Greencom Secunda Fibre Project

Wonder when Alex is going to realise that 123net, isn't really the company/ISP that he thinks they are. Much like PostManPot and mWeb.
 
3) You guys realise you're funding them with your contracts, right? So you'd better be damn sure you agree with their financial models. Which from a cursory glance make me go nope, nope ,nope...
Don't understand this part.
 
Wonder when Alex is going to realise that 123net, isn't really the company/ISP that he thinks they are. Much like PostManPot and mWeb.
It's a case of desperate times call for desperate measures. I hope this anonymous investor of theirs realise they are burning goodwill before it's too late. None of the other providers or ISPs offer anything worthwhile.
 
Correct. But people have signed contracts, which are being used to "secure" funding, for a model that involves non-recovery of capex, and free uptake on the bottom-end of the market. Any banker would request one to please go back to finish grade 4 mathematics before approving this. Any "funder" the same...
Could you explain a little more? How are they not recovering capex?

You either sign up with an installation fee, or a monthly fee, and this is supposed to help recover capex. Whether it does is another question and one you cannot answer definitively without access to their input costs.

As to the 'free' end of the market, presumably their prior experience of this kind of rollout points towards customers very quickly moving from the basic offering to more expensive options.
 
Could you explain a little more? How are they not recovering capex?

You either sign up with an installation fee, or a monthly fee, and this is supposed to help recover capex. Whether it does is another question and one you cannot answer definitively without access to their input costs.

As to the 'free' end of the market, presumably their prior experience of this kind of rollout points towards customers very quickly moving from the basic offering to more expensive options.

I am intimately familiar with the input costs of operating a fibre network, I can assure you of that. I have also written a dozen financial models for FTTH projects, and have been involved in capital funding for most of my working life, and over the last 3 years specifically on FTTH and FTTB. I have also signed off on agreements with most infrastructure providers in South Africa and know their product offerings very well.

Let me illustrate this like I did in a previous thread with basic numbers in a perfect world. Assume the following:

100 free-standing houses and all sign contracts
All houses are next to each other (perfect world) and are on average 20m apart (excluding verge to modem connection)
Dedicated fibre into each house
Backhaul to DC of 120km (Secunda has no x-connect carrier grade DC so you have to backhaul that traffic into your closest network POP or directly into your DC - i.e. you buy carriage on a DFA link for that distance or find a suitable Telkom product for this national backhauling)

Trenching costs per house are R10k
Material and labour are R2,5k per house
CPE is R1k per house
Backhaul is R250k per month on that distance and R15k once-off (probably more on quote though)
Core network capex including config is R300k
DC costs are excluded here as well, assuming you have your own network operating already, which in the case of 123net, they don't.

Total capital expenditure per household is R16,650 excl which you'd have to prove to investors you can recover from profit over the duration of the contract. Immediately, any free model will be excluded as potential recovery, so don't even consider it. At R3,500 per install (the proposed model) you're left with a funding shortfall of R13,150 per household. This excludes internet bandwidth - we haven't yet included that cost. So if you amortise that over the duration of the contract you're left with a baseline cost of R550 per month per house to recover. You must consider the present discounted value of money when performing these calculations at the very least, meaning the recovery is closer to R700 excluding interest rate exposure risk and risk due to time in the market, as well as excluding any profit yet, and the opportunity cost loss of your proposed venture. You still haven't added internet bandwidth yet either.

Now you're left with a problem as a new network. You either need a massive x-connect and peering network, or you're going to have to purchase a lot of local bandwidth. Either way you have to buy a lot of international capacity now depending on the speed variation uptake, or you can simply sell capped (but that's not what has been proposed here). So you adopt more risk and go for uncapped no FUP no shaping 100Mbps. You're marketing what amounts to a dedicated service but you add some contention there. You now require 1Gbps international breakout and 1Gbps local traffic. You arrange a deal at R200/Mbps for this on a 1:1 blended deal. Nice.

Now you have R700 p/m to recover for your initial capital investment per house. You have R2000 p/m to recover on bandwidth breakout on 10:1 contention (with a peak capacity load problem because you only sold this to home users, so bandwidth is only used at certain times of day - this also creates a price/GB pricing problem and imbalance). Now you have to find a retail model where your ARPU equals minimum R2,700 per month, or you are trading insolvent. And that's a fairly perfect world scenario. And requires massive pockets at first as well.

There are ways to get those costs down, but those are the indicative costs a brand new network would be forced to recover using the bold statements 123net for example have made. It goes even further seeing as the claims are that they'd install it all themselves and hire the staff to do so. You could never operate with the economies of scale that the traditional fibre installation companies do. It's taken even further when you claim you're purchasing the machinery to do this yourself too. You now require the capital funding of an infrastructure installer, a tier 1 network operator, and a tier 2 network operator.

So the scepticism is not borne from anything other than actually knowing the costs involved based on the statements made. Nothing more...
 
I am intimately familiar with the input costs of operating a fibre network, I can assure you of that. I have also written a dozen financial models for FTTH projects, and have been involved in capital funding for most of my working life, and over the last 3 years specifically on FTTH and FTTB. I have also signed off on agreements with most infrastructure providers in South Africa and know their product offerings very well.

Let me illustrate this like I did in a previous thread with basic numbers in a perfect world. Assume the following:

100 free-standing houses and all sign contracts
All houses are next to each other (perfect world) and are on average 20m apart (excluding verge to modem connection)
Dedicated fibre into each house
Backhaul to DC of 120km (Secunda has no x-connect carrier grade DC so you have to backhaul that traffic into your closest network POP or directly into your DC - i.e. you buy carriage on a DFA link for that distance or find a suitable Telkom product for this national backhauling)

Trenching costs per house are R10k
Material and labour are R2,5k per house
CPE is R1k per house
Backhaul is R250k per month on that distance and R15k once-off (probably more on quote though)
Core network capex including config is R300k
DC costs are excluded here as well, assuming you have your own network operating already, which in the case of 123net, they don't.

Total capital expenditure per household is R16,650 excl which you'd have to prove to investors you can recover from profit over the duration of the contract. Immediately, any free model will be excluded as potential recovery, so don't even consider it. At R3,500 per install (the proposed model) you're left with a funding shortfall of R13,150 per household. This excludes internet bandwidth - we haven't yet included that cost. So if you amortise that over the duration of the contract you're left with a baseline cost of R550 per month per house to recover. You must consider the present discounted value of money when performing these calculations at the very least, meaning the recovery is closer to R700 excluding interest rate exposure risk and risk due to time in the market, as well as excluding any profit yet, and the opportunity cost loss of your proposed venture. You still haven't added internet bandwidth yet either.

Now you're left with a problem as a new network. You either need a massive x-connect and peering network, or you're going to have to purchase a lot of local bandwidth. Either way you have to buy a lot of international capacity now depending on the speed variation uptake, or you can simply sell capped (but that's not what has been proposed here). So you adopt more risk and go for uncapped no FUP no shaping 100Mbps. You're marketing what amounts to a dedicated service but you add some contention there. You now require 1Gbps international breakout and 1Gbps local traffic. You arrange a deal at R200/Mbps for this on a 1:1 blended deal. Nice.

Now you have R700 p/m to recover for your initial capital investment per house. You have R2000 p/m to recover on bandwidth breakout on 10:1 contention (with a peak capacity load problem because you only sold this to home users, so bandwidth is only used at certain times of day - this also creates a price/GB pricing problem and imbalance). Now you have to find a retail model where your ARPU equals minimum R2,700 per month, or you are trading insolvent. And that's a fairly perfect world scenario. And requires massive pockets at first as well.

There are ways to get those costs down, but those are the indicative costs a brand new network would be forced to recover using the bold statements 123net for example have made. It goes even further seeing as the claims are that they'd install it all themselves and hire the staff to do so. You could never operate with the economies of scale that the traditional fibre installation companies do. It's taken even further when you claim you're purchasing the machinery to do this yourself too. You now require the capital funding of an infrastructure installer, a tier 1 network operator, and a tier 2 network operator.

So the scepticism is not borne from anything other than actually knowing the costs involved based on the statements made. Nothing more...

Well put.

For many Telkom FTTH is going to make the most long term sense. They have the capital, experience and logistics to handle such an expansion.

I personally do not see why Secunda does not put out an RFP like we did. Get the different providers to compete and choose the one with the best prices, track record and means to do the job.
 
Notice how DJ still hasn't said anything about how much actual trenching costs and is still applying an average as well as conflating the cost of high speed bandwidth with that of a low speed free service.
 
Notice how DJ still hasn't said anything about how much actual trenching costs and is still applying an average as well as conflating the cost of high speed bandwidth with that of a low speed free service.

I am not conflating anything, and I have provided you with indicative costing. If you'd like specifics, you can submit a request from our business channels.

I'm not trying to argue with you here at all. I'm trying to illustrate some facts. If you disagree and would like to correct me, do so...
 
I am not conflating anything, and I have provided you with indicative costing. If you'd like specifics, you can submit a request from our business channels.

I'm not trying to argue with you here at all. I'm trying to illustrate some facts. If you disagree and would like to correct me, do so...
You have. If you want to know the cost of a service you take all variables of that service into account and not others like for 100Mbps. I asked the cost of raw trenching which you don't want to provide. So by your estimation it would cost R10k to dig a hole from one house to the next. Seriously, I'm looking for the job that would pay me that kind of money.
 
Simply asserting as such means nothing. If you believe trenching costs less, go get a quote. I have provided you with the cost of trenching fibre. Around R500/m depending on the provider. I've never been coy about that, and I've posted it before for you...
 
So you're not going to give any actual figures and just continue with overall estimates. This is where things don't add up and why I'm becoming more distrusting of the industry with their claimed figures. You know when driving in a car distance goes by rather quickly as the speedometer ticks over every few seconds. Now when going you see this trench. One week there was nothing and the next week there's this long trench 100's of meters long. Let's make it 100m just for giving the benefit of the doubt.

Now let's say you have a team of 5 all working on it and the salary is around R10k per month for each. That's around R50k or R12.5k per week for 100m. Even being conservative here the figures don't come close to your estimation of R500/m. Manual labour is simply not that expensive. So what makes fibre again so special and different from the rest of the labour market. Are you using diamond shovels? Machinery? No if you were using machinery the cost would be a lot less. I'm still waiting for an explanation from someone that has some basis in reality.
 
So you're not going to give any actual figures and just continue with overall estimates. This is where things don't add up and why I'm becoming more distrusting of the industry with their claimed figures. You know when driving in a car distance goes by rather quickly as the speedometer ticks over every few seconds. Now when going you see this trench. One week there was nothing and the next week there's this long trench 100's of meters long. Let's make it 100m just for giving the benefit of the doubt.

Now let's say you have a team of 5 all working on it and the salary is around R10k per month for each. That's around R50k or R12.5k per week for 100m. Even being conservative here the figures don't come close to your estimation of R500/m. Manual labour is simply not that expensive. So what makes fibre again so special and different from the rest of the labour market. Are you using diamond shovels? Machinery? No if you were using machinery the cost would be a lot less. I'm still waiting for an explanation from someone that has some basis in reality.

We are an ISP. We are not a civil works company, and we do not set the pricing for the civil works. Trenching hard-ground is what dictates pricing, not the cabling, although fibre has certain bend parameters and other criteria that determines the type of reticulation required. Installing into soft-ground is cheaper, but poses its own challenges. The cost to trench into hard-ground like road surfaces, pavement, concrete and similar, is roughly R500/m (in fact one provider you all know charges around R750/m for this). You are asking me for a specific number here. I can tell you that no provider operates like this - they spec an RFP and provide a quote on the specific site/route/build path. Not sure why this angers you so much.

I also suspect you underestimate what trenching involves:

Route survey, design, and approval
Wayleave management and right of way application
Trench excavation
Ducting
Backfill of trench
Fibre blown through
Testing

Each step requires a set of specialists, with specific machinery, tools, and expertise. The above also excludes possible foundation work required. Also bear in mind the moment you cross a road you require very specialised machinery and skills. Once you hit the splicing point you also require the skills of an expensive specialist in his/her field.

You seem angry with me, and really all I'm doing is being open about the costs with you and other members here. You are more than welcome to contact a company like PHB Engineers if you would like an actual quote for such work to confirm for yourself. Or you can continue to think that I'm somehow part of some large conspiracy to drive up costs, which would be absurd as the lower our input costs as ISPs the better. There are solutions to these high costs coming soon. We are waiting on regulators to make the appropriate moves here...
 
Last edited:
^
You left out reinstatment costs, e.g. roads, driveways, pavements, flower beds, etc.

Also the cost to mitigate the risk of hitting water pipes, gas lines, electrical cables, etc.

Also take into account that you probably need to bring in your own backfill depending on the soil properties in the area, eg. clay or too sandy or hard rock.

And and and...
 
I'm lumping those in survey and backfill, as line iteming them would take forever...
 
I am not angry with you. I am simply asking questions which by your own admission you can't really provide the answers to. Quoting quoted figures by no means gives us a cost as we all know suppliers have massive profit margins. It's the raw cost of trenching I want to know and by an estimation of R500/m it would take 2 hours per meter which doesn't add up with the cost of labour. As I said if it was R500/m I would be there tomorrow but funny enough nobody would offer me that much.

Sure R500/m for roads and pavements is a more realistic figure. You're looking at it from the point of view that you're incurring these and other costs again with every connection. It doesn't work that way.
 
Can you explain why the cost is not incurred for every house? You're also only looking at the labour cost. As per previous posts, it's not only labour you pay for...
 
Top
Sign up to the MyBroadband newsletter
X