SEACOM pricing – how much does it really cost

[brianherlihy]

SEACOM Pricing

Dear All,

it is very interesting to read all of these posts. i think some very good points are being made but also would point out that there are multiple ways to look at pricing to the end user.

first, I come from a very different pricing structure in the US. We don't have the concept of Gigabit caps or pricing per data. we price on the throughput we receive from our carrier with unlimited rights to download data. The SEACOM model is also based on throughput, therefore our pricing and the end user pricing in SA are not an apples to apples comparison.

If we look at the throughput model for a minute to the end user, you have to look at the total cost as formulated below:

end user price = last mile + metro area access/national access + international access + services / (divided by) contention ratio. In other words, when you do the calculation of the SEACOM cost to the carrier, (as someone pointed out), you are doing it on a 1:1 contention ratio. This is not the service level that a carrier provides you (anywhere in the world) but better service level agreements have better contention ratios. The pricing model i am used to from the States (as an end user) would get anywhere from a 20:1 to 30:1 contention ratio - best in class i 10:1. To complicate the matter, carriers will do different contention ratio for different parts of the solution (i.e. last mile, metro, international) and use statistics to hope the customers don't overload usage at any peak time. What has been discouraging for me is not the data caps but that the through puts are a fraction of international standards.

The model SA is used to is limitation of data or price per data. the pricing that Suveer presented let's a person do a conversion from throughput to data usage by calculating the total amount of Gibabits available on an STM-1 circuit over a 24 hour period but this assumes 100% utilization which is not the case (hence why i said an apples to apples is difficult).

I have been the biggest proponent of SA needing more bandwidth because ultimately those through puts aren't what they should be (and i don't like living in a capped data world when i visit - FYI my girlfriend blew through 9 Gb/s in 10 days on her first trip to SA because she was downloading TV shows every night on ITunes - without our knowledge). But to some extent, SA is much better off, furthermore, carriers are re-investing in the national and metro networks and are using some of the savings from SEACOM to make those investments. i do understand frustration of the trickle down effect (and SEACOM is dependent on SA becoming a big broadband market) but i do believe that carriers and ISPs are making necessary investments now to launch a much bigger broadband world going forward. ICASA and the Altech case opened the door for that competition and we are starting to see it.


Brian Herlihy

 

[Sting]

Thanks Brian, great first post and good to see you here!!

 

[rpm]

Hi Brian

Thanks for the valuable feedback and giving us some hope that we are well positioned to improve the state of broadband in the country. I hope your girlfriend did not use a mobile broadband connection for her series downloads In future you should drop het off at the Neotel data centre for direct SEACOM access – you guys build the cable after all

 

[poolmania]

you contention ratio is proball somthing like 20:1

I knew I was missing something. Thanks.

 

[pmbellis]

Rather a shocker

Firstly, kudos to MyBB for these figures, they certainly shed a lot of light on the subject.

Now, unfortunately, I realise that cheap broadband is a pipedream for a long time to come. If R30.00/Gb +/- is the bottom figure for International bandwidth, then I don't see the retail price dropping much below R50.00/Gb in the near future. Even when the other cables come online, won't they have the same sort of pricing structure?

How do other cables (USA to Europe, for instance) compare with their pricing? Does anyone have a similar breakdown for them?

 

[greggpb]

Dear All,

it is very interesting to read all of these posts. i think some very good points are being made but also would point out that there are multiple ways to look at pricing to the end user.

first, I come from a very different pricing structure in the US. We don't have the concept of Gigabit caps or pricing per data. we price on the throughput we receive from our carrier with unlimited rights to download data. The SEACOM model is also based on throughput, therefore our pricing and the end user pricing in SA are not an apples to apples comparison.

If we look at the throughput model for a minute to the end user, you have to look at the total cost as formulated below:

end user price = last mile + metro area access/national access + international access + services / (divided by) contention ratio. In other words, when you do the calculation of the SEACOM cost to the carrier, (as someone pointed out), you are doing it on a 1:1 contention ratio. This is not the service level that a carrier provides you (anywhere in the world) but better service level agreements have better contention ratios. The pricing model i am used to from the States (as an end user) would get anywhere from a 20:1 to 30:1 contention ratio - best in class i 10:1. To complicate the matter, carriers will do different contention ratio for different parts of the solution (i.e. last mile, metro, international) and use statistics to hope the customers don't overload usage at any peak time. What has been discouraging for me is not the data caps but that the through puts are a fraction of international standards.

The model SA is used to is limitation of data or price per data. the pricing that Suveer presented let's a person do a conversion from throughput to data usage by calculating the total amount of Gibabits available on an STM-1 circuit over a 24 hour period but this assumes 100% utilization which is not the case (hence why i said an apples to apples is difficult).

I have been the biggest proponent of SA needing more bandwidth because ultimately those through puts aren't what they should be (and i don't like living in a capped data world when i visit - FYI my girlfriend blew through 9 Gb/s in 10 days on her first trip to SA because she was downloading TV shows every night on ITunes - without our knowledge). But to some extent, SA is much better off, furthermore, carriers are re-investing in the national and metro networks and are using some of the savings from SEACOM to make those investments. i do understand frustration of the trickle down effect (and SEACOM is dependent on SA becoming a big broadband market) but i do believe that carriers and ISPs are making necessary investments now to launch a much bigger broadband world going forward. ICASA and the Altech case opened the door for that competition and we are starting to see it.


Brian Herlihy

Thanks Brian, very informative.
++Rep for taking the time to respond on a public forum without going through a PR company

 

[Gaz{M}]

you contention ratio is proball somthing like 20:1

Even more sickening is that this uncapped bandwidth is simply unused business bandwidth that sits idle at night time which has already been paid for by some company. Mostly pure profit in my opinion.

 

[guest2013-1]

Dear All,

it is very interesting to read all of these posts. i think some very good points are being made but also would point out that there are multiple ways to look at pricing to the end user.

first, I come from a very different pricing structure in the US. We don't have the concept of Gigabit caps or pricing per data. we price on the throughput we receive from our carrier with unlimited rights to download data. The SEACOM model is also based on throughput, therefore our pricing and the end user pricing in SA are not an apples to apples comparison.

If we look at the throughput model for a minute to the end user, you have to look at the total cost as formulated below:

end user price = last mile + metro area access/national access + international access + services / (divided by) contention ratio. In other words, when you do the calculation of the SEACOM cost to the carrier, (as someone pointed out), you are doing it on a 1:1 contention ratio. This is not the service level that a carrier provides you (anywhere in the world) but better service level agreements have better contention ratios. The pricing model i am used to from the States (as an end user) would get anywhere from a 20:1 to 30:1 contention ratio - best in class i 10:1. To complicate the matter, carriers will do different contention ratio for different parts of the solution (i.e. last mile, metro, international) and use statistics to hope the customers don't overload usage at any peak time. What has been discouraging for me is not the data caps but that the through puts are a fraction of international standards.

The model SA is used to is limitation of data or price per data. the pricing that Suveer presented let's a person do a conversion from throughput to data usage by calculating the total amount of Gibabits available on an STM-1 circuit over a 24 hour period but this assumes 100% utilization which is not the case (hence why i said an apples to apples is difficult).

I have been the biggest proponent of SA needing more bandwidth because ultimately those through puts aren't what they should be (and i don't like living in a capped data world when i visit - FYI my girlfriend blew through 9 Gb/s in 10 days on her first trip to SA because she was downloading TV shows every night on ITunes - without our knowledge). But to some extent, SA is much better off, furthermore, carriers are re-investing in the national and metro networks and are using some of the savings from SEACOM to make those investments. i do understand frustration of the trickle down effect (and SEACOM is dependent on SA becoming a big broadband market) but i do believe that carriers and ISPs are making necessary investments now to launch a much bigger broadband world going forward. ICASA and the Altech case opened the door for that competition and we are starting to see it.


Brian Herlihy

Thanks Brian, just a quick note on SA Telecoms. Telkom was state owned until the late 90's when they broke off and became a "private" company.

My great grandad and grandad paid for the network Telkom is using to capitalize and overcharge us with. Having these companies re-investing my cash into their networks to keep charging me over-the-top prices just so they can provide better infrastructure *to who knows* is a bit stupid.

I would love to see your pricing model that you're used to and have implemented for Seacom be used in South Africa. It has long been discussed here that we would rather have shaped and uncapped access internationally OR unshaped and capped access.

We currently pay good money per GB but have to make do with shaping as well. And pay a premium to be unshaped. The two shouldn't go together.

I hope to see more posts from you in future as it's a breath of fresh air having someone as high up as yourself be involved in a forum that Telkom didn't even dare touch or acknowledge for the past 7 years

 

[ambo]

How do other cables (USA to Europe, for instance) compare with their pricing? Does anyone have a similar breakdown for them?

There isn't really a comparison to be made there. In the US and Europe most (more than 90%) of the traffic that they use is what we would refered to as 'local-only'.

 

[boerseun]

From what I can read the cost of R325 per month per Mbps does not mean that 1MB of bandwidth cost R325, it means that 1Mbps costs R325 per month. So basically You can get 1MB/s uncapped at R325 per month!!

 

[poolmania]

From what I can read the cost of R325 per month per Mbps does not mean that 1MB of bandwidth cost R325, it means that 1Mbps costs R325 per month. So basically You can get 1MB/s uncapped at R325 per month!!

But it doesn't cost that. The article clearly said that it costs +R700 per month and that doesn't include all the other extra charges that come on top of it. Read it again slowly and carefully.

 

[Cloud]

But it doesn't cost that. The article clearly said that it costs +R700 per month and that doesn't include all the other extra charges that come on top of it. Read it again slowly and carefully.

Ok, I might be wrong, but at 20:1 contention, even if it costs R1000 for 1mbps per month - that's still only R50 per month per user correct?

At 30:1 you are looking at R33 per 1mbps per user per month - uncapped.

So a 4mbps uncapped line should cost about R132.

I could be wrong.

 

[SaiyanZ]

Firstly, kudos to MyBB for these figures, they certainly shed a lot of light on the subject.

Now, unfortunately, I realise that cheap broadband is a pipedream for a long time to come. If R30.00/Gb +/- is the bottom figure for International bandwidth, then I don't see the retail price dropping much below R50.00/Gb in the near future. Even when the other cables come online, won't they have the same sort of pricing structure?

How do other cables (USA to Europe, for instance) compare with their pricing? Does anyone have a similar breakdown for them?

The R30.00/GB is calculated with a contention ratio of 1:1. Normally the ratio is around 20:1 so it would be R1.50/GB. With a ratio of 30:1 the price would be R1.00/GB. Much cheaper than we're paying for right now. Add in costs and a profit margin of 100% and R2.00/GB is reasonable at a contention ratio of 30:1.

 

[TeleGuru]

Missing the obvious?

I'd like to highlight some things that I've read that most of the poeple here, who I assume are all quite intelligent, seem to have missed.

1.) Seacom Proffit Margin
Its just interesting that no-one has picked up on this yet.

• Seacom is selling 1 x STM64 (10 Gbps) IRU for $99m, which is also the level at which you get the "best price" (R325/Mbps/Month calculated using Seacom Maths, at 100% utilisation from day 1 and assuming the money was free to borrow).
• They have a Total of 1280Gbps, which is 128 x STM-64s.
• The cable cost them $600m to build.

So if they sell 6 x STM-64's they basically have all their money back that they paid to build the entire cable, while they still have an additional 122 x STM-64's left, which is then essencially "free" for them.

Using Seacom Maths again, that gives them a 2011% profit margin.
Cost: $600m / 128 (STM-64s) = $4.69m/STM-64.
Profit Margin = (Sales Price - Cost) / Cost * 100 = ($99m - $4.69m)/$4.69m * 100
Profit Margin = 2010,8%

And people are calling Telkom greedy?


2. Cap vs Contention
I've also seen that people are confusing two different concepts:
Usage Based (Capped) based Costing
vs
Contention / Oversubscription based Costing

You have to choose either the one, or the other, but you cannot apply both.

Usage Based:
Capped Accounts / Usage Based Accounting is another way of oversubscribing your customers. While a 1Mbps link being used at 100% 24/7 wil lbe able to transfer 331GB/month, a 1Mbps link being used for usage based accounts will only transfer about 220GB/month. This is as a result of the mindset subscribed below, where you need to have space capacity in your network at all times. A usage based network, running at 100% utlilisation, will provide a horrible service.

The mindset for this type of Service Porvider is usually: How do I need to build my network to enable my customers to download as much as they can, in as little time as possible, seeing as I get paid for what they use. In essence: The slower my network, the less money I will be able to generate. To make more profit, I need to have enough bandwidth at all times.

Contention:
If you have 10 x 4Mbps Uncapped ADSL Subscribers which you serve using a Single 8Mbps pipe to the internet, you are effectively oversubscribing your customers 5:1. They are open to use as much as they like, but between them, they will fight for bandwidth on the 8Mbps pipe to the internet.

The mindset for this type of Service Porvider is usually: What is the minimum amount of bandiwth that I need to purchase, in order to provide a "good-enough" service to my customers, and be just "good-enough" for them not to complain. My income is fixed, independant of hum much my customers consume, to increase my profit, I need to reduce my input costs.

Hope my coments were usefull to those interested.

 

[MFour]

Very interesting indeed. Asuming SAT3 works on the same sort of principle, it again just illustrates how badly we have been done in by Telkom all these years, and that with the blessing of Gov. So they charge us for a unshaped, uncapped service which they resell to us as a capped, and shaped service, so severely limited that you can hardly call it broadband. Sad really.

 

[poolmania]

Ok, I might be wrong, but at 20:1 contention, even if it costs R1000 for 1mbps per month - that's still only R50 per month per user correct?

At 30:1 you are looking at R33 per 1mbps per user per month - uncapped.

So a 4mbps uncapped line should cost about R132.

I could be wrong.

So let's see:

A STM-1 (155Mbps) on a 1 year lease:

R737 446...... SEACOM Price
R12 000........ Cross connect fee
R9 000.......... Rack space
R755 000....... Local transit
R15 662........ London transit

Total R1 529 108 per month for a 155Mbps pipe to London.

Or R10 195 per Mbps per month. (With 150Mbps usable bandwidth)

1Mbps = 331 GB per month

So R10 195 / 331GB = R30.80 per GB

----------------

If the ISP buys an STM-1 IRU instead of leasing the capacity, then the cost is roughly R969 000 per month or R6462 per Mbps per month = R19.50/GB

There is your answer.

So at R10,000 per Mbps per month at a 20:1 contention ratio would be R500 per month.

If someone buys the capacity over the 20 year lifespan, then it would be about R350 per month.

This is however looking at the exact price. Of coarse it would be higher than that as ISP's need to add on their profits.

That's the way I understand it.

 

[Gaz{M}]

Using Seacom Maths again, that gives them a 2011% profit margin.
Cost: $600m / 128 (STM-64s) = $4.69m/STM-64.
Profit Margin = (Sales Price - Cost) / Cost * 100 = ($99m - $4.69m)/$4.69m * 100
Profit Margin = 2010,8%

And people are calling Telkom greedy?

Don't forget that SEACOM currently has around 80Gbps lit capacity. The 1.28Tbps capacity would require many more interface cards across the network at each landing station, at a large additional cost. So technically, you can't use the 1.28Tbps in your cost calculations.

If you use 80Gbps = 8 x 10 Gbps @ $99 Million, you get $792 million which is roughly the cost of the cable plus a margin and over 20 years, remember.

 

[Gaz{M}]

The R30.00/GB is calculated with a contention ratio of 1:1. Normally the ratio is around 20:1 so it would be R1.50/GB. With a ratio of 30:1 the price would be R1.00/GB. Much cheaper than we're paying for right now. Add in costs and a profit margin of 100% and R2.00/GB is reasonable at a contention ratio of 30:1.

One thing i don't quite understand is this usage vs. bandwidth argument. If I am an ISP and I have 100 000 customers on 3GB packages, don't I need 300 000GB worth of data at a minimum per month to serve them all? Using the contention reasoning, you are saying I only need (30:1) 10 000 GB to serve these customers? Or in effect, a 30 Mbit pipe 24/7?

I do see that having a 100% utilized pipe would give a pathetic, slow service to everyone at peak times, but how do you dimension otherwise?

If you look at things from the other angle, saying you have 100 000 customers on 512kbps ADSL @ 30:1, this means you need 1.7 Gbps to serve them. But look at all the wasted capacity:

1.7Gbps @ 100% 24/7 = roughly 500 000 GB per month = 5GB per person, not 3GB.

This whole cap model is a mess.

 

[TeleGuru]

Don't forget that SEACOM currently has around 80Gbps lit capacity. The 1.28Tbps capacity would require many more interface cards across the network at each landing station, at a large additional cost. So technically, you can't use the 1.28Tbps in your cost calculations.

If you use 80Gbps = 8 x 10 Gbps @ $99 Million, you get $792 million which is roughly the cost of the cable plus a margin and over 20 years, remember.

Hence my reason for saying the calculation was done using "Seacom Maths", as they are constatly leaving out details, which have a significant impact on the true answer.

That being said, the cost of an upgrade is $2 000 000/STM-64 (assuming it lands at every country along the way), that is an additional $240m.

Re-doing the calculation, then gives your a profit margin of:
($99m - $6.56m) / $6.56m * 100 = 1409.1% Profit.

And if they sell all 80Gbps now, they even by your numbers, they would have paid for the entire cable, and the additional 120 x STM-64's would still be "free"

 

[Gaz{M}]

Wow, undersea cables are a real money spinner! No wonder 3 more are on the way...