In reply to a post by RobertoS:

� I wonder if Openreach will ever achieve a positive ROI on G.Fast�

The development cost of G.fast was already sunk - originally it was conceived as something you would put in the basements of buildings and in DPs. That turned out to be impractical.

So they had the technology, and all they had to do is slap it on the side of a PCP, when there's already a powered VDSL cab nearby. Could just be a day's work in some cases.

OTOH, the long term support costs of G.fast, with increased engineer visits when people don't get the expected speeds, could eat into any profit.

I expect BTG aren't really interested in an absolute ROI, but it's more about propping up the share price based on having 2.4m more "ultrafast" lines available and thus keeping ahead of the altnets.

It's interesting that Tunbridge Wells, which already has quite a lot of G.fast coverage, has recently been announced in the Fibre First programme. I wonder if they'll be replacing G.fast, or using fibre to fill in other areas only.

In reply to a post by jchamier:

I feel OR has missed an opportunity for G.Fast in multi-occupancy-dwellings (MDU) e.g. small blocks of flats. Small blocks often have entry phone and fire detection systems (a lot have been installed since the Lanarkal House fire). This means communal power is available.

The lengths of copper wiring through the property to 6 / 12 / 18 flats from a interior DP would make G.Fast much more cost effective than drilling FTTP through the walls. Problem is most flats are wired to external DP's in ground chambers, with no power.

I hadn�t thought of that idea. It could make a lot of sense, particularly if for some such properties they could redeploy G.Fast hardware from new FTTP areas.

Additional G.Fast kit would probably be needed as well of course.

The outside chamber could well have room for miniaturised G.Fast hardware, with power back fed? By �miniaturised� I envisage pre-connection inside it�s box to external connections for the existing copper. Removing the existing DP.

Hmmm. Thinking more, maybe it wouldn�t be as good long term due to the maintenance requirements compared to passive fibre all the way to an outside wall mounted fibre DP and drilling into each flat from outside.

In reply to a post by candlerb:

The development cost of G.fast was already sunk - originally it was conceived as something you would put in the basements of buildings and in DPs. That turned out to be impractical.

I know that was the original idea, but unsure what you mean my �sunk�.

It has certainly never been recovered from direct income, so I regard it as still part of the equation for assessing whether it was a cost-effective spend. In other words, true ROI.

It could well have disappeared from overall accounts and written off within an R & D budget, but that is just for public accounting jiggers-pokery. Internally it is money that could well have been spent on something that provided real net profit. A positive ROI.

That is because it is a figure not pushed to labs as a direct availability figure

https://www.thinkbroadband.com/news/8565-pace-of-ope...

Compares different FTTP areas and there is little difference in over 80 Mbps take-up in Fibre First areas compared to G.fast areas

https://www.thinkbroadband.com/news/8579-bt-group-fi...
Has some maths from the revenue FTTP generates, higher revenue than FTTC but then you have a higher per premises roll-out cost, but this may balance in longer term with lower maintenance costs.

Retention against VM is probably not a big worry, I would suggest that the provider with the most to lose is Virgin Media in current climate from the competing FTTP roll-outs.

G.fast is FibreFirst areas - it will not vanish but you will find the footprint may change slightly with some premises losing the G.fast option even if in range of the pod, i.e. FTTP be the sole ultrafast option over the Openreach network. This is based on what I've seen in other urban areas with a high G.fast coverage when FibreFirst FTTP has appeared. Original position was that no-one would have both G.fast and FTTP options but plenty of addresses actually do (and I mean addresses not postcodes).

What G.fast will allow for is once the need for higher speeds becomes more common that those in old flats where landlords are blocking FTTP that they will get an option for better speeds.

In reply to a post by RobertoS:

I know that was the original idea, but unsure what you mean my �sunk�.

It has certainly never been recovered from direct income, so I regard it as still part of the equation for assessing whether it was a cost-effective spend. In other words, true ROI.

It could well have disappeared from overall accounts and written off within an R & D budget, but that is just for public accounting jiggers-pokery. Internally it is money that could well have been spent on something that provided real net profit. A positive ROI.

Well, we're talking about G.fast deployment to PCP pods.

A "sunk" cost is one which is not recoverable. If G.fast wasn't deployed to pods, then that money is gone. But I don't think it's fair to book that entire cost against the newer, tactical G.fast deployment.

The choice was either (1) bury G.fast entirely, or (2) do the PCP deployment. On that basis, the incremental cost of the PCP deployment is low.

We shall have to agree to differ about the original choice. Openreach were quite clear in all pre-announcements that the pods would be widely distributed well away from the main cabinets.

It was only when the penny dropped that wasn't cost-effective that they came up with the current deployment, so as not to write off the full project.

As the project has not been closed down and written off, the costs so far all count towards the overall project's ROI. No argument. Otherwise they would have to appear somewhere in the accounts as write-downs.

If you remove those development costs from the ROI of PCP-attached pods, you completely falsify the accounts. My bold:-

A sunk cost is a cost that was incurred in the past and cannot be undone. Since most transactions cannot be undone, most amounts spent in the past can be described as sunk. In other words, a past or sunk cost will be there regardless of what you decide to do today or in the future.

The cost is still there to be placed against the project.

ROI tries to directly measure the amount of return on a particular investment, relative to the investment's cost. To calculate ROI, the benefit (or return) of an investment is divided by the cost of the investment. The result is expressed as a percentage or a ratio.

I accept there is a difference of treatment between the fundamental research and the ongoing development costs, but both remain relevant to ROI in respect of management decision-making.

Openreach will never make a direct profit from having researched, developed and deployed the product. BT Group may, over many years.

In reply to a post by RobertoS:

We shall have to agree to differ about the original choice. Openreach were quite clear in all pre-announcements that the pods would be widely distributed well away from the main cabinets.

Absolutely. The idea was that it would go into DPs, leaving the last hundred meters or so as copper. In that scenario, using the high frequencies that don't overlap with ADSL and VDSL was quite reasonable.

That idea has failed. Some unknown but possibly large amount of money was burnt on developing, testing and trialling the idea.

Imagine that company went bust. Its remaining assets - the design of electronics for G.fast - were bought up in a "fire sale" by a different company. It takes those assets and deploys them in a way unrelated to what they were originally developed for.

The second company doesn't have to account for all the money spent by the first one that went bust. It paid what is accepted by both sides as a fair and reasonable price for what was left after the first company failed, and that's what shows on its books.

That's how I think of the tactical G.fast deployment. It's a separate project that doesn't have to account for the full costs of the blind alley that the first one went down. It would have to account for the fair value of the assets it took on, but not for example all the failed deployment trials.

If it weren't for LLU, then VDSL 35b could have been deployed instead, and the residual value of G.fast would have been zero.

Amounts spent internally and buying up assets and rights from companies in liquidation are two very different situations. You cannot equate them in that way.

If Openreach had been in a completely separate company from BT Group as opposed to wholly owned, then BT had bought its assets at a low cost either direct or into a newly-formed "BTOR Ltd" for instance your idea would be valid.

The same happens frequently in fact with (commonly retail) companies going into liquidation, sometime voluntary liquidation, and the directors setting up a new company the next day and buying the whole shebang cheaply from the liquidator.

Eradicating its debts to suppliers at the same time.

You cannot do that in the case of Openreach as a healthy company within BT Group. You may wish to think of it that way, but its internal accountants and auditors cannot.

Having said that, as I pointed out earlier, the waste of money will just be "disappeared" within the overall research and development budget. Some projects are bound to fail completely, some to be open to redeployment and the recouping of some of the initial loss. As in the case of G.Fast. Allowances are made in such budgets, and for reasonable variations.

For all such projects there will be some form of ROI calculation and also a risk assessment before authorising it to go ahead. If at some stage it looks like those assessments were badly wrong then the project is reconsidered. Think for instance of HS2 at the moment, where ongoing costs are continually reviewed and doubts about continuing it raised.

Changing the route of HS2; or cutting out plans to extend it after Phase 1; or terminating it just before London (which has been discussed); or lowering the expected maximum speed in order to lower costs as I think happened a while ago, do not and cannot result in the costs to the date of any such revision being zeroised.

In essence the two situations are the same.

In reply to a post by RobertoS:

Having said that, as I pointed out earlier, the waste of money will just be "disappeared" within the overall research and development budget. Some projects are bound to fail completely, some to be open to redeployment and the recouping of some of the initial loss

So you'd be happy that if G.Fast had been canned completely, the entire budget could be written off against failed R&D.

But you say that if *any* limited deployment of G.Fast takes place, its business case has to carry the *entire* cost of G.Fast R&D, including the blind alleys? That doesn't make sense to me. Surely a partial write-off is perfectly reasonable here.