe that we will meet our expectations and goals of having 1100 buildings directly connected to our network in December 31, 2006.

As of May 1, we had 106 direct quota-bearing sales representatives. We began Q1 at 88 representatives, and we ended the quarter with 100 direct quota-bearing sales representatives. We do have a goal to be at 130 quota-bearing reps by the end of the year and feel very confident that we will meet that target. We constantly monitor our sales force productivity. We view this as very important in containing our cost of revenue acquisition. Our sales reps added over 5.6 orders per rep, per month for the first quarter of 2006. When these reps are converted to full-time equivalency -- that is, reps that are adding based on a ramp period where they have become fully seasoned and matured, they have produced about 6.3 orders per rep per month. Again, this is in Q1 2006. We had approximately 64 -- or 74, excuse me, full-time equivalent sales reps at the end of Q1 2006. We believe that the additional resources and growth in traffic will help us achieve our business plans and continue to expand our profitability.

With regard to the scale of our network, we today operate now in over 95 markets. We have recently expanded our network to serve Stockholm, Sweden, and have expansions right now underway to serve Milan, Italy, Zurich, Switzerland, and Montreal, Canada. These expansions were all driven by customers that we have in other markets that have requested services in those markets. We today have over 9300 metro fiber miles in our network, and over 22,500 miles of intercity fiber. Our network is very robust in terms of its interconnections with other networks. We are today connected to approximately 1840 other autonomous systems or unique networks around the world. Approximately 380 of these networks are settlement-free peers. Meaning we exchange traffic on an equitable basis with these partners equally with no revenues changing hands. Approximately 1460 networks are, in fact, exchanging traffic with Cogent and paying us as customers.

We believe that our network expansions continue to allow us to have substantial excess capacity in the locations in which our customers are demanding service, and there is the greatest market opportunity. Traffic continues to grow on our network at a pace that exceeds that of the Internet. Our customer connections continue to grow. Our on-net revenues, our growth businesses, continue to increase as a percentage of our total revenue, and most importantly, are growing at an increasing rate. These metrics are resulting in expansions in our gross margins, our EBITDA margins, and I think are a continued validation of the operating leverage of our business. With that, I'd like to maybe open it up for questions and see if we can try to take a few minutes and try to answer questions from the audience.

Thank you. [OPERATOR INSTRUCTIONS] We'll go first to Michael Bowen with FBR.

Good morning Dave. Good morning, Tad. How are you?

Good, Michael. How are you doing today?

Doing fine. So couple questions for you, I was doing some calculations last night that I wanted to see if you had any data for me. I was actually using a little bit lower sales force productivity number, but now that you said it was about 5.6 during the quarter, I was wondering if you -- I came up with a churn rate of about 2.7% for the on-net customers in this quarter versus about 3.1% churn in the fourth quarter, and without getting into all the granularity, I wanted to find out if at least I'm on the right side of that trend. And then the second question I have is, it looked like the ARPU on the off-net business surprised us, and maybe even you a little bit, by stabilizing. Wanted to find out if you could give us a little bit of an idea more of why that occurred and then how we should think about that for the rest of the year in going forward. Thanks.

Hey, Michael. It's Tad. I can give you the churn for Q1 '06. You are essentially in line with what happened. Our on-net churn number was about 2.6%, and as a reminder that includes corporate and NetCentric. We have a much lower churn rate on the corporate side. It's less than 1%. So that averages out. On the off-net side it was close to 4% for Q1 '06.

So we did see some improvement there. To the ARPU question, Michael, we were a little bit surprised by the stabilization in the off-net ARPU. We had in previous calls indicated that we had a one-year cycle to go through and reprice the customer base that we acquired from Verio in December of 2004. And to remind the group this is a customer base that we acquired along with a series of network assets for a cost of $1. We did anticipate some repricing, but it was more dramatic in 2005 than we had initially anticipated, based on market conditions. Now that that repricing exercise is behind us, we have seen more stability.

We've also seen a slight uptick in the number of sales of higher capacity off-net services, and these fall into kind of two camps. One, some T3s, which are substantially more expensive than the T1 product that we sell, which is still the preponderance of our off-net customer base, which I think the increased number of T3 sales did pull the ARPU up a little bit, and we've also seen some uptick in the ability to use ethernet-based services from the RBOC. While their footprint is very limited on the availability of their Gigaman or GigE product, depending whether it be Verizon or AT&T territory we have had some applicability of using that product and therefore had the ability to sell even higher cap off-net services. So I think going forward we should see continued stabilization in that off-net revenue. I think, the trends that I've just spoke about are probably too embryonic for to us predict any kind of increase but I think the rate of decline has pretty much ended at this point.

If I could follow up-with just one other one, if you could maybe break down for us a little bit what you're seeing as far as your penetration rates in the on-net buildings and then also with regard to the data centers and what trends you're seeing there, that would be great.

Sure. Obviously our unit growth in on-net was quite good for the quarter, and we're very pleased with that. We're continuing to add sales representatives that focus on each of these different addressable markets. In the corporate space, we're a little better than 4.3 units per building per month. The average on-net building for Cogent, corporate building, is about 570,000 square feet with a total inventory of on-net square footage of about 470 million square feet, about 9% of the U.S. office footprint is directly on-net for Cogent. In those buildings, there is an average of about 56 unique tenants. Our survey work in canvassing those -- indicates that somewhere between 30 and 35 of those 56 are legitimate prospects, meaning that they buy today a T1 or greater of Internet connectivity. So we believe that we will continue to see expansion in that corporate footprint, and will ultimately reach at least 8 or 9 customers per building, perhaps even more than that over time, but that's kind of what our thinking is at this point in time, based on our penetration rates.

Now, moving over to the data centers, in those centers, again, we're connected to about 220 centers, the centers have about 200 customers per center. There is a great deal of variance around that number. Some centers may have 3 or 400, and some may have only 100 in some of the more rural markets or smaller cities, but on average, there are about 200 logos in each of those data centers. Today we have about 12 customers in each of those data centers. We will expect to see that unit productivity and penetration increase dramatically. However, in those data centers, we have historically focused on the very largest bandwidth consuming customers. So we have a disproportionate amount of bandwidth out of those facilities. In many of those facilities we're north of 30% of the total bandwidth coming out of the facility, yet we may have only 6% of the customers. So this will result in probably some decline in our ARPU in that service provider NetCentric customer base as we focus on some of these less bandwidth-intensive customers in those facilities.

And then the final point about those data centers, the footprint there is a little more elastic in the sense that in many cases, we have actually been able to bring customers to those facilities, bringing a new logo that never existed before there, as customers contact us because of our pricing model, and they may not have yet chosen a location in which they want to house their equipment. We will typically, in a market direct them to one of those carrier-neutral facilities, bringing a new tenant to that facility operator but also a customer to Cogent. So different than the corporate buildings, we actually have some ability to increase the size of that addressable market in that data center footprint.

Okay. Great, Dave. Thanks a lot. We'll see you out on your coming up road show.

Thanks a lot, Michael.

We'll go next to Yergen Yuslin with Wachovia Securities.

Good morning, Dave. Good morning, Pat. How are you guys today.

Great, Yergen.

A couple of questions. One is on the sales of new logo versus existing logo. Maybe you can elaborate more on that as well as the break out between NetCentric customers and corporate customers, where that stands as of today. And also, if I look at your off-net customers I think the decline rate is a lot faster versus the previous past three or four quarters. Maybe you can give a little bit of color on that. Thanks.

Okay, sure. In terms of the decline rate, I'll let Tad take that one first.

The unit decline is included in the metrics and I think it was about 10% quarter to quarter. What we are giving you is the revenue churn number, which did improve quarter to quarter, and that was close to 4% that I cited earlier.

In terms of the breakout between NetCentric and corporate customers, we're continuing to see that breakout be approximately 55% of our revenue coming from our corporate customers, and about 45% of our revenue coming from our service provider customers. So we have not seen any material variation in that ratio over the past year, and as we add additional sales people, we are doing so in both target markets. So we are focusing on both groups.

So you don't expect this to change going forward. What about this quarter? Did you add more corporate customers in NetCentric?

Yes. In every quarter, we add more corporate customers than NetCentric customers by unit count. The corporate customers typically buy a slightly less expensive product. The service providers buy a more expensive product, and that's why the ARPU for our on-net services remains substantially above the product price point for our key corporate product, which is a 100-megabit-per-second for $1,000 a month. So in the corporate space the vast preponderance of our customers buy that product.

We do have some lower end products that are focused on the very smaller customers, a 2-meg product for $500, and a, actually a 500 kilobit product for 249 that is focused on the very lowest end customers. Then with the service provider customers, historically we have focused on those customers that either buy a full gigabit or multiple gigabits. What we are seeing now is we are going to smaller service providers in the data centers. Many of them are buying just the 100 meg product, i.e. $1,000 a month. Or some of them are buying maybe two of those connections, a 200-meg product, and we're selling less of the full $10,000 a month product, the full gig. Again, we do have a few customers who buy a $100,000 connection and a 10-gig product, but those are very rare.

In terms of the first question you asked, which is kind of the mix of new logos versus existing customers, again, we continue to see that split being roughly 70% of all sales each month are to brand-new logos who have not bought from Cogent previously. About 30% of our sales are existing customers. Now, for those existing customers, the corporate customers are typically buying additional locations as opposed to buying greater bandwidth at an existing location. For the service provider customers, who are included in that 30% that are increasing their spend with Cogent, they are typically buying a larger connection. So the buy dynamic is slightly different in the two universes, but about a third or 30% of our sales each month are to existing customers. And again, as a reminder, we do not have a suite of value-add products. We believe that our core competency is at the network layer, and, therefore, have allowed many of our customers to add the application layer. So different than other service providers, we are not deriving more revenue by selling less margin intensive products to those customers, but rather we are -- the 30% of our sales that are coming from existing customers have identical margin characteristics to our new sales and are really allowing us to leverage our network.

Okay. So if I can ask again on the off-net customers, the decline rate about 10% quarter over quarter. Where do you expect that to go going forward?

Yergen, we don't expect a material change between what we experienced in the first quarter regarding the revenue churn. The unit churn number that you see is the -- also the net change quarter to quarter but what we experienced this quarter appears that the Verio customer mix has stabilized, so where we are and what we experienced in the first quarter is kind of what we expect to see going forward.

Okay. Thank you.

Thanks a lot, Yergen.

We'll go next to Nick Netvulidoff with Lehman Brothers.

Hey, guys. How are you?

Good, Nick.

Congratulations on the quarter.

Thank you, sir.

My first question is just a clarification on the 3.5% off-net -- or the 2.6% on-net churn and the 4% off-net churn that you quoted before. Is that customer or revenue?

Revenue.

Revenue. And then as far as the off-net is concerned what portion of your off-net business has an on-net connection? And is that what's driving the high cap both the ethernet based and the T3 based business that you're getting, is it on-net customers signing up for an off-net connection with you because of your backbone?

Yes, I'm going to define our off-net business into two buckets, Nick. Roughly 90% of the off-net connections that are in our installed base today were acquired, not organically sold by Cogent. So they came to us through some acquisition, and most likely had no linkage to our existing on-net business. In all of the acquisitions that we have done, we have been able to convert about 3% of the customers acquired who were off net, into on-net. We were just lucky enough that they fell into a footprint that we had facilities to. But the vast majority of these acquired customer bases were kind of autonomous off-net customers. On a going forward basis, however, as we sell new business, approximately 80% of our sales productivity is focused on our on-net business.

Sure.

20% is in the off-net business.

Right.

For that 20% that is in the off-net business, about half, or 50% of new sales of off-net services, are to someone who also has an on-net connection with us. Now, those customers who have an off-net connection have already made the decision that they chose Cogent in large part because they have a business need for the higher capacity bandwidth.

Right.

If those higher capacities are commercially available, many of them have elected to buy those, we are somewhat constrained by the outside plant limitations of the incumbent providers. We buy off-net facilities from nearly 60 different providers, many of them competitive, but the reality is the underlying last mile facilities are almost always that of the incumbent operator, the ILAC. And in many cases T3 or ethernet services are just not available. But in the opportunities where they are available, we do propose those to the customers and have had some success rate with them. And, therefore, I think that's a partial answer to the increase in off-net, ARPU.

Right.

And it's also being helped by the fact that the repricing has been fully mitigated with the one-year contract cycle.

Right. And can I assume that the churn for those off-net customers that have an on-net connection is lower?

We have not tracked that historically as a separate data point.

But your gut.

Intuitively, you would believe that to be true, but I don't have any empirical data.

Then one other question. You've maintained your on-net building goal of 1100 by the end of the year. Do you expect to change that after the offering?

The answer is yes. We have specifically issued a press release describing our proposed offering. In that offering, there are two primary uses of proceeds, one being to add additional sales people, and again, we're very encouraged by the productivity of the sales force and will continue to add people, because we have so much underutilized or baking capacity in our network that the operating leverage dictates that it makes sense to add sales people provided the cost of revenue acquisition doesn't go up.

Secondly, we do believe that there is an ability to selectively identify additional buildings in which our footprint passes in front of that meet all of our criteria. We post all of our buildings on our website, and people can go in and look, and I think you can see that we've been focused on some very high-profile additions recently. I think as we raise capital, it's reasonable to expect that we will add -- continue to add buildings in 2007, probably at the pace of about a building a week that is our intention.

In 2007?

That's correct. Okay. Great, guys. Thanks a lot. Thanks, Nick.

We'll go next to David Barden with Banc of America Securities.

Thanks for the call.

Thanks, Dave.

I just had a couple of questions. One, I think we've been talking a lot about the NetCentric versus corporate user base. If you could give us, if you're sharing that, a revenue breakdown between the two groups, I guess last year was about 46% NetCentric. The second question would be just in the carrier hotels, Dave, a lot of the bigger players have been trying to assert or message to people that the prices are starting to level out. Are you seeing the rate of narrowing of the gap between yourself and your pricing at the $10 per megabit level and other carriers kind of stopping its convergence, if you will? And I guess the third one is, we've talked a little bit about getting the handle around the churn metrics. What is the primary reason why churn is occurring at the rate at which it is at the 4% level? Is it companies going out of business? Are they changing carriers because of price or some other reason? What is the driver of that? Thanks a lot.

Okay, great. Dave. I think I'll try to grab those in reverse order. Starting with the churn in the service provider space, what we have seen is really, I think three key drivers there. One, there is a higher propensity for those entities to have either payment problems or bankruptcy because our service provider definition is fairly broad, and it includes e-commerce companies, content companies, as well as distribution networks and ISPs, so we have seen a greater need to turn people off for nonpayment. As Tad indicated. We've had a lowering in our total bad debt and our collections have improved, but we still monitor that.

Secondly, we've seen, I think, continued consolidation in our industry, which has resulted in just less people, sometimes we've had situations where two or three customers who each bought independently from us merged and we then had one customer, in some cases one much larger customer buying more. In other cases each of them had some wastage, and they were able to groom that and therefore had less unused capacity.

Then the final point is, in the service provider space, we maintain a very vigilant view in terms of our acceptable use policy. This is standard in the Internet industry where service providers on their websites post what is their acceptable use, and if a customer violates that, an example would be someone who is accused of running spam or harboring DOS attacks on someone else, it is necessary for us to terminate that customer because they then have the ability, if we did not terminate them to pollute the IP address space of innocent customers, and therefore perhaps create a scenario in which those innocent customers are blocked by fire walls due to the fact that they're associated with a spammer. So it's sometimes unfortunate that you've got a great paying customer who has a business model that unfortunately makes them kind of a bad net citizen, and we have to terminate them. So I think for those three reasons we will continue to see that churn rate among the service provider universe to be much higher than the Corporate end user churn rate for our on-net business, which is sub 1%.

To the second question, which is the ARPU question -- well, actually, the pricing question, over our entire history we have been very steadfast in our pricing model at a $10 a meg price point. When we entered the market, the market average was obviously substantially higher than it is today, probably north of $300 a megabit. Today that market average has fallen to about 50, representing about a 50% per year price reduction among our competitors. Throughout that entire period in the data centers, our price point has remained constant at $10 a megabit. We have seen no reason to drop price or to raise price, for that matter. We have seen some mitigation in our competitors and their willingness ton continue to drop prices to match our price point. I think that's been reflected in our continued gain in market share.

What we don't have visibility to is why they are not dropping prices. It may be that they're unable to based on their cost structure, or they are unwilling. But we feel that the average 5-to-1 cost advantage that we have is still substantial and puts tremendous pressure on the market. And it's been demonstrated by the maintenance of our revenue per employee, per sales person and sales productivity remaining very high and constant, even though we've been adding a significant number of sales resources, increasing our sales force by 2.5 fold in the past five quarters. So we maintain a commitment to our pricing strategy. Today we are substantially below the competitors. We have seen them slow the rate of decline. If, however, they accelerated it, we may then review our pricing policy and decide to reduce prices even further, but at this point we've seen no reason to do that.

And then finally, in terms of the breakdown of corporate versus NetCentric revenue for the quarter what we saw is about 48% of revenue sales were corporate, and 52% came from our service provider customers. By unit, roughly 70% of unit sales for the quarter were corporate, and about 30% of unit sales were service provider or NetCentric, and again, that's kind of demonstrating the fact that the service providers bay fewer but buy much larger connections. Again, the overall mix remains pretty close to that 55/45 that we've been speaking about.

Thanks a lot, Dave.

Thank you, Dave.

We'll go next to Daniel Furniture with Tier One Research.

Thank you. I have a question, I think it's just a clarification of questions already asked. To some extent Cogent is still a transformation story as we watch on-net go up, off-net go down, and the sales team grow. Do you have a sense of what the steady state will be in those, like and ratio of on-net to off-net at steady state and where you think you'll have enough sales people?

Sure, Dan, this is Dave. Two points. First of all, in terms of steady state I think we have a great daily of visibility to that. I think we will eventually get to a revenue break down of 80% of our revenues coming from our on-net business, 20% from our off-net, and none of our revenues coming from noncore products, since we don't sell these products and they are churning, as Tad said, at about 10% a quarter. So we eventually see that revenue going away. We do sell off-net services, it's about 20% of our revenue sales going forward, 80% being on-net, so I think that's through steady state. When we get there is probably a couple of years out. But I think our guidance again has said that we should be well north of 70% of our revenues coming from our on-net business by the end of this year.

With regard to the number of sales people that are appropriate for when we will reach saturation, we have a huge addressable market in front of us. We continue to add sales people at a very measured pace. We constantly look at that unit productivity and our cost of revenue acquisition. We've been very measured in the way in which we've added those people, and again, we've indicated a clear plan to get there to 130 by the end of this year, and we think that's a prudent plan, considering that we're managing out the underperformers on a very regular basis. I think we'll continue that growth strategy into 2007. We'll keep adding at about that same pace, and we'll evaluate along the way constantly what is the appropriate number. But today, as long as that productivity remains high, we're going to keep adding people.

Okay. Got it. Thanks.

Okay, thanks, Dan.

We'll go next to Colby Synesael with Merriman.

I just want a clarification. You guys were talking about competition earlier and how you see some of your competitors moving more up the -- I guess the OSI stack towards the application layer. I think you mentioned that that was a lower margin business and that's why you guys aren't really going to do that. Can you guys talk a little bit about that margin? It's been my impression that adding more services to either the basic network would actually improve your margin.

Sure. First of all, when you look at the network, it is effectively a fixed cost asset, and our utilization rates are exceedingly low, kind of in that 5% range, so we have a great deal of surplus capacity in the network to add. When we add on-net revenues, we're getting about $0.95 of each marginal dollar flowing through to EBITDA. That obviously gives us tremendous operating leverage. There are virtually no service or application layer businesses that have that type of incremental margin contribution. There's either direct cost of goods sold or in most cases, these applications are very labor intensive, and, in fact, can be very custom in that IT is very customer specific and may not be able to be productized as efficiently as network capacity. So I think when other service providers have indicated that they are moving to higher value services up the OSI stack what they are really saying is they either don't have the operating leverage in their network that we have, they may either be out of capacity or have a higher cost of goods sold, when adding traffic to the network, or they believe that the network business is only a inducement to get someone into some other business that will be a much larger business than their network business.

Our strategic asset is the network, and with so much vacancy in that network, we need to focus our sales resources and our operational resources on increasing the utilization of that network with the margins that we have. When you move into things like firewall or managed security or managed IP VPNs, they are more labor intensive and will require more capital expenditures, and most importantly the sales cycles are longer and more complex. Bandwidth is easy for people to understand, and allows our sales force to be exceedingly productive. So I think a big part of our success has been the fact that the sales force is able to give a clear message to the customer and close the sale quickly.

Have you guys -- you kind of bring up a point though, I mean when you look at the business, I think from an investment standpoint, everybody's looking, I guess the bottom line would be free cash flow. Have you guys done any type of analysis that would say on an absolute basis we might lose a little bit of margin, but we gain that in the sheer amount of sales that we'd be getting by selling more applications on top of our basic service, or do you think that in the long run, sticking to the model that you have today, it might take a little bit longer to approach those levels but you will have better margin at the end of the day?

First of all, I want to reiterate what we have said before, which is we expect to be true free cash flow by the end of this year on a quarterly basis, which I think is an admirable goal and one that we can absolutely achieve. As we look at the distractions to our ability to leverage the asset we have, we think that moving into these other businesses would require a lot of management resources, a lot of reengineering of our processes, and may, in fact, retard our ability get to cash flow break even sooner. Many companies suffer from the fact that they try to be all things to all people, and I think for us, with so much idle capacity and so much competitive advantage in our network business, that's really what we need to be leveraging and focusing our resources on.

Okay. Then my last question, you talk about a better way to utilize your network is obviously to grow your sales force. Can you talk a little bit about the value-added third party sales channel and what you guys are seeing there, and what the strategy might be to utilize that going forward?

Sure. There's kind of two parts to that question, Colby. First of all, the vast majority of Cogent's sales are direct sales initiated by our sales team. However, we do have a limited agent program that accounts for a very small percentage of our sales, probably 5 to 7% of our sales come through agents. We encourage an agent channel. However, with a transparent pricing policy, such as the one we have adopted, our brand is almost synonymous with our price, it's so well-known in the market, does it make it difficult for people to add mark-up or margin on top of our product. So we don't have a traditional reseller base that others have who have more opaque pricing strategies.

However, to the other part of your question, which is value-added services, when we talk about our service providers, they're, in fact, buying our bandwidth and adding value to it and selling a different product to their customer, perhaps at a huge increase in margin. And that value add can come in really two different dimensions. One can be they are basically parsing our bandwidth in a smaller increment and distributing it over their own networks to large number of end users. We see that constantly with our cable companies, our PTTs, and our CLAC customers who all buy bandwidth from Cogent, then productize that to their retail end users, and that is how we benefit from the uptick in consumer demand, not just business demand.

On the application side, we have over 50 VoIP carriers. We have security companies, we have video distribution companies, conferencing companies, IP VPN operators, firewall and security companies, remote storage companies, all as Cogent customers. Each of them are buying our bandwidth, adding an application to it, and selling that bundled product to their end customers. They are then creating value and have their own value prop to win those customers. We get the best of both worlds. We get the application, but we don't have the costs associated with it, and that we can then sell more bandwidth. We think that's how we best leverage the asset we've built.

Thanks, Dave.

Thanks.

We'll take our last question from Arian Monart with Credit Suisse. Ms. Monart, your line is open. Please go ahead.

Hello. Can you hear me?

Absolutely, Arian.

I apologize for my voice. Two questions. [Inaudible] can you explain why? Also SG&A didn't go up very much, you had 12 more -- quite a few more sales people. What should we expect for SG&A, if you do increase to 130, and is that in your guidance? And finally, the offering, is the plan to increase the number of sales people beyond what you've just mentioned? Thank you.

Yes, sure. Regarding SG&A from the change from the fourth quarter to the first quarter as I mentioned on the call we did experience a lower rate of bad debt expense due to increased customer collections and a drop in our days sales outstanding. We also had a reduction in professional fees because we had Sarbanes-Oxley fees in the fourth quarter of '05. Did have some of those fees in the first quarter of '06, but they declined. And those amounts offset the increase in compensation from the increase in headcount regarding the sales force. The guidance is for SG&A expense to be 30 to 33% of revenues for '06.

Okay. And will the offering -- I mean, the primary piece of the offering also be applied to increasing the number of sales people beyond the 130 you mentioned, or you still mention -- still targeting 130?

The answer is, we are looking to do 130 by the end of the year. It is our intention, however, that we will probably continue to increase that throughout 2007, and what that final number is we don't have perfect visibility to yet because we are going to continue to monitor the productivity, but we absolutely expect to continue to increase the sales force beyond 130, and our best estimate at this time would probably be another 50 people in '07, but I want to couch that as a very preliminary estimate because we are going to continue to monitor the addressable market and the efficiency, but that is -- the two primary uses of proceeds from the proposed offering that we have indicated we are going to commence on are, one to add additional sales people beyond the 130. But to an indeterminate amount at this point in time. And then secondly, giving us the capital to continue to expand the footprint beyond probably the 500 million square feet or so of corporate space we'll end the year with in the data center footprint and continue to add buildings selectively to the network at the pace of one per day. Per week.

Per week. Excuse me, I'm sorry, one per week.

Thank you. One other little thing. You mentioned the real debt being 10 million. So is it because the rest -- what's in the 10-Q -- [Inaudible] could you explain that a little bit? Because it impacts your valuation obviously.

Sure, yes. The true debt is the Allied Riser notes which, again,are the 10 million due in 2007. It showed on the balance sheet that it's a discounted amount that will accrete up to 10 million in June '07, and I think that's 7.2 million at the end of March. The remaining long-term liabilities are the IRU obligations and because of the term of those leases they meet the U.S. GAAP test to be recorded as capital lease, so you put the assets recorded in PP&E as an IRU asset, and the debt obligation at its present value is put on the balance sheet and as those payments are made you charge interest expense and relieve the obligation. As I mentioned earlier the average life of those obligations is about 13 years going forward. Remaining life.

So the 7 million interest expense includes those IRU payments?

Yes, does it.

And so the debt itself should not be -- the 85 million -- 83 million in your Q, we should not count that as debt, assuming you keep paying your interest expense?

Under GAAP, we have to count it as--.

I'm only hesitating because I think we're just stuck with vernacular here. They are a true debt that I would consider a debt obligation as the Allied Riser notes. The others are long-term leases that are recorded as liabilities. I agree with you, they would not be true debt.

Thank you.

Very good. I think we may have time for one last question. I saw there were still a number of people on the monitor. We could take one more maybe and then wrap it up.

We'll go to Steve Brown with Hergemann Asset Management.

Good morning, guys. Quick question on currency assumptions and changing and how it impacts numbers moving forward and what assumptions you made.

Okay. Well, first of all, I wish I had a crystal ball to tell what you currencies are going to do. As Tad indicated the impact this quarter was about $100,000. In each of our Qs and in each of our earnings releases we are trying to be totally transparent about that impact. I think we've been very realistic in looking at the current market price of the dollar versus the multiple currencies which we sell services under, and we think that we baked reasonable assumptions into our model, but obviously if there are radical movements in currency, we're going to see some fluctuation, either positively or negatively, our plan is based on the euro to dollar at 1.20, and the Canadian dollar at $0.86. I mean, these are approximately correct today. Obviously I think the dollar is a little weaker than that. Bit we don't modify our plan on a daily basis as currencies fluctuate.

Great.

Thanks, Steve. Again, thanks, everyone, for attending the call today and their interest in the Company, and as always, management remains open for follow-on questions, and again, we thank everyone. Take care now.

Thank you, everyone. That does conclude today's conference. You may now disconnect.