Home » Broadband "Shortage" »Consumer News »Internet Overcharging » Currently Reading:

Broadband Costs Continue Accelerated Decline; Provider’s Real Cost for Your Usage: $1/Month

Phillip Dampier August 7, 2012 Broadband "Shortage", Consumer News, Internet Overcharging 12 Comments

Broadband transport costs continue to decline, at an accelerating pace, according to researcher Telegeography.

Prices to move data across the Internet continue to decline throughout the world. According to new data from TeleGeography’s IP Transit Pricing Service, price declines in most locations accelerated over the past year, at an accelerating pace. But none of those savings are showing up on customer bills. In fact, while providers have been increasing broadband prices over the past three years, their costs to provide the service continue to plummet.

“IP transit prices have reached extremely low levels in developed markets, but remain high in many developing markets and in countries that are remote from major IP transit hubs,” said TeleGeography analyst Erik Kreifeldt. “Nevertheless, few places remain where transit prices exceed $100 per Mbps. As carriers expand into emerging markets and establish new price floors in developed markets, global IP transit prices will continue to fall.”

The median monthly lease price for a full GigE port in London dropped 57 percent between Q2 2011 and Q2 2012 to $3.13 per Mbps, compared with a 31 percent decline compounded annually from Q2 2007 to Q2 2012. In New York, the comparable price dropped 50 percent to $3.50 per Mbps over the past year, and 26 percent compounded annually over the five-year period. Pricing for short term promotions and high capacities have dropped below $1.00 per Mbps per month.

DSL Prime‘s Dave Burstein says that translates to Internet backbone wholesale pricing of less than $0.50 per broadband customer per month in New York or London.

Burstein also notes router and switch prices are also matching the predicted pace of Moore’s Law, declining 25-40 percent per annum. With competition for backbone connectivity robust in North America, the reduced costs are passed along to large broadband providers, but not to customers.

Burstein reports that while Internet traffic continues to expand at “ferocious rates,” your broadband provider’s net cost has been generally flat or even down. In fact, he estimates that when providers add up the cost of backbone transport costs and moving traffic from their network to individual customers, they end up spending less than $1 per month on traffic per customer. But they charge you $40-50 or more for the service.

Burstein also notes that broadband usage has almost no impact on provider costs, whether they offer 3Mbps or 1,000Mbps service, have caps of 50GB, 500GB, or no caps at all.

“With bandwidth costs this low, we’re talking dimes or at most a couple of dollars per month to handle any likely traffic flow,” Burstein reports.

Even accounting for perennial predictions of data tsunamis from equipment manufacturers like Cisco, their own data shows the primary cost of Internet traffic per customer is falling, according to Burstein, even as data consumption increases.

Share

Currently there are 12 comments on this Article:

  1. txpatriot says:

    1. Transit costs are not the same as access costs. The difference is similar to the difference between being a long-distance provider and being a local telephone company. The two are entirely different.

    2. If the access business was as cheap as this story implies, everybody would be getting into it.

    3. @Danny Burstein: ask google how cheap are their access costs in KC. Again, if it were cheap, everyone would be doing it.

    • I know you know perfectly well the reason why everyone is not getting into the ISP business: the initial start-up infrastructure costs.

      The cable and phone companies benefit from the fact their outside plant has been already built and in most cases paid for. The biggest costs are wiring the last mile. Verizon is the one company willing to massively upgrade that last mile in larger populated segments of their service area and they were pummeled for it on Wall Street, and now the grand upgrade is largely over.

      The larger point here is there is absolutely zero justification for incumbent providers to be raising rates, much less blaming consumption for increased costs that don’t exist. The phantom case for usage caps or usage billing has been exposed for the fraud it is.

      BTW, the dollar a month cost per customer noted above covers backbone -and- last mile distribution costs. These costs are obviously higher for very small ISPs who cannot get the volume discounts Comcast, Time Warner Cable, AT&T and Verizon get, but most Americans are vastly overpaying for Internet service and will continue to do so until deep pockets like Google can ram the case for competing networks through the Wall Street nabobs of negativism.

      • txpatriot says:

        No doubt the initial infrastructure start-up costs are high — but not so high as to stop google from entering the market.

    • Scott says:

      Obviously the $1 doesn’t include the “cost of business” to operate a ISP and is just the transit, nor as Phillip pointed out is it accurate for smaller ISP’s – especially those not located near large distribution points with access to cheaper transit.

      A smaller ISP may be paying anywhere from $10 to $80 per Mbps, compared to the quoted $3.50 rate for the big guys which carry the majority of the traffic over their own fiber or free peering agreements.

      What it does show is the declining and low cost of the large corporate Cable and Telco’s that do have access to those resources who have consistently been raising subscription costs, forcing bundles to get discounts, and charging punitive fees for “overage”.

      Cheap transit (which no new ISP will get until your commit is well into the 1000Mbps+ range except in very rare markets) has little to do with starting up an ISP since you’d be going up against the AT&T’s, Time Warner, and Comcasts of the industry. Those guys have the political connections and money to lobby and run you out of town, and they have the deep pockets to discount their services for up to 2 years locking in customers with contracts that you can’t ever match without losing money yourself. Once you get noticed by those guys, the gloves come off.

      You need to be at the level Google is to do this and self-finance and fight against entrenched players.

      If I thought I could make a difference again, I would startup another local ISP but there’s too many of the above issues plus the drawbacks of being in a remote city where my only options for transit are the 2 local competitors. For now I’m much more content with the profit margin from racks of servers in the US and Europe with transit

  2. Eddie says:

    Part of what this shows, I’m sure, is that in the US, the actual cost of providing the service is only a fraction of the cost of running the business. I am not defending the major US ISPs… the US pays more for less than most developed countries, and that is deplorable. For example, the cost of crude oil is between 1/3 and 1/2 of the cost of gasoline, the rest is refining, transportation, overhead, and so on. I would love to see a cost breakdown of the US major ISPs, cable, wireless, telco, and others, that shows what fraction of their cost is actually carrying network traffic, what fraction of their costs is overhead (labor, office buildings, etc), and so on.

    Your data suggests that a smaller and smaller fraction of their cost is the actual cost of carrying the data, but it is not enough to prove that their total overall costs are dropping. I’m not saying you are wrong about that. I honestly don’t know. You may be right. But your data doesn’t prove either way, because it’s possible their other costs are going up by enough to compensate.

  3. Phil
    Thanks for including me in your reporting.

    The comments are right – the costs I’m covering in this article are for the bandwidth itself, as well as ferrying it from the Internet access point to the neighborhood DSLAM. There’s plenty more involved in the cost of delivering the service, which has a marginal cost of $5-10 at any large carrier and a fully-loaded cost that’s of course higher.

    The bandwidth costs (transit/peering + provider’s backhaul to DSLAM) are a crucial number because that is the only difference in cost between slow service and the fastest service your local loop can provide. They are so low that in more competitive countries (like France) everyone gets higher speeds for little or no extra cost.

    That’s why Google is giving all paying customers a gig. It would save so little giving them slower rates it doesn’t make sense. I know two other (small) fiber carriers that have made the same decision to give a gig to everyone, soon to announce.

    I’ve actually looked at Google’s costs and other fiber builds. Running new fiber is brutally expensive, far outweighing anything like bandwidth costs. The $70/month price Google is seeking is a little higher than I hoped but not crazy considering the costs. Fiber in favorable conditions costs about $1000/home passed. With a 33% take rate, that’s about $3,000 per subscriber. Amortized over 5 years, that’s $50/month. Actual numbers vary considerably neighborhood to neighborhood.

    Small and rural carriers are often totally screwed, as I’ve often reported. The numbers here are for the medium and large carriers which serve 80-95% of the customers in almost every country.

    —————-

    As Eddie notes, bandwidth is such a small factor it doesn’t tell you overall costs. Net of increased traffic per user, the cost per month is flat to slightly down.

    I’ve elsewhere looked at overall as well, and they appear to be marginally down over the last few years. Modems and DSLAMs keep getting cheaper. Support costs have been going down as well, because new customers need much more support and the percentage of new customers is dropping. Labor rates are generally up and probably electricity.

    The overall conclusion on costs is that the change the last few years has been modest. That means that changes in prices are primarily due to something else. My impression is that competition is the key driver. Where competition is weak (U.S., now Japan) prices have gone up, with Verizon’s 30-40% increases in DSL the most dramatic. In most of the world, they’ve generally drifted down.

    Strong direct government action can make a difference, and has in Taiwan and now China. Indirect government action, encouraging competition, also can work – see the French wireless prices. Most regulators are so influenced by the industry they don’t take strong action, however.

    db

    • I think we’re both on the same page here although I may have tried to simplify the concept for readers to the point of fuzzing the distinction. I realize the piece I wrote only speaks about transport/bandwidth costs. It does not talk about infrastructure, construction, maintenance, marketing, et al.

      Last I had seen Verizon had gotten FiOS fiber builds down closer to the $700 level, but that is probably because of vicious control of their vendor costs, quantity buys, experience, etc.

      The larger point I am making is the ludicrousness of slapping caps and consumption billing for the traffic element which has simply gotten ridiculously cheap for larger carriers.

      As far as costs are concerned, I base that statement on the financial reports of the major providers, which I read every quarter. Outside of a minor uptick in CapEx when DOCSIS 3 was rolled out, and some fiber/backbone expenses for phone companies, spending on broadband is either flat or declining across the board except in one case — spending on infrastructure to wire businesses and office parks. The cable industry sees new business marketing commercial services to business customers and is spending on infrastructure to reach the currently unreached.

      Even providers have said their major infrastructure spends are over… time to reap the rewards.

      • txpatriot says:

        Dave writes: “The overall conclusion on costs is that the change the last few years has been modest. That means that changes in prices are primarily due to something else. My impression is that competition is the key driver.”

        I appreciate the fact that Dave draws a distinction between costs to the carrier and prices to the end-user. Far too many pundits toss those two terms around as if they were the same thing, which they are not.

        Costs are costs and don’t change just because new competitors enter the market. And as Dave notes, putting fiber in the ground is a VERY expensive proposition. Trenching costs do not follow Moore’s Law, and no amount of wishing by anyone will change that.

        The next time someone says costs are coming down therefore prices should come down needs to have their head examined.

  4. Austin says:

    “DSL Prime‘s Dave Burstein says that translates to Internet backbone wholesale pricing of less than $0.50 per broadband customer per month in New York or London.”

    I’ve always been critical of how ISPs add up their expenses. What surprises me about this is that it seems that when you divide their backbone bill among all their customers, one comes to such a low number. I’ve always considered that this is where the prices were conditionally justified.

    Two networks will peer and, on a scoundrel’s handshake, agree to bill each other an astronomical amount of money per gigabyte, minus credit for traffic peered. When this comes out uneven, one peer gets billed a fairly insignificant amount of money (for them). But small ISPs are driven out of the market by their inability to negate the bill through peering, or to afford the arbitrary fee which the big boys pretend is the market rate. It was my thought that ISPs added their backbone bill without including their credits, much the same way corporations complain about their high effective tax, despite the fact that they exploit enough loopholes that they never pay that much.

    No one but the backbone providers knows how much it costs to run their operation, but it’s not linked in any way to the amount of data they pump. It’s linked to the cost of building out (probably long paid off now), maintenance, and paying the power bill. The cost of these things, we know, are plummeting due to better technology like faster, cooler switches that take up far less room than earlier generations of hardware. It costs the ISPs no more to provide that “extreme” 12Mbps connection than it does to provide the “pro” (“up to”) 3Mbps tier which it sabotages so that it never exceeds 85%.

    So if, for example, so many small DSL ISPs could run their operations on a shoestring, billing customers just a little more for the bandwidth than it costs them to buy it from their upstream provider at full price, then it’s reasonable to assume the cost of operations and maintaining the last mile is relatively small. It’s likely that monster ISPs are operating at an even higher level of efficiency. And if the unmolested cost of the bandwidth is miniscule for the monster ISPs, then there can be no conclusion to make other than that they are raping their customers by setting service prices so high. This is not the market at work, this is monopolizing.

    And of course, with no real competition in a monopoly (or duopoly), the big ISPs have no incentive to charge fair prices or to use their gargantuan profits to upgrade their network. They make more profit by keeping prices up, and pretending bandwidth is scarce. That’s called an embargo. They spend a pinch of their profits on perverting the law in their favor and keeping the cash cow’s blood spilling into their slop troughs.

    What we need is a game changer like Google in every city. Google probably forecasts paying off it’s network in a relatively short period of time. And they’ll do it without the government subsidies which current ISPs enjoyed, and they’ll do it with service prices far lower than the current ISPs demand. More evidence that current cable and DSL prices aren’t just a little unjustified, but criminally inflated.







Search This Site:

Contributions:

Recent Comments:

  • RG: You can thank the greedy cable companies for that. Cox for one, was supposed to support VOD on the Premiere with an agreement signed with Tivo about 4...
  • Tony: We cannot use On Demand (VOD) via Roamio. It is maddening....
  • KF: Tony, I also have Charter and have not been able to get On Demand (VOD) to work through the Roamio. Charter insists that VOD is impossible with the R...
  • Phillip Dampier: If enough U-verse customers say goodbye over data caps, there probably won't be data caps. They are just there to further monetize their operation, wh...
  • txpatriot: Did anyone notice a slowdown yesterday? I didn't (other than everything being normally slow)....
  • Andrew Letson: I can't believe I got cited by you guys - imagine my surprise when I was looking through my WordPress trackbacks to see a link back to here! I feel li...
  • Dave Hancock: OK, did that, but it seems that Verizon is slowing down submissions to the FCC or something!! When I hit confirm, the site just sits there waiting for...
  • tacitus: Thank you for your reply, Phillip, and for investigating the deal further. Yes, the line rental is a fact of life for the vast majority of the British...
  • Bill A: This is great news! Finally some competition to the monopolies! I just signed up. Installation is in a couple weeks. I'm so happy to support a com...
  • Phillip Dampier: The presumption here is that the vast majority of the British public already maintains a landline connection primarily for broadband (DSL and its vari...
  • m1817: I was paying $25.50 for 12Mbps on a 12 month term. When the 12 months were up, I was offered and accepted an upgrade to 24Mbps for $28 per month on a...
  • tacitus: Phillip, as much as I admire your diligence and accuracy when it comes to reporting on the broadband and telecom industry in the USA, you have consist...

Your Account: