Cogent Communications Holdings Inc (CCOI) 2008 Q2 法說會逐字稿

Good morning, and welcome to the Cogent Communications Group second quarter 2008 earnings conference call. As a reminder, this conference is being recorded, and it will be available for replay at www.Cogentco.com.

I would now like to turn the conference over to Mr. Dave Schaeffer, Chair and Chief Executive Officer of Cogent Communications. Please go ahead, sir.

Thank you and good morning. Welcome to our second quarter 2008 earnings conference call. I am Dave Schaeffer, Cogent's Chief Executive Officer. With me on this morning's call is Tad Weed, our Chief Financial Officer.

Overall, we are relatively pleased with our results for the quarter. While our revenue and EBITDA did not exceed our second quarter guidance, we continue to grow our on-net revenues, and our EBITDA margin expanded at a record level of 270 basis points, the greatest margin expansion ever in the Company's history in a given quarter. During the quarter, we further expanded our revenue and sales productivity initiatives, reemphasized our position as an industry price leader by expanding our volume-based contract discount program to our net centric customers. We ended the quarter with over 200 sales representatives, and we have lowered our sales force turnover rate.

We are extremely pleased with the initial results of these initiatives, and continue to invest in the improvements of our sales organization. During the quarter, we announced an increase in our stock buyback program of an additional $50 million, with a purchase authorization through the end of the year. In the second quarter we returned over $28 million to our shareholders, by purchasing a total of approximately 1.8 million shares, at an average price of $15.26.

As of today, we have approximately $40 million remaining available under our buyback plan. Through yesterday and since June of 2007 under our total of $150 million of combined stock buyback programs, we have purchased a total of 5.3 million shares of our common stock for approximately $110 million. These purchases represent over 10% of the outstanding shares in the Company when these programs began. Throughout these discussions, as in the past, we will continue to focus on our results and impact of our on-net business, which accounts for approximately 82% of our revenues.

We will highlight several operational statistics that we believe demonstrate our increasing market share, expanding sale, and the operating leverage of our business. I will review with you certain operational highlights, also outline the results to-date of our revenue improvement and sales productivity efforts, and our continued expansion plan. Tad will provide additional details on our financial guidance. Tad will also walk through our guidance for the third quarter of 2008, as well as update and modify our guidance for full year 2008. Following our remarks, we will open up the floor for questions and answers.

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

Thank you Dave, and good morning everyone. This second quarter 2008 earnings report and this earnings conference call will discuss Cogent's business outlook and contain forward-looking statements within the meaning of Section 27A and Section 21E of the Securities Act. The 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, acquisitions, other business combinations, or financial 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 in Reg G, you will find these reconciled to the GAAP measurement in our earnings release, and on our website at Cogentco.com.

I will turn the call back over to Dave.

Thanks, Tad. I would like to give you some highlights of our second quarter results. Hopefully you have had a chance to review our earnings press release. As within previous quarters, our press release includes a number of historical metrics. These metrics will be added to our website. Hopefully you will find these metrics informative, and help you better understand our financial results, and the trends of our operations. As always, if you have any suggestions on the metrics that you would like to see added or refined, please let us know.

Our second quarter 2008 revenue was $53.9 million. This was 1.1% below our guidance target of $54.5 million. Revenue increase represented a 3.4% increase over revenue in the first quarter sequentially, and a 19.4% increase over our revenue in the second quarter of 2007. EBITDA as adjusted of $16.6 million was 2.4% below our guidance for the quarter of $17 million. This EBITDA increase represented a 13.5% increase in EBITDA over the first quarter of 2008, and a 50% increase in EBITDA over the second quarter of 2007. Our loss per share of $0.12 was within our guidance range of $0.08 to $0.12 per share. This loss was impacted by our share repurchases throughout the quarter, having less shares available concentrated that loss in fewer shares.

Despite a 1% decline in traffic on our network quarter-over-quarter since last quarter, our on-net revenue in business grew by 3.3% in the first quarter and by 25.3% from Q2 2007. We believe that this revenue growth continues to be greater than the revenue growth rates of our competitors. On-net revenue was approximately 82% of total service revenue for the second quarter, essentially unchanged as a percentage of revenue from the first quarter.

Approximately 85% of our new sales in the quarter were for on-net services. Our on-net customer connections in the quarter increased sequentially by 5.5%. Revenue from our off-net business also increased by 5.8% quarter-over-quarter, and our non-core revenues declined for the same periods by 9.2%. Off-net, the off-net quarter to quarter revenue growth was consistent with our guidance for off-net revenue growth of over 5%, and the highest off-net revenue growth rate we have experienced since the first quarter of 2005.

Now I would like to take a moment and talk about pricing. Since our inception, Cogent has been committed to being an industry price leader. We intend to remain committed to that policy. In the second quarter, we reemphasized our price leadership position by initiating additional discounting structures that included volume-based discounting, as well as additional contract term discounts for customers taking multi-year contracts of greater than two years in length, in particular, three-year contract discounts.

In May of 2008, we completed a pilot test program where we offer volume-based discounting to certain net centric customers. We were very encouraged by the results of that program. As a result, on June 2, we expanded this program to include all new Cogent net centric customers, and to existing net centric customers who increased their total contract value with us.

Let me take a moment and define that increase in total contract value. This is the number of months remaining under which a customer has an existing contract, times the monthly recurring revenue that they are contractually obligated to spend. Under this program, and with the assistance of our other productivity initiatives, we are encouraged by the results of our June and July activities, and the increase in sales force productivity.

Pricing for our most widely sold product remains at $1000 a month, for our month-to-month 100-megabit product, or $10 a megabit. Discounts are available related to contract term, as well as volume for our net centric customers. Our average price per megabit of effective sold traffic remains at about $9 per megabit when these programs are taken into account. The net centric pricing discount program did not and does not impact our corporate customers. Our corporate customers represent over 60% of our total customers, and actually 44% of our revenues are today derived from our corporate customers. This is up from the 42% of our aggregate revenue that was the result of our corporate customers in the first quarter.

Our provisioning cycles remain consistent. We remain guaranteeing our customers 17 business days to provision on-net services. In the second quarter, our provisioning team again delivered better results in provisioning our on-net customers in an average of 10 days. Traffic, as I mentioned earlier, traffic declined on the network by 1% for the full quarter. Traffic did increase in the months of May and April, each by about 1%, but then in the month of June, we saw a significant decline in traffic of approximately 3%.

First, we have a significant concentration in university and K-12 customers, and due to the seasonality of traffic requirements, not revenue spend by those customers, we saw a decline in traffic. Secondly, we saw a series of video and social networking sites, who exhibited much more modest traffic growth than they previously have been exhibiting.

Now I would like to talk about our ARPU. ARPU for our first quarter on-net customers was approximately $1,239. That did decline to $1,210 in the second quarter, a decline of approximately 2.3%. ARPU for our off-net customers continued to increase. Off-net ARPU increased from $890 in the first quarter of 2008, to $940 in the second quarter of 2008, an increase of approximately 5.7%, due primarily to off-net customers taking larger connections.

Total on-net churn remained consistent at about 2%. On-net churn is at the same rate that we have experienced for the past two years. Off-net churn also remains consistent at approximately 2.5%. Contract terms, during the quarter, we again extended our average contract length by approximately 3%. Our customers are entering into longer-term contracts, as they are more and more confident with Cogent's quality, and expressing their commitment to Cogent and the value that we deliver. We anticipate further increases in our average contract length, due to the expansion of our term discount program.

Tad will now cover some additional details related to our second quarter results. Tad will also provide guidance for third quarter 2008, as well as update our full year guidance for 2008.

Thank you Dave, and again good morning to everyone. I would like to thank our team on their hard work and efforts this quarter.

EBITDA and gross margin, as Dave mentioned, EBITDA as adjusted was $16.6 million for the quarter. That was an increase of 13.5% from the $14.6 million for the first quarter, and the margin expanded, EBITDA margin expanded by 270 basis points, from 28.1% to 30.8% for the second quarter. The margin expansion was better than our guidance of 200 basis points. Gross margin decreased by 50 basis points for the quarter.

Our second quarter gross margin was 57.4%, compared to 57.9% for the first quarter. The decline was in part due to our revenue mix, and increased cost of goods sold related to our data centers. Our direct incremental on-net gross margins continue to be almost 100%, with direct incremental on-net EBITDA margins of about 95%. Our loss per share, basic and diluted common share of $0.12 was within the range, but at the high end of our range of $0.08 to $0.12 a share.

On-net, our net loss and loss per share for the quarter was impacted by a combination of less cash from accelerated share repurchases, and a reduction in interest rates in our invested cash, and as Dave mentioned, fewer shares outstanding due to share repurchases. Each of these factors increased our loss per share by about $0.015 per share from our plan for the quarter.

Our weighted average shares decreased by about 870,000 shares, as a result of shares purchased under our stock buyback program. And we purchased 1.8 million shares in the second quarter for $28 million, and our plan for the remainder of 2008, anticipates and includes the estimated impact of additional share repurchases. Noncash equity based comp expense for the quarter was $4.2 million, or $0.09 a share. That amount was slightly less than our guidance of $4.5 million for the quarter.

Depreciation and amortization was $15.8 million, which was consistent with our guidance of $16 million. The Euro to dollar conversion rate positively impacted our comparable quarterly revenues by about $500,000. For the second quarter of 2008, about 23% of our revenues were based in Europe, and 8% of our revenues are related to our Canadian operations. These geographical percentages for our revenue mix is consistent with the percentages we have experienced since the fourth quarter of 2007.

CapEx totaled $9 million for the second quarter, versus $9.8 million for the first quarter. On a quarterly basis, we can and historically experience seasonal variation and lumpiness in capital expenditures and construction activities. Typically we experience our lowest level of CapEx in our fourth quarter. We continue to expect our 2008 CapEx to be approximately $30 million.

At June 30 on balance sheet items, our cash and cash equivalents totaled $129.2 million. Capital lease IRU obligations totaled $105.4 million at June 30, and about $8 million of that is a current liability. These obligations are paid over remaining weighted average life of more than 10 years. Our Days Sales Outstanding for worldwide Accounts Receivable continued to improve, and was only 32 days at June 30, significantly better than our target of 40 days, an improvement from the 34 days at the end of March. It seems that every quarter I need to mention this, but again, I need to personally thank and recognize our worldwide billing and collections team, for doing a great job on customer collections.

Operating cash flow, cash flow from operations improved by 24%, and was $14.2 million for the second quarter; it was $11.5 million for the first quarter. Now I'm going to provide guidance for the third quarter of 2008, and update our 2008 guidance. We prepare our guidance based upon the current and expected run rates of our business. Our guidance also includes the estimated impact of our share repurchase program, planned increases in sales and marketing programs, and the estimated impact of our current and anticipated network expansion projects and current traffic trends.

For the third quarter of 2008, we expect our total revenue to be over $55 million. From the second quarter to the third quarter, the components of revenue are expected to change as follows. We expect on-net revenues to increase by over 2%. We expect off-net revenues to increase by over 5%. We expect non-core revenues to continue to decline by approximately 15%.

We expect EBITDA to be over $16.5 million for the quarter. We expect noncash equity-based comp expense to be about $4.5 million. We expect depreciation and amortization to be about $16 million, and we expect net interest expense to be about $2.5 million, which is reflecting a lower interest rate on invested cash, and the additional estimated share repurchases. These results are expected to result in a loss per share of between $0.10 and $0.15. Our loss per share assumes 44 million weighted average shares outstanding for the quarter.

For the year of 2008, we are updating the previously released estimates for the year. Total revenue for 2008 is now expected to be over $218 million; our previously released guidance was between $225 million and $235 million. We expect the following annual revenue growth rates from 2007 to 2008. On-net revenue growth of approximately 22%, our previously issued guidance was an increase of approximately 30%. Off-net revenue growth of approximately 5.5% to 10%, that is unchanged from the previous guidance, and non-core revenue decline of 35% to 40% is also unchanged. We expect EBITDA as adjusted to be over $65 million; the previous range was $75 million to $80 million. We expect net interest expense to be between $7 million and $8 million, from our previous estimate of $4.5 million to $5.5 million, again impacted by less cash due to share repurchases.

Due to these changes, we expect our 2008 loss per share to be between $0.50 and $0.60 a share. Our previous estimate was $0.20 to $0.30 a share. The loss per share estimate assumes about 45 million weighted shares outstanding, and the previous estimate was 46 million weighted average shares outstanding. Due to these revisions, we no longer anticipate achieving positive net income by the fourth quarter of 2008.

Now I will turn the call back over to Dave.

Thanks, Tad. I would like to take a moment and talk about sales force productivity. We began the second quarter of 2008 with 190 quota bearing sales reps, and ended the quarter with 208 quota bearing sales representatives. We hired a total of 59 sales reps in the quarter, the most reps we have ever hired in a single quarter in the Company's history. 41 sales representatives left the Company during the quarter, or were promoted to management positions, down from the 54 sales individuals who left the sales role in the first quarter. This improved our monthly sales force churn number to about 6.9% per month in the second quarter.

We began the second quarter of 2008 with 156 full-time equivalent sales representatives, and ended the quarter with 177 full-time sales representatives. As of yesterday, we had 221 quota-bearing sales reps selling our services. We have added additional tools and resources and continue our disciplined approach to monitoring our sales force productivity.

On a full-time equivalent basis, our sales force productivity per FTE was approximately 3.4. This was less than the 3.6 that we had experienced in first quarters. As a reminder, these rates are customer installations, and are not necessarily directly correlated to immediate sales activity, because of the lag of 10 business days between sales and installations for on-net, and about 30 days for off-net services.

As mentioned on the last call, we have taken a number of measures and steps to improve our sales force performance and rep productivity. We expect to continue to see improvements from those initiatives. As an outlook, we will continue to add sales reps in 2008. At the pace we are currently adding reps, we now expect to have a total of 250 quota-bearing reps by the end of the year, an increase from our previous forecast of 240 sales representatives by year end.

Now with regard to our footprint and our scale, we added 27 buildings in the second quarter of 2008 to the network, bringing our total footprint to 1,274 buildings on the network as of the end of the quarter, and that is defined as fiber connectivity into the building all the way to the customer. For 2008, we expect to fulfill our goal of adding 100 buildings to our network, and for the first six months of 2008, we have added 57 buildings towards that objective. We have secured additional fiber and construction continues in expanding our network. We have added additional markets, such as Tulsa, Oklahoma, Munster, Germany, Malmo, Sweden, Basil, Switzerland, and Bern, Switzerland to the network. We continue to add additional markets, and we continue to evaluate different routes and new extension opportunities in Europe and in North America.

The scale and size of our network continues to grow. We now have over 12,000 fiber miles of metro connectivity. We have over 32,000 inner-city route miles of fiber. We are directly interconnected to approximately 2,400 networks throughout the world, approximately 330 of these networks are settlement-free peers. The remaining connections are Cogent customers. We believe that our network has substantial available capacity continued to accommodate our growth plans. We are currently and consistently utilizing approximately 21% of our network, unchanged from the previous quarter.

So in summary, our on-net revenue, our growth business continues to grow, and is growing at a rate greater than our competition, albeit slower than it has for Cogent previously. Cogent is the low cost provider and our most recent pricing changes emphasize this position in the market. We have revised our 2008 plan to reflect traffic growth rates, the stock buyback program, and the expected increases in our sales force.

Our business remains completely internet-based and internet-focused, and we believe that we are isolated from the current economic turmoil. Our business provides a necessary utility to customers. Our simple product set and limited number of SKUs, allow us to continue to surpass our 17-business day provisioning guarantee of our on-net services.

We have a very strong balance sheet, especially when compared to others in our industry. We have $67 million of true debt. During 2008, we intend to increase the lid capacity in our IP network by approximately 60%, all within the 2008 fiscal year capital budget of $30 million. Under our revised 2008 guidance plan, we now expect to generate approximately $0.80 a share of free cash flow to the benefit of our equity holders. This is down from our previous estimate of approximately $1.00 per share of free cash flow.

Finally, we continue to be extremely encouraged by the recent results of some of our sales force productivity initiatives, and our volume-based pricing program. For the month of July, we sold over 200 customer connections to existing net centric customers under this program. The total contract value to those 200 existing customers increased by greater than 100%, and resulted in approximately $10 million of additional contract value for Cogent. I think this continues to demonstrate the wisdom of the plan that we put in place.

I would like to now open the floor for questions.

(OPERATOR INSTRUCTIONS) And our first question will come from Jonathan Atkin with RBC Capital Markets.

Yes, good morning. I had a couple of questions. First of all, with regard to ARPU, can you comment on the ARPU you are seeing from customers that have taken advantage of the new pricing that you introduced on a wide scale basis in June?

Sure, John. We have actually seen an increase in ARPU from those customers taking advantage of the new volume-based pricing. That ARPU increase came from both existing customers, as well as new customers taking advantage of that program. However, we saw an increase in the percentage of our business that comes from corporate customers. Corporate customers increased from 42% to 44% of our total revenue.

Corporate customer ARPU is substantially lower than that of our net centric customers, and that was the primary reason for the reduction in ARPU. Our corporate business remains extremely strong, and is not impacted by the traffic trends, because our average corporate customer is utilizing only approximately 8% of the bandwidth that they purchase, and traditionally does not come back and buy more bandwidth at a same location, but can increase their purchases from Cogent by taking additional locations.

Can you also give us a refresh on the mix of new revenues coming from existing customers versus new logos? In the revised guidance, what are your assumptions on that mix going forward? And then finally, a question about the lat expansion that has been ongoing and continues, particularly the new long haul routes, as opposed to the laterals, are the revenues that you are seeing along the new routes meeting your expectations? Is the revenue growth there just proportionately high or low, does it take a while to ramp up the revenues along those new routes? Thanks.

Sure, John. Two very different questions. In terms of existing customers taking additional business with Cogent, on the corporate side, that rate of mix has not materially changed, with approximately 30% of incremental revenues in a given month coming from an existing customer, and approximately 70% coming from a new customer.

For our net centric business, however, because of the slowdown in traffic, the percentage of business that we are getting from existing customers has declined to only about 15% of incremental business, with about 85% of business coming from brand-new businesses that do business with Cogent. And I think this is a result of those companies taking less bandwidth because their underlying businesses are not growing at as fast of a pace.

Now shifting gears and looking at the route expansion strategy, Cogent has always been focused on the largest and most desirable pockets of traffic that are economically [served]. We go through a detailed analysis as we evaluate new route opportunities. We look for existing customers who will purchase more in those markets, as well as customers who have expressed interest in Cogent's services, but have not been able to connect to us, because of the lack of interconnect facilities to get to a Cogent point of presence. We also evaluate whether or not the fiber that would be led, is in our inventory, which does exist in some cases, or needs to be purchased.

Then finally, we evaluate the addition of the new market in enhancing the quality of service to the entire base by increasing Cogent's level of interconnectivity. As I mentioned, Cogent's network interconnectivity increased by about 5% in the most recent quarter, going from approximately 2,300 networks globally to 2,400, and that does have an impact on the entire organization, and all of our customers. So all of those factors weigh into our evaluation of new markets. We expect this to continue, but at a slower pace than it has been in the past. We also do expect to continue to add buildings at a comparable pace at least for the next several years.

Finally to your question of revenue production on these new routes, many of the routes are very new. That revenue production is lumpy. Some of the routes have brought revenue on faster than we have anticipated. Others have been slower. Some of the more eastern European countries tend to have a little bit more regulatory burden, in terms of dealing with customers, and being able to get those customers to quickly move to Cogent.

So in some cases, customers sign orders, but we have not been able to provision them as quickly, because they need certain types of regulatory approvals to connect to us. All of these factors have created a situation where some markets have developed new revenues slower than we would like. Others, maybe a little bit faster, but on balance, we think our strategy has been correct, and will remain effectively consistent.

Thank you.

Thank you. We will now take our next question from Frank Louthan with Raymond James.

Thank you. Looking at your business here, obviously you are growing your percentage of corporate customers. Does that also imply that you're growing, is that where you are seeing some mix shift that's putting pressure on margins as those corporate customers are more and more off-net? And can you give us an idea of how quickly the traffic has been falling off, and remind us what the traffic growth was in Q1 and Q4, and when did you start to see this traffic fall-off, and what has been the experience with some of your larger embedded customers so far this quarter? Has that continued to get worse, or is that stabilizing, or what are the thoughts there?

Sure. Two very different questions, first of all, hi Frank.

In terms of margin pressure, we delivered the largest sequential EBITDA margin expansion quarter-over-quarter in the Company's history. It is unrealistic to expect 270 basis points of EBITDA margin expansion quarter-over-quarter. We have told investors to think of Cogent as delivering about 100 basis points quarter-over-quarter. We did about 500 basis points for the past four quarters, and we will expect to kind of continue to do that going forward.

On the next quarter, we have guided to a lower margin expansion rate, simply because we had such great performance in the most recent quarter, and are quite honestly not sure it can continue, as well as we have some seasonally expected costs, particularly on the cost of goods sold line, where our data centers are now all being billed at summer power rates, which are on a per megawatt basis for the locations in which we directly buy the power, as opposed to a third party are seasonally impacted, and are generally charged at a higher rate.

Corporate customers do generally generate less margin, because we have the cost of the local loop. Our corporate business that is off-net has increased, which has also dampened the rate of margin expansion. But our corporate on-net business has identical margin characteristics to that of our net centric business, which is 100% incremental gross margin, and 95% incremental EBITDA margins. We expect those types of margin expansions to continue going forward, as we add additional customers.

Now let me shift gears and focus on the traffic question for you. In terms of traffic, one, we saw in both April and May sequential traffic growths of 1%. That is far below our historical growth rate. Cogent's compounded traffic growth rate for the past five years is about 120%, even counting in this most recent downturn in traffic growth. So clearly 1% monthly sequential is far below the rate we have seen historically.

Secondly, we have in the past seen a slowdown in the rate of traffic growth in the summer, and in fact, last year saw one month, the month of August where we had negative traffic growth. This year, that negative traffic growth came in the month of June, where we saw that traffic growth go down 3% month-over-month, resulting in the full quarter being down 1%, which is the first time in our history we have seen traffic growth be negative for a full quarter.

Our revenue growth is partially impacted by that, but it is still possible to grow revenue substantially faster than traffic. You saw that with the mix shift to a greater number of corporate customers, which are not particularly traffic sensitive, and the fact that we grew revenues in the quarter 3.4% in aggregate, even though we experienced this traffic growth. It does mean the sales force has to work harder. There is not that natural uplift from those existing net centric customers. But if we look at our trends over the past year or 18 months, we began to see a deceleration in the rate of traffic growth really at the very beginning of the year.

When we built our full year guidance and issued it in November, we were basing it on run rate data that was compiled in October and September of 2007. A big part of our modification and reduction in our full year guidance is a result of the observations that we have made in the market of the actual traffic trends that we are seeing. While many of our competitors comment on traffic anecdotally, I think a number of them have demonstrated, and some of them forthright and others maybe not so forthright in acknowledging this slowdown in traffic growth. And it is supported by unique impression views on the Internet over the past year. The Internet will continue to grow, but it is in a period of transition.

Your specific question on the last quarter to quarter growth, the previous quarter, that was 6.4%.

Okay. So what is in the experience in July? Has this traffic continued to worsen, and can you give us some color on that?

Sure, Frank. In July, traffic actually increased 1%. That is clearly better than the June performance of a negative 3%. That is not, however, robust growth, and we do believe that part of the reason why it didn't continue to decline, was in fact the volume based pricing initiatives that we put in place.

Normally we would see traffic continue to deteriorate over the summer, and in fact that we saw a 1% sequential growth rate was encouraging, and we think in looking at the specific customer data that a large portion of that growth came from customers who did take advantage of the volume-based pricing program.

Can you give us an idea of just sort of a range that we should think about with what sort of margin pressure you get in the summer months from the peak power rates, is there any sort of rule of thumb, or how much your power rates have been going up, and how we should be thinking about that from a modeling perspective?

On kind of a quarter-to-quarter basis, it is in the neighborhood of $250,000, if you look at the aggregate kind of quarter-over-quarter change. When we build our forecast, what we have done historically, continue to do is look at the trailing six-month averages, not blindly take that figure, but also incorporate the trends, and then use that rate going forward. We have kind of incorporated that increase into the anticipated rate for the second half of the year.

So we basically took a $250,000 per quarter hit to our gross margin, and an increase in cost of goods sold, due to the increased power costs.

Okay, great. And then just last question, on the corporate side with the economy, are you seeing customers taking, harder to sell, because they are laying people off, so they don't need as much incremental bandwidth? Or are they pushing out the sales? Is the sales cycle lengthening at all? Any changes in that direction, and how so far have you progressed, as far as bad debt and so forth in this quarter? Thanks.

It would be very easy to use the economy as a shield to explain the slower traffic growth and we do not see that. That is just not our experience. In talking to our sales force, in measuring opportunities within our sales funnel, and the amount of time that they spend in the funnel to move to closure, the actual sales cycle time has not expanded. If anything, because of our value position in the market, we may actually be benefiting from a tough economic environment, as customers look to get more value.

The internet is a utility and very few businesses today can function without internet connectivity on the corporate side. On the net centric side, people buy bandwidth because they need it. They incorporate that bandwidth into the final good or service they sell, or they distribute it to their end users in smaller increments. They are more likely to buy from Cogent than less likely, if their margins are compressed. I know some of our cable customers have actually commented publicly that their upstream costs have fallen, and that was in part due to the fact that they have shifted traffic to Cogent.

I am going to let Tad talk about the bad debt and collections experience.

Sure. If you look back through the first quarter of 2006 through the end of the most current quarter, bad debt expense as a percent of revenue has ranged anywhere from 0.5%, which was the low in the first quarter of 2007, to 2.3% in the third quarter of 2006. So while those numbers are not large in terms of percentage of revenue, it is a decent range. We are very pleased with the performance in the current quarter, where bad debt was 1.3% of revenues.

Do you think that is going to increase going through the rest of the year?

Again as I said, we take kind of the last six-month average and anticipate that trend remaining constant. I have not incorporated a reduction nor an increase in bad debt expense.

Okay, great. Thank you.

Thank you. Moving on to John Schildkraut with Jefferies.

That was close.

That was pretty good, John. How are you doing?

I am all right, Dave. I would like to go through a couple of things. In the past, you have talked about sales force productivity numbers in terms of net and gross revenue dollars, when we visited with you in June, you pointed to about $400,000 of net sales productivity in the prior quarter per month, and that you had given us some numbers around your launch of the new price card, or at least a test period, where net sales force productivity had gone up to about 750,000.

I would like to get a sense as to where the net and gross sales force productivity numbers are coming out of the quarter. As we look into your guidance for next quarter, you are implying about $150,000 per month of net sales force productivity. Then I will come back with another question.

Sure. You are correct. We saw and were very encouraged by the initial test results, and saw a tremendous level of take rate from those net centric reps who were given the test products. That did result in an increase in the number of sales. Remember, for revenue, there is a bit of a lag between those sales, and our ability to recognize that revenue, so as we said, you would see very little impact in the second quarter.

On the counter negative side, what we have seen is this drag, from a lack of traffic growth from existing customers. We factored that into our guidance, and that is why we came up with the lower sequential revenue growth in the guidance, simply because we have experienced a reduction in the rate of traffic growth from customers actually went negative, because we also gained some customers.

So you can obviously see that if total traffic was negative 1% in the quarter, some customers were negative even more than that, considering we added customers that didn't exist in the quarter, and we ended up with more customers. We also derived a portion of our revenues from customers who exceed their commitments, and that rate of burse penalty actually declined by about 25%, and we factored in that decline into our analysis.

Now clearly if traffic accelerates, or even if the 1% trend that we saw in just the month of July, is not impacted in August and September, again we will be great and do better than what we said. But to Tad's point around guidance, we wanted to be very conservative, and take kind of the most recent data we had, and project it going forward.

All right. So no specific numbers around some gross and net sales force productivity?

Yes. I reaffirmed the analysis that you said of net, and the churn rate remains the same. So obviously the gross number has come down, and part of the reason for that is there is a negative gross number that comes from the reduction in that penalty rev.

Okay. Also, as we look into the back half of the year, EBITDA guidance, I know that Tad mentioned some incremental costs due to power, but still it looks like the operating leverage implied from an incremental EBITDA margin perspective, is significantly less in the back half of the year than the front half of the year. Beyond power, are there other things we should be thinking about? There was a fairly substantial reduction in SG&A in the quarter. Perhaps that will scale up again as you add more head count?

Yes, if you look at the kind of quarter to quarter trends, as anticipated, we did experience a reduction in SG&A this quarter compared to the first quarter, and the planned increases that we will incur from Q2 to Q3 primarily include two things, one, obviously the increase in head count from the expansion of the sales force, but secondly, we frankly have lumpiness in terms of the timing of the professional services that are required to run a public company, specifically audit work and SOX-related work, which is more in the third quarter than in the second quarter, where the bulk of that is performed. So all that, timing is scheduled out based upon, as our experience, what the contractual amounts are, and how we expect those services to perform. So that is the two major components for the change Q2 to Q3 on SG&A.

All right.

John, I would like to follow up on margin for a second. If you think about it, we have a fixed cost business, and because the growth rate of our off-net business has slowed, that does impact incremental margins, and the fact that our off-net revenue growth has actually accelerated, which is less margin favorable, I think that needs to be calculated in as well.

Okay. Final question, a bit of a housekeeping item here, could we actually get one more significant digit on the percent of revenue that comes from your corporate customers, please?

I don't have that on my sheet, Tad. It was just rounded to 44%.

It was 44%.

Yes, he wants to get, like, is it 43.9% or--

Bear with me. I have it.

I have to admit, John, I don't have the numbers to that level of detail.

Okay. We can always follow up with Keisha.

That is all right. It is 44.0%.

Thank you.

Thanks, John.

Thank you, sir. Moving on to Tim Horan with Oppenheimer.

Hi, guys. Had a couple of questions. Hi, Dave.

Hi, Tim.

Some of the sources I used to measure internet traffic growth, show that we haven't really seen a slowdown. Now some of these sources more, they measure out at the edges of the network, where I think it is a little bit easier to measure. Can you maybe give us your thoughts on why you think the volumes have been lower, does it have anything to do with maybe just exchanging traffic within data centers more, or more CDN traffic, just curious on some of your thoughts? I guess Level 3 would argue traffic go from IP back to private lines in some form, or other forms of traffic, which is directly opposite from what we have seen in the last decade, just kind of curious on your thoughts? Thanks.

Sure. Again, thanks, Tim. Let me try to take those in reverse order.

First of all, I think traffic will go to the technology and platform that delivers it at the lowest cost per bit mile, provided the quality is equal, and that bit mile includes the routing or switching as well as the transport. With internet transit prices where they are today, there is generally about an order of magnitude lower cost, to deliver traffic over a transit pipe, than either a private IP network, a wavelength network based on the utilization, or an ATM network.

So unless customers are going to pick the more expensive, not the less expensive alternative, I don't believe there is substitution of IP transit by those other products. I think the trend is exactly the opposite. IP transit, as you pointed out for the past decade, and will continue to become the dominant way in which connectivity occurs and traffic passes.

Now to the point of CDNs, some CDNs have experienced deceleration in their rate of traffic growth as well. Others have maybe been less transparent about that. CDNs are really hosting companies that still need to purchase transit from someone to connect. You need to get to the machine, and then you need to get from the machine to other networks and customers. The average bit travels about 2,700 miles across the Internet, and is distributed to literally thousands of locations. Few if any CDNs, have all of their content available at all of those locations.

So CDNs typically charge a premium per megabit mile for their services. They may be able to extract that premium because they are adding additional services, but if anything, I think transit, again, well, first of all, it double counts, because it counts the CDN business as well, and second of all, it continues to be the lower cost mechanism and capture more market share.

The final point was could cross-connects within a data center somehow cannibalize the transit market? Well, in fact cross-connects within the data center is critical to the data market. It is critical for two reasons. 96% of our traffic comes from net centric customers. Those customers buy cross-connects at about 350 data centers, to connect to Cogent, or some other network. So the cross-connect exists in the data center so the customer can get to a network operator.

Secondly cross-connects in the data center are critical because of peering or interconnection agreements. We are interconnected, as I said, with over 2,400 networks around the world. I think we are the most interconnected network when you look at size, connection, number of locations, and number of networks connected. That is very important.

However, it is interesting to note that the top 16 networks that connect to Cogent, actually exchange about 60% of the traffic with us. That means 40% is exchanged with all of those other networks. Smaller networks may elect to exchange traffic among one another, but major networks generally look for someone to pick up the load, and carry that bit, their fair share. Someone has got to carry the bit 2,700 miles. The content and the networks do not all coexist in the same location, so either you buy wavelength and connect, which is more expensive per bit mile when you look at the utilization rates and the current pricing, or you buy transit. That is why I think transit continues.

I think traffic slowdown is a different issue. Traffic slowdown is driven by business models. We have seen a proliferation of broadband connectivity, where we now have almost 80% broadband penetration in the western world. Line rates on those last mile connections have increased to close to 5 megabits of download speed, which is sufficient for most applications, particularly video. So you are not going to get an uplift from more broadband penetration, or greater download speeds.

What you need are more applications that consumers want to use more and more, and you have to look at the size of the installed base. At about 70 pedabytes a day, that base is very large. Many applications that people point to, could migrate to the Internet or increase, and not materially move the needle because the base is large. There are applications that will drive growth, and we have been pretty clear that that is really video. We see that with a number of customers and we see that trend continuing, but we have not seen the massive migration of video consumption over the Internet.

Today video is consumed about 4.05 minutes a day on the Internet. That is [pure] data research. And television, which is traditionally delivered via broadcast, satellite, cable or DVD is consumed 4.5 hours a day. Until those trends, there is some shifting in that relative mix, we will see I think smaller growth. It is not software-as-a-service, or e-mail, or even voice migration, that will materially drive growth. The Internet is not going to decelerate. It is not going to go away, but it is going to be a bit lumpy in the way in which it reaccelerates.

Thanks. Maybe we can talk offline about those trends. Appreciate your--

Sure. Give me a call.

Great. Have you assumed any decline in revenues on the net centric snide your guidance going forward? You are not seeing much in the way of revenue growth on the margin, kind of pricing is coming down because of the term and volume contracts. Just curious, what is embedded in that going forward?

I think as Tad said, we took all of the pricing trends that we implemented. We looked at the customers that took the contracts with us. So when we saw those customers increase their contract value, we also looked at how much of that came from more spend per customer, and how much of it came from contract lengthening. For those 200 customers that we talked about, in July sales that increased total contract value by $10 million, we saw their average contract length lengthen to about 18 months, actually 17.9, our installed base is about 11.8 months, so we did see some contract lengthening, but we did not see a reduction in their monthly spend.

And I also commented on our price per megabit sold, which actually had come down slightly from the previous quarter, from about $9.20 to about $9.00, so we are seeing some reduction in the price per megabit, as a result of the volume-based pricing and the term-base discounts, but the average monthly spend is actually going up. The only reason ARPU declined by the 2%, was because the mix of corporate customers as a percentage of on-net customers increased.

Can you give us those numbers again? What was the corporate customers of the percent of revenue a year ago, and maybe last quarter, and now, and what are you expecting going forward?

It was exactly 40% for the quarter, and historically if you look back--

44%.

I am sorry, 44%. If you go to the Q2 2007, it was 43.5 of revenue. Then it trended down, and then it has trended back up.

And if your volume is not growing, it seems like it is hard to grow the net centric revenue base overall. So it would seem like your corporate as a percent of revenue has to expand to have volume, to have revenue growth. Is that what you are expecting going forward, and are you maybe shifting your sales force towards that?

The mix of sales people has been pretty consistent at roughly two-thirds of the sales force has been focused on corporate customers, and one-third net centric. We expect that to continue.

I do disagree with your assertion that without volume growth, we cannot grow the net centric business, because we had the ability to capture a significant amount of market share. So we can grow either by existing customers taking more bandwidth. Clearly that is easier, or we can grow by winning customers away from other providers, which we are doing.

Part of the structure behind our volume-based pricing model was to accomplish both of those goals, and I think we are seeing that and as I said, we would have expected a bigger traffic decline in July than June, and in fact, because of the K-12 exposure, and in fact we saw a traffic increase in the month.

And it is hard to read into this a lot of data from monthly trends, but there are clearly a number of new business models, and many of them Cogent customers, who are exploring different ways of bringing video to the Internet, professional video. And some of them are more successful than others. So to directly answer your question, we can grow our net centric revenue, even in a moderate or slow traffic environment for net centric aggregate growth.

Thanks, Dave.

Thank you. Moving on to David Dixon with FBR Capital Markets.

Thanks for taking the question, Dave. Just wondered, I have three questions here. The first one is really, I apologize if I missed it, but could you comment on the magnitude of the increase that w ha've seen from new business from net centric customers in the third quarter so far, relative to the 15% that you had mentioned in the second quarter? And on this note also, just looking at the outlook for cost per meg, I noted that the average price on the basis is $9 per meg currently, I imagine that is a second quarter point of reference. What is your expectation for that as we look out over the next 12 months?

Sure. I will take the pricing one first. I would expect that number to go down. It has consistently gone down. That will be as a result of average contract lengthening. As we said, the average contract in the quarter lengthened by about 3%. That is pretty consistent with the rate of lengthening, and not nearly as extensive as we saw in the middle of last year.

Secondly, we would expect to see customers taking more and more of the volume based discounts on the net centric side, which would get them lower prices. However, that will probably increase the amount of non-used bandwidth that they purchase. So the effective price per megabit, i.e., not the price sold, but the price used, may in fact go up based on that. It is somewhat irrelevant to Cogent. All that matters to us is getting more revenue, with the relatively low utilization rate in the network.

Now with regard to incremental sales for the month of July to net centric customers, existing customers, and new business, I would say it is equivalent to what we saw in the most recent quarter, at that roughly 85/15% split. However, it can be a little lumpy within a month, and I am not willing to say that one month makes the quarter.

Just trying to reconcile, we have got obviously some larger customers that we have been speaking to that indicate that the price per megabit is in the $4 range now, and I am kind of trying to back into the outlook for the base going forward? I was trying to get a point estimate perhaps. I appreciate your point about used versus non-used, but where would you see it as a kind of point estimate in 12 months?

Sure. So first of all, you are absolutely correct. Our largest customers are at $4 a meg, and even some of our customers that are not our largest customers can be at that level, because to get that price point, a customer would enter into a three-year contract, and they would take a full 10-gig port, whether they used it or not.

Now clearly they would expect to have a high degree of utilization, whether or not that actually came to fruition could change. If I looked out over the next year, I would say that the rate of decline over the previous 12 months, which obviously did not have the volume-based pricing in it, was probably from about $9.60 to $9.00, based almost exclusively on the term discounts, not on the volume. I would expect that rate to actually accelerate and probably be greater.

However, there is a counterbalancing force, which is the continued focus on smaller net centric customers. That does have the tendency of pulling ARPU down. That will tend to mitigate that rate of decline. So I hate to predict the exact number, and be held to it because, I will be honest, I am happy with the revenue. I am not as concerned about the number, but I would suspect that average number probably will be in the $8.15 to $8.20 range a year from now, but that is just a guess.

Okay, Dave. And then just as a quick follow-up, just wondered if you could give us a sense of where you see the waves, the pricing for waves to be at more currently versus IP transit, because I thought that was very interesting commentary about IP transit, and potential at these levels we could actually start to see a shift away from peering back to IP transit, which obviously could be a benefit.

We have some major cable company customers, for example, who previously had a wavelength based network, and elected to shift back to transit because it was a lower cost when they figured their utilization rates, the cost of colocating and purchasing routers, maintaining them, and then maintaining peering connections, and buying transit for those carriers that would not peer with them.

And as I said, there is kind of a world of haves and have-nots. I think wavelength pricing falls because of the technology changes. However, that competitive universe has changed quite a bit, and there are many less players in the wavelength space. But there are also many less customers as well. Because the majority of the wavelength market was sold to companies who tried to build IP networks on top of wavelength.

So when the price per megabit was much higher, you could afford to buy routers, buy wavelengths and use them relatively inefficiently, and run an IP network and make money. What we have seen is a continued reduction in the number of global backbone, and particularly most of that reduction has come from those companies that buy wavelength.

I was talking to an EVP at a competitor who runs the sales organization, and we were just chatting, and anecdotally, we both concurred that the entire market globally for wavelength customers, this being corporate, as well as service providers, is probably no more than 300 to 400 organizations. CIearly Internet access and transit is a much bigger addressable market.

Okay, thanks very much.

Thanks, Dave.

And from Thomas Weisel Partners, we have James Breen.

I was wondering if you could just comment sort of on the competitive environment, and as you are going through the process, either bidding for new business or with existing customers, are you seeing the gap closing between you and the next highest competitor from a pricing perspective, how willing are the guys that had previously been pricing at a premium to come down to your price points? Thanks.

Thanks, Jim. Prices do continue to fall, and the gap has closed. Remember, when Cogent started selling, we had a 30:1 advantage. Today we have probably a 2.5:1 advantage, so that is a much smaller gap. With that said, there are a whole lot less competitors. So we see a lot less bid activity on large opportunities, with generally only two or three other bidders, generally two other bidders and ourselves in most opportunities. Sometimes there is a second tier player in there as well.

I would say those other carriers have gotten more aggressive, yet they have not matched Cogent's price points. They have not even matched twice our price point. I can't really comment on why that is, but I do know that they are lowering their rates faster than the 30% that they have indicated, but still not fast enough to get down to even 2X our price point. But we anticipate pricing falling, and that is built into our model, and kind of today's point when he got me out on the limb, that $8.00 number kind of assumes all of these trends.

Great, thank you.

Thanks, Jim.

Thank you. Moving on to Tom Watts with Cowen and Company.

Good morning, Dave.

Hi, Tom.

I know in your guidance, in the revenue guidance, a number of different factors, and it sounds like you assume the same revenue, the same traffic flowing you saw this quarter continuing. If there were further traffic flowing, would that lead to revenue projections lower than your guidance?

Sure. We have built some margin in Tom, this is not a precise science. As I said, with a greater percentage of our business coming from corporate, where there is no correlation between traffic, we have seen the ability of the business model, to basically immunize itself from a slowing traffic environment, and that is why in fact with negative 1% traffic growth, revenues grew 3.4%. On our projections, we have assumed a pretty pessimistic view. We took kind of July's type numbers and projected them out.

Remember, we don't have any kind of long-term history of negative traffic growth. We did factor the negative numbers from June in as well. I think we could sustain a 5 or 10% reduction in traffic quarterly, and meet our numbers that we have outlined. If Internet traffic fell by 50%, we would be back having a different conversation, but we have no imperical evidence that that is happening, and in fact, the 1%growth that we saw in July, while not great, is clearly better than the negative 3% that we saw in June.

Okay. Then you talked about some of the factors that are going to pressure EBITDA margins in the second half. To the extent that you wanted to, they turned out great, or you wanted to take offsetting factors to reduce costs, are there areas other than sales where you could have additional cost reductions?

I will take part of that, and I will maybe let Tad take a little bit as well. We do have a fixed cost business. There is little variability other than our off-net business, and in that off-net business we do have 50% gross margins, so that is still additive to cash flow. It may be dilutive to incremental margins, but it is still in aggregate a good business. The only variable cost really in Cogent's business beyond those customers, is the cost of sales, which is relatively small. I look at our business and don't believe we have a cost issue. It is really can we continue to grow the top line, and can we even accelerate that rate of top line growth. Tad, maybe you want to comment.

Yes, maybe just to reemphasize, it really is a fixed cost business, with the exception of the variability with the hiring rate. We talked about the range of bad debt expense, although the performance is good, you can be from 0.5% to more than 1.5%, and then the lumpiness in terms of the timing of professional fees. I mean when I look at what we had in the plan for the second quarter ,versus what was actually achieved, with the combination of cost of goods sold and SG&A combined, it was a 1% difference, and that is very small on about $36 million of aggregate costs.

So fortunately for us, I think within a very small range, it is easy to forecast what we expect. Unfortunately I guess to your direct question on offsetting factors, there are not a great deal of offsetting factors. I think it is relatively fixed base with the increases that we talked about. I guess we didn't mention, the timing of expansion of routes. As soon as you accept a fiber route, you will have the maintenance charges incurred with that, albeit those numbers are not very significant to the second half of the year. That would offset to the extent that that timing is delayed, or that would accelerate to the extent that that timing is accelerated.

Okay, and then just finally, particularly with the off-net business, and the potential for growth there, are there any trends that you are seeing in tailed circuit pricing? Is there any upward price around that? Is there any potential to get more leverage on that to make your off-net more competitive?

We buy generally at tariff. We do shop. We buy from over 150 different tail circuit vendors, where there is facilities-based competition, but for the most part there isn't a lot of competition. I would say tail circuit pricing is basically flat, which gives us a stable price point.

Our ability to lower costs would require us to take on some additional regulatory burden, and also make additional volume commitments. We are uncomfortable with either of those strategies, and are content with the 50% gross margin in the off-net business, and are happy to grow with that. I just think the mechanisms to improve that margin are not worth the risk.

Okay. Thanks very much, Dave.

And Robert Weaver with Forest Investments has our next question. Mr. Weaver, your line is open. Mr. Weaver, perhaps you have your mute function turned on?

You may want to go on to the next question, if there are any others.

Hearing no response, we will move on to Andrew Beale with Arete Research.

Hi. I have got three questions, if that is okay. First of all, in terms of the net centric volume trends that you gave us from April through July, are you sure that those are sort of the end user experiences of volume trends, or do you think that there is a market share overlay that we should be thinking of on top of those?

Sure. I will start with that one. There is also a market share overlay, but because we are the lowest price provider, and because maybe the economy is a little more difficult than it was in the past, I think we are probably getting a bigger share.

Now we also have anecdotal evidence from customers that we are getting a bigger share of their business, and then finally, with the volume-based initiatives that we put in place, the early adopters of those initiatives were existing customers, who were shifting us even more share, because they couldn't generate more traffic in that short period of time, almost all of that increased volume that they gave us came through shifting. So I do actually think it is buffeted somewhat, and I think the end user experience is actually probably a little more extreme.

Okay. Then secondly, in terms of your volume pricing, can you give us an idea of the drag on on that revenue that that caused you in Q2, and what the effect will be in the full quarter in Q3, perhaps you want to tell us on a gross and net basis, after the volume improvements that you anticipate?

First of all, it had no impact on our results in Q2, simply because there were virtually none of the orders installed, and when we looked at the orders installed to date, we actually see no negative impact in our guidance for Q3, simply because of the contract value requirement for existing customers taking more bandwidth. So there is little or no exposure to getting less monthly revenue from an existing customer. We have no data that says that is in fact what is happening.

In fact, most of the customers are our largest customers who have taken advantage of this, have done so by increasing their monthly spend with Cogent, and with regard to winning new business in the market, it is really hard for me to tell whether or not it is business that the volume-based pricing was responsible for us winning, or we would have won it without it. But we have factored no drag in our guidance.

So essentially it just gives you less sensitivity to whatever the volume trend is through the year, through the summer quarter, is that right?

That is correct.

And then just moving to the sales force, how have you refocused them on corporate, in particularly selling off-net corporate? Is there something that you have done differently in terms of the incentives that you offered them perhaps? I don't know.

Yes, it is actually the exact opposite of what you would have expect, which we have disincented them from selling off-net, because they only get paid on the gross margin, which does not include the local loop. It means they only get paid on the port that they sell.

Now we had not done that until March, so what you saw was business that was sold in part of the first quarter, fully take effect for the full second quarter, but we also continued to see a significant set of new bookings of off-net services, and particularly those with higher loop capacity, and therefore higher ARPU for us, so in fact the sales force is disincented to sell off-net, yet we continue to see that as a growth business. It is a huge addressable market, and it is a market in which traffic volumes play little or no role.

So I mean is it fair to say that the sales force see it as just a bigger opportunity, even if they get paid less dollars per dollars monthly revenue?

That is correct, although the majority of our sales, as Tad said, continues to be our on-net product, with about 85% of sales being on-net.

Okay. Thank you.

Thanks.

And gentlemen, we have no further questions. Mr. Schaeffer, I will turn the conference back over to you.

All right. Well, thanks everyone. I know it was a long call, but a lot of good topics covered, and we look forward to continuing to deliver good results, and chat with you guys next quarter. Take care. Bye-bye.

And that does conclude today's conference call. Thank you for your participation. Have a great day.