A sample SEACOM cost calculation

[Gaz{M}]

Hi all,

I decided to cut through the sea of confusion about SEACOM by doing an example calculation based on the financial terms of buying a 155Mbps STM-1 link from SEACOM and see what the actual costs are.

Firstly, the terms of the SEACOM IRU (basically, you 'own' the bandwidth for 20 years):

5% downpayment as a deposit
35% of the cost paid in Year 1
30% in Year 2
30% in Year 3
Thereafter, you have paid for your IRU and are only liable for the annual maintenance cost of 4% of the original price.

An STM-1 155Mbps IRU from SEACOM costs $4 350 000.

So here are the figures for Years 1, 2 and 3.

Year 1:

5% deposit = $217 500
35% of IRU = $1 522 500
4% maintenance = $174 000
TOTAL = $1 914 000 @ R8.30/$ = R15 886 200

150Mbps usable bandwidth @ R15 886 200/year = R8825.67/ Mbps / month

Year 2:

30% of IRU = $1 305 000
4% maintenance = $174 000
TOTAL = $1 479 000 @ R8.30/$ = R12 275 700

150Mbps usable bandwidth @ R12 275 700/year = R6819.83/ Mbps / month

Year 3:

30% of IRU = $1 305 000
4% maintenance = $174 000
TOTAL = $1 479 000 @ R8.30/$ = R12 275 700

150Mbps usable bandwidth @ R12 275 700/year = R6819.83/ Mbps / month

Year n+3:

4% maintenance = $174 000 @ R8.30/$ = R1 444 200 <-- (EDIT: This figure reduces to 3% for Years 5-7 and to 2.5% thereafter)

150Mbps usable bandwidth @ R1 444 200/year = R802.33/ Mbps / month


Year 1: R8825.67/ Mbps / month = about R30 /GIG
Year 2: R6819.83/ Mbps / month = about R22 /GIG
Year 4: R802.33 / Mbps / month = about R3 / GIG


This does NOT look very rosy at all!

 

[Gaz{M}]

In addition to this, we've all seen these prices quoted by SEACOM in their marketing:

STM-1 ......R 673/Mbps/Month
STM-4 ......R 575/Mbps/Month
STM-16 ....R 435/Mbps/Month
STM-64 ....R 267/Mbps/Month

Please note, this LOOKs impressive, but is the total effective cost per Mbps/month over the ENTIRE 20 year period you own the cable.

What they do NOT say is that for the first 3 years the cost is WAY more as I have shown above. Only after year 4 do you see significant cost savings on international bandwidth. This proves that you should not believe everything you read at face value.

What does this mean? Prices will likely not come down much in Year 1, a bit more in year 2, and only drastically in 3 years time (2012)!

 

[Roman4604]

This does NOT look very rosy at all!

You left out a few expenses;

1. Pay for getting the 150 Mbps from where Seacom ends in SA to your premises

2. Pay an ISP(s) to hook your end of Seacom to the Internet in London

3. Pay finance charges, as few will have this amount of cash lying around.

4. Pay some bright spark(s) to look after it

5. Buy some (big ass) routers etc.

 

[Nerfherder]

The only way this could be correct is if you compared it to how much SAt3 charges for bandwidth.

Also.... A good business model will let you run at cost until the cheaper bandwidth kicks in.
For example you say that is is going to cost R30 per gig at first but then R3 per gig in 4 years time. So look at it like this... you sell bandwith at a loss now and in 4 years time you will be selling 10x more and making double the profit.

Its not as gloomy as you paint it..... I know that you can get R30/gig wholesale at the moment, so its not possible that they are going to come in at the same price.

 

[Bobb]

Is that cost a 1-1 contention ratio? What is the average ISP contention ratio?

And what does a 155Mbps STM-1 link cost on SAT3? This would be a more interesting comparison.

 

[Gaz{M}]

Well for SAT3, I've seen figures of between R8000 and R11 000 per Mbps per month thrown around various forums.

ie: SEACOM is not really much cheaper in the first 3 years.

 

[bawrak]

Sorry for my ignorance, but how did you arrive at:

Year 1: R8825.67/ Mbps / month = about R30 /GIG
Year 2: R6819.83/ Mbps / month = about R22 /GIG
Year 4: R802.33 / Mbps / month = about R3 / GIG

Then again I suck at maths

 

[sjm]

R8825.67/ Mbps / month = about R30 /GIG

How do you work this out? I'm getting R0.03/GIG roughly

 

[Gaz{M}]

1 Mbps = 110KB/sec usable download speed x 3600 sec x 24 hrs x 30 days = 300 GB/month

8825.67/300 = about R30/GB

 

[DeesCent]

Well, i guess that is business. The 1st 3 years the costs are high, but you will make that back in time. If a provider averages their cost/price from year 1, they will eventually make a profit. Is that not the way it works? you start running at a loss then turn profitable in a few years. If they keep prices too high at the outset they will stand to lose more as less people will use their services as they wont be running at or close to capacity.

 

[sjm]

1 Mbps = 110KB/sec usable download speed x 3600 sec x 24 hrs x 30 days = 300 GB/month

8825.67/300 = about R30/GB

My bad!

 

[Nerfherder]

1 Mbps = 110KB/sec usable download speed x 3600 sec x 24 hrs x 30 days = 300 GB/month

8825.67/300 = about R30/GB

If that is the case then you are wrong... not in your calculation but in your assumption.

 

[Gaz{M}]

If that is the case then you are wrong... not in your calculation but in your assumption.

Care to elaborate on this assumption?

 

[NaushadE]

Sorry for my ignorance, but how did you arrive at:



Then again I suck at maths

Can you please explain it too.. i have always wondered about the cost structure of isp's this is a perfect opportunity to find out.

 

[Nerfherder]

Care to elaborate on this assumption?

Well bandwidth is not costed like that. If you were able to get a gig at that speed then you might be able to. All it tells us is that we can't all have uncapped 1 meg downloading all the time.

Even so, the fact that you can get 4 meg uncapped for 3k means that the assumptions must be wrong. Work backwards from that and you will see how the costing will be different.
We know that SEACOM will be cheaper bandwidth, if your calculations tell you that an uncapped 1 meg line will be 8k at cost then you know that it is wrong.

Why would any of the ISP's pay for more expensive bandwidth if they could get SIAX cheaper ??
The Kenyan costing will also show you that you have made the wrong assumptions.

 

[Pox]

I remember that telkom once said their contention ratios are 20:1...so that would make the cost about R1.5 per GB

 

[Nerfherder]

I remember that telkom once said their contention ratios are 20:1...so that would make the cost about R1.5 per GB

Yes... I think that is more realistic. Its the cost price but more realistic.

 

[NaushadE]

can some1 please put up a brief explanation of all the terms used, ie; Cotention Ratios, LLU, etc. some of us noobz are interested in this stuff....

from what i can see, if a business can absorb the losses for 3-4 years, theres quite alot of money to be made even at the low cost level.

 

[Gaz{M}]

Fair enough Nerfherder. You make valid points. Without insider knowledge of the ISP industry, I can only do basic, back of the envelope calculations.

If currently 4Mbps uncapped services are sold for around R3000 a month, based on a thumb suck oversell of 50:1, then that requires 4096/50 = 82kbps per subscriber.

Based on a 100% markup for costs/profit, that's R1500/82kbps or R18800 per Mbps/month cost which sounds about right.

So, let's say we go with the new SEACOM pricing of R8900/Mbps/month, meaning we should see uncapped 4 Meg services at R1400 per month in the next year or so or afterhours uncapped services for 50% of what they are now (R900), which would be around R450 per month.

This 50% odd reduction in prices is what has been recently quoted in various SEACOM articles.

 

[Nerfherder]

This 50% odd reduction in prices is what has been recently quoted in various SEACOM articles.

Well done In that case your maths is spot on !

I can live with those prices for now... next year I expect another 50%