Cogent Communications Holdings Inc (CCOI) 2009 Q3 法說會逐字稿

Good day, everyone, and welcome to today's Cogent Communications Group third-quarter 2009 earnings conference call. Today's call will be available for replay after the conclusion of the call at www.cogentco.com. As a reminder, today's call is being recorded.

At this time I would like to turn the call over to Mr. Dave Schaeffer, Chairman and Chief Executive Officer of Cogent Communications. Please go ahead.

Thank you and good morning. Welcome to our third-quarter 2009 earnings conference call. I'm Dave Schaeffer, Cogent's Chief Executive Officer. With me on this morning's call is Tad Weed, our Chief Financial Officer.

We are pleased with our results for the quarter in an improving but challenging economic environment. During the quarter traffic on our network grew by approximately 15% sequentially quarter-over-quarter, and our revenues grew at a faster rate than most of our competitors. We continued to be encouraged by the increase in traffic on our network, as well as our revenue growth.

We continue to be optimistic about our outlook for the remainder of 2009 and beyond. During the quarter we again significantly expanded our footprint by adding 32 on-net buildings to our network and additionally by adding over 3000 route miles of both metro and intercity fiber to our network. We achieved another major corporate milestone in generating a quarterly operating income for the first time in our 10-year history.

As we discussed in previous earnings calls, in 2008 we extended a volume and term discount program to new Cogent customers, as well as to our existing net-centric customers if they increased their total contract value with us.

During this quarter this program continued to be successful. In the quarter this program resulted in over 716 net-centric customers representing approximately $10 million of remaining contract value, increasing their remaining contract value by an additional $10 million or 100% increase. Additionally, as more customers extend their contract terms and show their confidence in Cogent, our average contract length has continued to increase this quarter on a quarter-over-quarter basis by another 2.6%.

Throughout this discussion we will highlight several operational statistics that we believe demonstrate our increasing market share, expanding scale and operating leverage of our business. I will review in greater detail the operational highlights of our continued expansion plans. Tad will provide additional details on our financial performance, and then following our prepared remarks, we will open the floor for questions and answers.

Now I would like Tad to read our Safe Harbor language.

Thank you, Dave, and good morning, everyone. This third-quarter 2009 earnings report and 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. 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 the potential impact of mergers and 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, you'll find these reconciled to the GAAP measurement in our earnings release and on our website at cogentco.com.

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

Thanks, Tad. Hopefully you have had a chance to review our earnings press release. As with previous quarters, our press release includes a number of historical metrics. These metrics will be added to our website. Hopefully you find the consistent presentation of these metrics informative and helpful in understanding the financial results and the trends of our operations.

Our third-quarter 2009 revenue was $60.2 million and represented an increase of 3.9% from our second-quarter 2009 revenues of $58 million. Our revenue growth rate was significantly greater than the announced growth rates of most of our competitors. In fact, many of our competitors continue to see sequential declines in their revenues. We evaluate our revenue based on product class -- on-net, off-net and non-core, which Tad will cover -- and we also evaluate our revenues by type of customers. Our customer types include two major groups, our net-centric customers and our corporate customers. Net-centric customers buy large amounts of bandwidth from us and data centers. Corporate customers buy bandwidth from us in large multitenant office buildings.

Our corporate customers at the end of the quarter represented 58.5% of our total number of customer connections. Corporate customers represented 47.2% of our revenues in the third-quarter 2009, and revenue from our corporate customers grew at 2.2% sequentially from the second quarter of 2009 to the third quarter of 2009. Our net-centric customers represent 41.5% of our total customer connections at the end of the third-quarter 2009. Revenue from our net-centric customers represents 52.8% of our revenues for Q3 2009, an increase of 5.4% from the second quarter of 2009. Fluctuations in foreign currency exchange rates materially impact our net-centric revenues. About 30% of our revenues are derived from international operations.

From Q2 2009 to Q3 2009, fluctuations in foreign exchange positively impacted our revenues by $880,000. On a comparable quarter, however, looking at Q3 2008 to Q3 2009, fluctuations in foreign exchange actually resulted in a reduction of revenue of approximately $930,000.

Now for some overall highlights and trends, pricing and provisioning in our network. Pricing for our most widely sold corporate product remains at $900 per month for a 100 Mb connection or $9 per megabit. We continue to offer discounts related to term contract for all of our corporate and net-centric customers.

We also offer volume commitment discounts for our net-centric customers. As I mentioned earlier, over the quarter 760 existing net-centric customers took advantage of these volume and term discounts, as well as many new customers. These discounts for net-centric customers contributed to a 7.3% reduction in our average price per megabit of our installed base. In Q2 2009 that number was $6.82. In the third quarter of 2009, that number was $6.33. The average price for a new megabit of sold services in the third quarter of 2009 was $4.82.

One of our competitive advantages is our ability to rapidly install our on-net orders. We guarantee the installation of any on-net service in 17 days or less. In the third quarter, our provisioning team consistently outperformed our guarantee and averaged slightly less than 10 days to install new services. On-net ARPU declined, however, at a much lower rate than we have experienced in previous quarters. On-net ARPU declined 2.8% from the second quarter of 2009 to the third quarter of 2009 due to increased contract length, net-centric volume discounts and partially offset by the foreign exchange impact. Our on-net ARPU was approximately $1010 in the second quarter of 2009 and $982 for the third quarter of 2009. Our off-net ARPU continued to increase and increased by approximately 8/10 of 1% as the size of our average off-net customer connection continues to increase. Off-net ARPU increased from $1118 in the second quarter of 2009 to $1127 in the third quarter of 2009. Substantially all of our international revenues are on-net, so no variations in foreign exchange impacted our off-net ARPU.

Now to discuss churn. Our on-net churn rate actually improved to 2.9% for the third quarter as compared to 3.1% in the second quarter of 2009. As a reminder, we consistently report our churn numbers on a gross basis. As a result, as a customer remains with us but amends their Cogent contract, they are counted in our churn and are included in this churn rate. (inaudible) and changes are included.

Our off-net churn rate also improved to 3% for the third quarter of 2009 as compared to 3.2% in the previous quarter.

Our churn rates have been impacted by our continued stringent monitoring of our customers' collections and credit quality. Traffic on our network continued to increase and increased by approximately 15% quarter over quarter. The increase in traffic was substantially greater than the 10% increase that we experienced Q1 '09 to Q2 '09 and much better than the comparable period last year when, in fact, our traffic grew Q2 2008 to Q3 2008 by 5%.

In these cost-conscious times, we believe that as a low-cost provider, Cogent has a distinct advantage. We continue to anticipate growth in network traffic as more and more of our customers take advantage of our value proposition.

Before Tad provides some additional details on the quarter, I would like to address our results against our long-term revenue targets and expectations. As a reminder, we have previously announced we expect topline revenue growth of between 10% and 20% annually. For the third quarter, our revenues grew at 3.9% quarter over quarter or approximately a 17% annualized growth rate. This is well within the range that we have forecasted.

Now I would like to turn it over to Tad to provide some additional details on the third-quarter 2009.

Thank you, Dave, and good morning to everyone. I would also like to thank and congratulate our entire team on their hard work and efforts during the third quarter. I will begin with providing additional details on our revenue by product class, which is on-net, off-net and non-core.

With respect to on-net revenue, our on-net revenue was $48.1 million for the quarter and increased 3.4% from the second quarter. About 75% of our new sales in the third quarter were for our on-net services. Our on-net customer connections increased by 645 customer connections or by 4% from the previous quarter, and we ended the quarter with about 16,600 on-net customer connections on our network. Our revenues from our off-net business were $11.1 million for the third quarter and increased 5.3%. Off-net customer connections were flat at about 3300 customer connections. Our non-core revenues were up about $1.1 million and represent less than 2% of our revenues.

EBITDA and gross margin. The operating leverage of our business continued to result in healthy gross and EBITDA margins. Our gross profit margin, however, declined by 150 basis points from the quarter from 57.7% from the second quarter to 56.2% for the third quarter. The decline was primarily due to an increase in investment in our expansion activities, including an increase in seasonal power costs.

EBITDA as adjusted was $17 million for the quarter and increased 2% from $16.7 million for the second quarter. Our EBITDA margin declined by 50 basis points. The EBITDA margin was 28.2% for the third quarter and 28.7% for the second quarter.

Quarterly lumpiness and the expansion of our gross margin and EBITDA margin can and does occur. If you examine our 18 quarters of quarterly metrics, you will notice occasionally lumpiness in our quarterly gross margin and EBITDA margin expansion. This lumpiness can occur due to seasonal factors such as the timing of the performance of certain professional services, including audit and tax services and more importantly to the timing of our expansion activities. However, our long-term trend is for increasing gross margin and EBITDA margin expansion, and we expect that long-term trend to continue.

Earnings-per-share. Our basic and diluted loss per share for the quarter was $0.07. This was a significant improvement from the loss of $0.10 per share for the second quarter. Dave touched on foreign exchange. Some additional details. Approximately 30% of our business is outside of the US. For the third quarter of 2009, about 22% of our revenues were based in Europe, and about 7% of our revenues were related to our Canadian operations. These percentages are similar to the percentages for the first and second quarter of this year.

Continued volatility in the year on Canadian dollar to US dollar conversion rates impact our comparable quarterly revenues and financial results. The impact on our revenues was an increase of about $880,000 from the second to third quarter of '09. However, if you look at the comparable quarters for 2008, it was actually a decrease of revenues of $930,000. The euro to US dollar rates were for the second quarter $1.36 and it increased to $1.43 for the third quarter. If you look back to the third quarter of '08, the rate was higher at $1.50. That Canadian dollar rate was $0.86 for the second quarter and increased to $0.91 for the third quarter, but again if you look back to the third quarter of 2008, it was $0.96.

Capital expenditures. Our capital expenditures were $16.7 million for the third quarter, an increase from the $13.4 million for the second quarter. On a quarterly basis, we can and have historically experienced seasonal variations in our CapEx and our construction activities. Our quarterly CapEx is in part dependent upon the number of buildings we connect to our network each quarter and to the timing and scope of our expansion activities, including adding route miles for the network. Typically we experience our lowest level of CapEx in our fourth quarter, and we expect that to continue.

We added another 32 buildings to the network in the quarter, which was similar to the 34 buildings we added in Q2. However, we also added over 3000 metro and intercity fiber miles to our network during the quarter, which is a material increase. We now expect to add the previous addition was about 100 buildings to the network, and that has increased to about 120 buildings to our network and further increase our fiber miles for the remainder of 2009.

As a result, our 2009 CapEx will be greater than our 2008 CapEx, and the increase is due to an acceleration of our expansion plans and expansion opportunities, including the impact of adding more buildings to the network. Cash, debt and the balance sheet. As of September 30 quarter-end, our cash and cash equivalents were $51.1 million. We have about $92 million of the original $200 million of face value of our convertible notes remaining, and those notes mature in June of 2027. The notes are reported net of their discount on our balance sheet at $65.1 million.

Capital lease obligations were $110 million at September 30, and about $10.8 million of that is a current liability. As a reminder, these obligations are entirely for long-term dark fiber leases and are being paid over a remaining average period of more than 10 years.

In October we entered into a $20 million revolving credit facility that may be used for general corporate purposes, acquisitions and/or share and note purchases. We have not borrowed under this facility. We have $30.1 million of authorized purchases remaining under our existing stock purchase program, and we did not purchase any shares during the quarter. Days sales outstanding for accounts receivable was 33 days at September 30, 2009, which was significantly better than our targeted rate of 40 days. However, it did increase from our Q2 DSO of 27.

I want to personally again thank and recognize our worldwide billing and collections team for continuing to do a fantastic job on customer collections and credit monitoring. In this economic environment, we will continue to closely monitor our credit standards and our Accounts Receivable. While the DSO did increase, bad debt expense declined as a percentage of revenues. Bad debt expense was less than 2% of our revenues for the quarter, an improvement from the 2.4% of revenues for the second quarter.

Increases in our average contract term, which Dave mentioned, increased about 2.5% from the second quarter, and our credit monitoring activities have an impact on our future revenue performance under US GAAP. For example an increase in our average contract term and estimated customer life results in a longer period over which to recognize nonrecurring installation revenues. An increase in the amortization period results in a reduction in the recognition of revenues under US GAAP. Additionally closer credit monitoring activities can result in the rejection of certain orders or the termination of an account.

Operating cash flow. Cash flow from operations was $14.8 million for the third quarter, an increase of 13% from the $13 million for the first quarter of -- second quarter rather of '09.

Operating income. Finally, as Dave mentioned for the first time in our 10-year corporate history, we achieved positive operating income of $458,000 for the third quarter. One of our next corporate financial milestones will be achieving net income and positive EPS.

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

Thanks, Tad. Let me take a moment and talk a little bit about salesforce activity and productivity. We began the third quarter of 2009 with 230 quota-bearing reps and ended the quarter with 250 quota-bearing sales reps, an increase of 20 in the quarter. We hired 76 reps in the quarter, which is the most we have ever hired in a single quarter. 56 reps left the Company during the quarter. This number was similar to the 57 reps that had left the Company in the second quarter of 2009.

Our rep churn rate for the quarter improved slightly to 7.8% from the 8% in the previous quarter. We began the third quarter with 214 full-time equivalent sales reps, those reps that have ramped to productivity and ended the quarter with 215 full-time equivalent sales reps. Productivity on an FTE basis for the third quarter was 3.8% per FTE per month. The performance was relatively in line with our organic productivity rates.

As a reminder, our sales rep productivity rates are not based upon contract signing but rather are based on delivered and installed services.

Now to our network scale. We added 32 buildings to the network and have 1421 buildings directly connected to our network at the end of the third-quarter 2009. In 2009 we now expect to increase our target for a number of buildings to add to 120 for the year, approximately 20% more than we had originally projected. We now have over 12,900 miles of metro fiber and over 40,000 route miles of intercity fiber on our network, an increase of over 3000 miles from the second quarter of 2009. We continue to evaluate additional dark fiber acquisitions for our ability to expand our network both in Europe and in North America.

Cogent remains one of the most interconnected networks in the world. This is critical because we sell Internet connectivity, and that service is dependent on our connections to other networks. Today we are directly connected to over 2840 other networks. We believe that our network has substantial available capacity to accommodate additional growth. We are currently utilizing approximately 19% of our lit capacity in our network.

In summary, traffic continues to grow on our network, and revenues grow as well. In fact, we are growing both traffic and revenue faster than the market and our competitors, demonstrating our increase in market share. We believe that Cogent is the lowest cost provider, and it is our value proposition that is unmatched in the industry. Our pricing program has attracted many new customers. Existing customers have increased their average volumes and contracts with us, and our ability to grow revenues continues. Our business remains completely focused on the Internet and, therefore, provides a necessary utility to our customers as the Internet becomes more critical to our customers' businesses and their lives.

We have a strong balance sheet compared to others within our industry. We are putting some of our cash flow to work to expand our footprint and expand the number of markets we serve, as well as expand our sales force. We continue to be encouraged by the results of our sales initiatives and our ability to attract high-quality sales representatives. We are committed to providing top-line revenue growth of between 10% and 20% annually and to expand our both gross and EBITDA margins.

With that, I would like to open the floor for questions.

(Operator Instructions). David Dixon, FBR Capital Markets.

So just a question on the pricing environment. If you could give us a sense of how we are positioned today in the net-centric space? Specifically some of the [checks] have been hollering some more aggressive pricing there, as well as some updates this quarter in terms of benefits that we are seeing from some of the content players in the market.

And then secondly, if you could perhaps just delve into the lowest price guarantee a little bit more for us. I wanted to understand the mechanics of this. And, in particular, trying to get a sense of the extent to which you have some flexibility there to reject some of these lower prices on the basis of non-comparable networks and how that is determined. That would be very helpful. Thanks so much.

Sure. So the net-centric market is one in which we and other providers are selling a commodity-based service. We offer the greatest amount of connectivity and capacity operating our network on a non-oversubscribed and non-balked basis. We have a standard pricing policy that prices, services anywhere from $4 to $10 per megabit depending upon contract term and length. Occasionally we do offer special promotional programs either on a given period of time or in a given geographic region. We remain the lowest cost provider. We believe that our price at $6.33 for our installed base is at about 50% of where the market is.

In addition to that, our average new sale in the quarter at $4.82, we believe, is less than 50% of the average price in the market.

We have seen situations a couple of times in this most recent quarter where we have offered lower prices in response to discounts from others. We have a policy where if a customer shows us an invoice from another Tier 1 network operator, some other network operator that is offering global connectivity, we will beat that price, and we did that a couple of times in the quarter. We have actually seen a decline in the rate in which we have had to offer those very aggressive prices. If, in fact, a customer provides a price and it is not documented, we will push back on them and request that they provide us documentation under a nondisclosure agreement because we have no intention of being in the business of negotiating against ourselves.

And then finally, we are looking for networks that can provide the service. So it really comes down to the locations in which the service is requested and the scope of the network. But we remain the lowest cost provider. And, in fact, we have seen less competition as we have seen the number of companies that we actively compete with continue to decline.

And just to clarify then, that is where they would have to show an invoice from a Tier 1 global provider?

That is correct, David.

Jonathan Schildkraut, Jefferies.

A couple of things here, Dave. First, this was a pretty nice uptick in sequential traffic growth, and I was wondering if there were certain things that drove the traffic to expand at the rate that it did whether it was seasonal, whether we are seeing some secular changes, uptake of online video or whether this was just kind of a continued market share shift.

Beyond that, if you could also talk about the acceleration in the number of buildings that you're adding to the network, I'm trying to get a sense as to whether that is being driven by opportunity that you see out in the market or whether you feel like you need to accelerate new business locations in order to maintain your growth rate.

First of all, with regard to our rate of traffic growth, I think seasonally we do see acceleration in the fall and winter months. But, as I stated last year, that acceleration resulted only in a 5% sequential growth now off of a much larger base we grew at 15%. I think this is a result of some of the take or pay contracts that we had put in place earlier in the year. It also is a result of some business models getting real traction and, therefore, customers requiring a lot more bandwidth. We are continuing to see a proliferation and acceleration in the growth of various video business models.

The Internet is morphing from a symmetrical communications network to one that is delivering video competing with broadcast and cable networks for the distribution of that video content. We are seeing significant pickup from a number of customers that have various video business models, and we are still at the early stages of that growth. And I think the fact that our traffic grew at 15% and we actually saw an acceleration in the rate of growth in our net-centric revenues as well shows that we are able to grow our revenues, even in an environment where there is volume-based pricing pressure.

We will continue I think to see further acceleration in traffic growth in the fourth quarter much as we did last year in the fourth quarter as well as that tends to be seasonally probably one of our strongest quarters historically.

Now to the building growth. We have chosen opportunistically to take advantage of the dislocations in the market both from suppliers of dark fiber, as well as from construction companies that help us connect our network to the end buildings. We chose to plow back in some of our additional cash flow to accelerate the growth in both of those areas. We added over 3000 route miles to our network. That is about a roughly 6% expansion in the network just in one quarter.

We have been able to take advantage of the distress of other companies in the past by buying the whole companies. Our strategy has changed somewhat, and now we are trying to just buy dark fiber as opposed to buying the whole company and then use our equipment to light that fiber. That is, in fact, what we did in the quarter that did result in an increase in our rate of capital spending. But we think this is transitory.

Also, with regard to adding additional buildings, we have been very clear about the criteria that our buildings have. They either have to be large carrier neutral data centers where we believe there is a sufficient addressable market or they are multitenant office buildings. And to refresh the group's memory, the average building on our network is approximately 580,000 square feet, about 41 stories in height and approximately 51 tenants. Today we have just under 590 million square feet on net that represents a corporate footprint of about 1/10 of 1% of the buildings in North America, but in excess of all 8% of all rentable office space in North America. And we expect to continue this accelerated pace of building connections probably for the remainder of this year.

As Tad said, our capital spending will go down as we have spent much of that capital and are finishing up the final connections, and then we should expect to see that continued accelerated rate at least in the first half of next year. Until we see a general improvement in the employment picture, we intend to take advantage of better pricing to accelerate our footprint expansion.

And to your growth question, John, we will be able to continue our growth out of our existing footprint. As in 2009 and each of our previous years, over 90% of our growth in that given year comes from the pre-existing footprint. That is the footprint that existed on the Cogent network at the beginning of the year. We expect that same trend in 2010 as most of our 2010 growth will come from our existing footprint, and we are really looking to expand our addressable market with these route expansions, as well as building expansions.

Wonderful. Are you open to talk about who you picked up some of these dark fibers from?

We prefer not to disclose those. Today we have over 135 different suppliers of both metro and long-haul fiber. Some of them are well-known companies. Some are companies we compete with. Sometimes it is a cable operator or a local utility, as well as international governmental-backed entities. But I would not feel comfortable in disclosing any specific names.

Michael Funk, Bank of America.

Two quick questions. For the on-net corporate revenue growth, it appears that there was a deceleration sequentially on that line. I don't know if you can call out or highlight the impact -- determine the volume discounting had on that. If there was, I know you mentioned some accounting impacts earlier.

And then second, just on the share repurchase, clearly you are choosing to invest your cash back into the network over the next six months or so. How should we think about the share repurchases given the accelerated or greater network purchases relative to the previous expectations?

Okay. Michael. A couple of points. First of all, you are correct. Our corporate on-net revenue growth did decelerate. We have seen a lengthening of the sales cycles.

For the quarter our corporate growth in aggregate was 2.2%. As Tad mentioned, we do account for the installation over a longer and longer period. However, we did not see any material impact from that longer period of accounting. So I mean it was minor at best. Our off-net corporate business continues to grow well as our average connection size increases.

On the corporate on-net, we have seen very little volume increases from corporate customers as the average corporate customer is using only about 10% of the capacity they purchase. So the only decrease in ARPU from those customers is as a result of longer-term contracts. And, as we said, the contract lengths extended about 2.9% quarter over quarter and are slightly longer than one year now.

Now with regard to the share repurchase, we do have approximately $30 million still remaining under our program. We do have additional availability with the line of credit that we have if we so chose to use that to purchase either debt or stock back. But we believe in these turbulent economic times with unemployment over 10% that it is prudent for us to redeploy capital and accelerating the expansion of our network into as large of an addressable market as makes sense, but also remain disciplined about where we will connect the network only going after the highest trafficked locations.

So I would expect for the short-term to see us plow most of that capital back into either the network or the sales force. As we sit down with the board each quarter we outline the uses of our free cash and the highest and best use continues to be to grow the sales force. After that, we then look to divert cash into the growth of the network, and then only after we are comfortable that we have optimized those two growths do we look for a mechanisms to return cash to shareholders. But we are committed to not sitting on a large pile of cash, but rather be efficient in either deploying it towards growth or to return it to shareholders as we have demonstrated in the past.

Great. And one quick follow-up if I could please print. Sorry, the off-net corporate growth that was mainly due to existing customers adding more satellite offices, or was that due to some of the newer on-net corporate customers coming on board with more off-net offices?

It was actually our number of connections was virtually flat. So while we have churned some customers, the customers that replace them tended to buy larger connections, and we also tended to see some existing off-net customers increase their connection size.

What we are seeing in our off-net business is a significant acceleration of ethernet off-net services. So previously our off-net products were TDM-based, either T1s or T3s, and for some customers they would take multiple T1s as they bonded product.

As the local network operators increase the availability of point-to-point ethernet services, we then can purchase either 10 meg or 100 meg off-net loops, connect them into our network and cost effectively offer IP transit over those circuits.

So the reason for the growth in that off-net revenue was the number of connections remained the same, but the ARPU went up and the ARPU went up because the average connection size is getting larger. And that is both a combination of the natural evolution of some smaller customers churning larger customers and replacing them, as well as some existing customers upgrading to larger connections.

Colby Synesael, Kaufman Brothers.

Regarding the fiber purchases, I think you have mentioned 3000 route miles purchased. Considering that your network is essentially or mostly is in the form of cap leases and dark fiber IRUs, is part of those purchases because you think that you may be losing other routes in your future because some of your contracts might be expiring? So this is kind of to backfill that before that happens, or is this truly just for route expansions?

And then I have another question after that.

Yes, our average remaining IRU life is in excess of 20 years, and virtually all of our IRU agreements, as I mentioned, those agreements are with over 135 different suppliers. The vast majority of those agreements have extension rights in them. Most of those extensions actually come at no charge. They give us the ability to exit if we so choose because of increases in operation and maintenance expenses.

We do not anticipate any loss of fiber in the near term or even over the next 10 years from any of our existing IRUs. What we are looking at is the ability to add new routes either for new markets to be served or to add additional redundancy to the network. Our network is built as a series of rings. But, as our traffic volume increases, it does make sense to potentially bifurcate these rings. And two great examples of that are here in North America where we have taken our Western ring and cut it in half, creating a West-West ring and a kind of Midwest ring. We have done the same thing in East where we've bifurcated our Eastern ring into two rings. We have done similar things in Europe, and in those bifurcations we also ended up picking up new markets as well.

Just one follow-up there. The only route that has been canceled, if you will, was back in Q1 of '08 we had a route and one IRU agreement. That is the only impairment we have had on an asset in our corporate history.

So to answer your question, the new routes are all expansion. We have not had any cancellation of routes really since that period.

That is helpful. My other question is you mentioned in your comments -- and I know this is a big part of the Cogent strategy that your are 50% below the market book for the installed base, as well as for new sales. Can you just explain to us what -- or at least remind us what the benefit of being so far below the market is?

In other words, if you were just 25% below the market, what is the concern if that was what your strategy was? I mean I guess what I'm getting at is, can you remind us again why you're so far below the market and what the advantage is? Is there concern that these customers would no longer look at your services as valuable if you were to be just a little below the market?

In our pricing strategy, we compete with other Internet service providers, but there is also a more global question, which is the Internet competition with other forms of telecommunications. And that is broadly defined to also include things such as DVD distribution via the mail. And I think it is critical that Internet transit prices continue to fall in order to accelerate the adoption of these new business models and increase traffic growth on the Internet in its entirety.

In addition to that, it is I think a key part of Cogent's strategy to continue to accelerate its gain in market share. And you see that in either the amount of [grid] capacity we have, the number of routes that we are carrying, the number of networks that we are interconnected to, and there is a natural advantage of continuing to gain scale over our competitors, helping us further get down the price curve.

So I think the pricing model that we have adopted and continue to support helps us lower our cost of revenue acquisition and accelerate our gain in market share. And what we have seen is a pattern where a number of companies we compete with chose not to compete with Cogent. Not because they cannot compete but because they have other streams of revenue that they are looking to invest their capital into and kind of move up the value chain. It is a constant refrain of service providers that they believe that they want to move up the value stack.

We at Cogent are very comfortable where we are at at the bottom of that value chain, and we have demonstrated our ability to operate profitably, expand our margins and generate free cash in that portion of the value gradation. So I think our pricing strategy continues to be one that helps us gain market share and lower our cost of revenue acquisitions, and we think kind of the differentials we maintain are appropriate for those two goals.

Tim Horan, Oppenheimer.

Just out of curiosity, how much do you pay for like an ethernet 10 meg connection versus T1?

Yes, typically depending on loop length, a dry Tier 1 -- and we are not a CLEC. So we buy either under tariff or under negotiated contract. I would say the average T1 cost for us is in the neighborhood of about $150. That does not have the IP layer on top of it. That is just the raw point-to-point T1. And again, there is a great deal of variance depending on loop length.

For the ethernet services, the 10 meg ethernet services can range anywhere from about $500 to about $1500. And for 100 meg, which I know you did not ask but becomes fairly common, we see pricing of anywhere from about $800 up to about $2500. All of those variances are dependent on service territory and loop length.

And do you see your customers putting more voice over IP on top of the ethernet or other applications like existing private lines? Where are we in that process do you think?

Yes, I think there is a long-term structural change going on in the industry that all other technologies whether they be frame relay, X.25 or dedicated WAN services are migrating to IP and are generally put over a converged facility. So voice continues to migrate to our network as packets running where the customer -- where a third party is running a VOIP service. Teleconferencing is another application for our corporate customers, disaster recovery, and then obviously for our net-centric customers, it is really their business models whether it be a social networking site, a gaming site, a video distribution site, all of those applications are only enabled by the public Internet which is IP.

So I think over time ethernet will become the dominant wide-area interface just like it is already the dominant local area interface. IP becomes the common transport mechanism, and I think it is the growth in traffic is coming both from new applications that previously did not exist, as well as existing applications moving to our network.

Today people consume about 1/60th of their video over the Internet. The other 59/60ths are consumed by some other distribution mechanism. And I think over the next few years it will be that increase in video consumption. So that is not a new application, but it is rather the Internet and ethernet and IP displacing other usually QAM-based delivery mechanisms as the primary way of delivering that content.

And then on the pricing front, I was a little confused on what you were going at with your ARPU. Were you saying the ARPU trends have improved substantially? It seemed like they worsen a little versus second quarter. I know they are a lot better versus third quarter of last year. I guess more importantly maybe roughly what do you think ARPU on-net is going to be going forward?

And related to that -- sorry, for the long question -- but the $4.82 on new sales pricing, how has that been trending, which, I guess, is kind of all tied in?

So there are really two different questions to focus on. The first question is the price per megabit, which is very relevant for understanding our ability to grow our net-centric revenues. And to remind you, those revenues grew sequentially actually at an accelerated pace in this quarter of 5.4% quarter over quarter. That new price in the third quarter of $4.82 compares to a new sale in the second quarter of $5.05. So that is coming down, but coming down at a slower pace than it previously has come down.

The installed base is converging to that lower price coming from $6.82 per meg in the second quarter to $6.33 in this quarter. But that was more than offset by the volume increases resulting in that increase in net-centric revenue.

ARPU for on-net services combines corporate as well as net-centric. It looks at the contract length, the volume and type of connection and is impacted by FX. It did decline by about 2.8% in the quarter. I would expect that as net-centric has increased as a percentage of revenues we should see that start to moderate and even head up.

So the average connection size gets bigger and the ARPU goes up. We actually have seen that happen in the off-net business about six quarters ago. About a year and a half ago we had seen previously sequential revenue ARPU decline in that business, but then we started to see it go up. And the reason it went up was the average connection got bigger. We should start to see that as well in the on-net business.

Do you think you can see out in the fourth quarter, or is that more of a 2010?

You know, it is really difficult for me to predict accurately. I think I would feel more comfortable with the 2010 estimate than a fourth-quarter estimate.

Frank Louthan, Raymond James.

A couple of questions. Just to clarify, on the 3000 route miles that you added, are those actually lit now? When you say you have added them, you bought dark fiber and you have actually lit them? And then how many additional buildings do those 3000 miles give you an opportunity to add to?

And then the last thing is, can you give us an idea on verticals where you're getting some traction or seeing any changes in what the top three customer verticals are that you are selling to currently?

Yes, Frank, first of all, all 3000 route miles are lit at this point in time, and it's actually in access of 3000. That is, in fact, part of the reason why the utilization rate in our network declined from 20% to 19%. So while traffic grew, our utilization rate actually declined because we were able to light these additional miles, and it is a combination of both metro and long-haul.

You know, our network today serves approximately 1420 buildings. It literally passes within 1000 feet, hundreds of thousands of buildings. But many of those buildings do not meet our stringent criteria in terms of prospectively building into them. If we get a customer request for a building that does not meet our requirement, we will build in if the customer either guarantees sufficient recurring revenue or gives us a large nonrecurring fee that would cover our cost of construction. That rarely happens because our sales force is really only focused on the lit buildings that we prospectively build into.

So while we connect to 1400 buildings past hundreds of thousands, I would suspect there will only be a few hundred more buildings either data centers or corporate that we will proactively connect to.

Now with respect to verticals, I'm going to answer that question both on the corporate and net-centric side.

On the corporate side, our verticals continued to be, number one, law firms; number two, financial services companies; and number three, general consulting. Those tend to be the most common types of businesses that locate in our addressable market.

Remember we tend to focus on the largest and usually the most expensive real estate in a given market, and these have been our biggest verticals and will continue to be that.

On the net-centric side, our business is actually divided almost equally between access network operators. So today we have literally dozens, multiple dozens of international phone companies that are our customers. We have hundreds of competitive carriers. We have got hundreds if not thousands of regional ISPs and cable companies as customers and over half of the universities and colleges in North America, coupled with a large number of K-12 systems.

So the access network operators account for about half of our net-centric revenue. The other half comes from a variety of content businesses, and there it is hard to pick any one business model. Some are social networking. Some are purely video distribution models. Some are major well known existing video producers. Others are brand-new video producers. So it is just a very -- it is hard to pick a specific vertical inside of the content other than video being the most common theme.

Jonathan Atkin, RBC Capital Markets.

I wondered if you could talk a little bit about seasonality that you saw during the quarter, as well as any commentary in October on the traffic trends and sales force productivity? As well as the sales reps that you hired, the 20 net new, was that mostly on the corporate side or mostly on net-centric?

I will let Tad take some of the seasonality on the cost side. I will take the rep and traffic question.

From the cost side, I would say, if you look at the two lines that are EBITDA impacting, so cost of goods sold and SG&A, on the cost of goods sold side, little seasonality with the exception of power costs. The remainder increase is associated with our expansion activities. If you look at the SG&A line, the seasonality impact there would be an increase in professional fees associated with audit and tax activities, which are increased in the third quarter from the second quarter, and the second quarter is also up from the first quarter. Absent that, not a great deal of seasonal impact on the cost components of EBITDA.

And on the demand side and revenue side, the corporate business remains relatively steady on a year-round basis. The net-centric business is susceptible to increases in traffic, and as we have stated in the past, the Internet is primarily a Northern Hemisphere phenomenon, and much of that traffic growth occurs in colder months.

We believe that you will continue to see an acceleration in traffic growth, even though the base is getting larger. While we do not disclose specific monthly trends, I will say that October has trended actually better than what we saw on average in the third quarter. Much as last year we saw a continued acceleration in the fourth quarter and first quarter in our rate of traffic growth, and we should expect to see that even though the base is getting larger.

To the rep question, roughly 2/3 of our reps are corporate. About 1/3 of our reps are net-centric. The hiring that we did in that quarter mirrors that ratio as we have kind of a fixed regime of openings on a global basis. We are looking to hire specific reps for specific offices, and that roughly 2/3, 1/3 ratio will be maintained on a going forward basis.

Michael Rollins, Citigroup.

I was just wondering if you can give a little bit of an update on some of the I think they were take-or-pay contracts that you described a couple of quarters ago in terms of you were looking to get some benefit in the ramp of revenue in the back half of the year?

We absolutely saw the benefit of that, and that was, in fact, part of the reason both for the traffic growth and the 5.4% sequential revenue growth in our net-centric business. Those take-or-pay contracts did not really impact our corporate customers and will not on a going forward basis.

This was not a new program instituted this year, but rather a part of Cogent's normal business activity. We continue to sign those contracts in both the second and third quarters. We continue to see the benefit of those contracts that were signed either at the end of last year or the beginning of this year, and we expect to be able to comfortably meet the top-line revenue guidance that we have had in part because of these take-or-pay contracts. But we often forward price customers who are testing new business models and give them the greater volume discounted price in order to help stimulate their business in exchange for a commitment from them to take additional bandwidth for a greater period of time, and that trend remains intact.

We have no further questions at this time. I would like to turn the call back over to our presenters for any additional or closing remarks.

Well, thank you all very much for taking time out on a Monday to join us. Hopefully you are as pleased with our results as we are, and we expect to continue to see both growth in traffic and revenue, along the same lines as we have delivered this quarter.

Thank you all very much.

That does conclude today's call. Thank you for your participation.