& cplSiteName &

Figuring FiOS

Phil Harvey
9/27/2006
50%
50%

Will it really cost Verizon Communications Inc. (NYSE: VZ) more than $9,000 to bring fiber to your home?

Yes, it could. And bear with me, please: I'm about to attempt some FiOS math.

The company today said it plans to pass 18 million premises with its fiber network by the end of 2010. It also says it expects to invest $18 billion in net capital from 2004 through 2010 in deploying that fiber-fed network. (See Verizon to Pump $18B Into FiOS by 2010.)

I’m on my second bourbon and coffee, but it seems like Verizon has just copped to spending about $1,000 a home, on average, to pass 18 million homes, over a six-year stretch. Earlier, Verizon said it cost around $900 to pass a home, so let's go with that.

Verizon also says it will cost only about $650 to connect a "passed" home to its network by 2010.

So what does it cost to hook up a neighborhood? These aren't absolute figures, mind you, but let's assume that Verizon passes each home in a 400-home neighborhood, then nabs 10 percent of the homes (40 homes) as customers.

Take $900 and multiply it by 400 homes. That's $360,000.

Now let's hook up those 40 homes. That's 40 multiplied by $650. That's $26,000 added back to the cost to pass the homes, which was $360,000.

So now we have a figure of $386,000 spent in just one neighborhood. But what has Verizon spent per customer? Take $386,000 and divide it by the 40 homes and you get $9,650.

Millennium Marketing principal Kermit Ross worked out some very similar figures for me on a notebook after a session at Optical Expo 2006 last week. I recall his numbers/assumptions were also in the ballpark of $9,000 per subscriber.

Okay, now I'm on my third bourbon and coffee. But I think the larger point here is that even with all the cost savings Verizon has managed to achieve, this stuff is still really expensive.

"Obviously, in the early stages of a network, the cost to connect each home is astronomical and there's really nothing you can do about that," says Graham Finnie, an analyst at Heavy Reading.

So slice it anyway you like -- fiber to the home is damned expensive. And the payback takes years, maybe decades. But without a next-generation access network, carriers simply won't have a business.

— Phil Harvey, News Editor, Light Reading

(22)  | 
Comment  | 
Print  | 
Threaded  |  Newest First  |  Oldest First        ADD A COMMENT
mfratto
50%
50%
mfratto,
User Rank: Light Beer
12/5/2012 | 3:38:53 AM
re: Figuring FiOS
Are you sure that your wireless if good to go? When I had FiOS installed, I was getting less than half of the performance I paid for 20Mbps down/5 Mbps up. The installer, a nice guy, couldn't get it fixed. One of the local reps wanted me to "tune" my TCP stack (a process that hasn't been necessary since Windows 98).

I finally fired up a wireless analyzer and determine that there were a few other wireless devices in the 'hood competing for the same channel. I reconfigured my Actiontec for only 802.11G and fixed the channel to a non-overlapping one. There is a decent chart at http://www.unixwiz.net/techtip....

Those two changes doubled my performance to an acceptable limit. To be honest, if I take another wireless hit, I am giving up and running Cat 5 everywhere.

I can't imagine the problem FiOS customers are going to have in higher density neighborhoods where there will be more compettion for spectrum. Verizon need to spend some money training it's field staff in 802.11 wireless network and stop turing to DSL Reports forums for tech advice.
optical
50%
50%
optical,
User Rank: Light Beer
12/5/2012 | 3:39:10 AM
re: Figuring FiOS
The heck with the dollars and cents, as a consumer i just switched over to FiOS and the service stinks. My Verizin DSL was more dependable and faster. Verizon's attempt at fixing is talking to someone who can barely speak english wanting to go though a checklist ... I've already checked the connections dude and the wireless LAN in my home just fine with DSL, so give me back my dependable DSL !
davrich
50%
50%
davrich,
User Rank: Light Beer
12/5/2012 | 3:39:12 AM
re: Figuring FiOS
Regarding the STB CAPex, found the answer in the transcript at: http://investor.verizon.com/ne... . I quote:

Unidentified Audience Member
Can you give us a general sense of the cost per home for just the set-top box piece of the equation, what we should add to that $2,500?
Doreen Toben - Verizon - EVP, CFO
I think I would just tell you industry normal, what you think a normal set-top box is.
stephencooke
50%
50%
stephencooke,
User Rank: Light Beer
12/5/2012 | 3:39:12 AM
re: Figuring FiOS
davrich,

I listened to the webcast, though the part with the Fios demo was obviously useless...

Doreen explained that the STB was totally separate from the cost-to-connect numbers because it is a separate line item (ie: it is leased by the customer on a monthly basis). See slide 37 where it lists equipment pricing.

A couple of extra notes on the slide package:

- Slide 40 basically says that Phil's $9k is actually $2535 on average through 2010. This assumes a 39% take rate, 18M houses passed (which apparently takes $4.4B out of their copper-based capex in that interval), $500M in support cost reductions, and that they meet their cost reduction targets for cost-to-pass & cost-to-connect.

- It seems that they are assuming that anyone who connects, whether they sign up for video or not, are paying for the video head-ends, etc.

- The 25k CO over 5 years case seems a bit inconsistant though... They assume the cost to connect is $733 yet they also say that 70% of the customers come on in the first 3 years. The point is that the cost to connect is $1200 in 2004, $1163 in 2005, hopefully $880 by YE 2006, and maybe $650 by 2010. It appears that the cost numbers are backend weighted but the customer numbers are frontend weighted...?

If they can meet the 39% "success" rate by 2010 (assuming 18M houses passed. If they don't pass this many the average costs/house change) they may do OK. Up until those conditions occur I'd say their stock will be hanging by a thread.

Steve.
davrich
50%
50%
davrich,
User Rank: Light Beer
12/5/2012 | 3:39:14 AM
re: Figuring FiOS
From rjmcmahon
>Costs didn't include set top boxes. Customers
>will demand them to be included. TIVO showed us
>there is little consumer demand for customer
>owned STB equipment, even if it's novel.

But then VZ's FAQ page for FIOS says the STB is leased by the customer. See http://www22.verizon.com/conte... . In which case, why doesn't VZ include the STB in their costs? Are they maybe doing something creative, like a lease arrangement through another party so that it's off of VZ's books?
DCITDave
50%
50%
DCITDave,
User Rank: Light Beer
12/5/2012 | 3:39:15 AM
re: Figuring FiOS
re: "Real humans don't care if their email MySpace access and porn are delivered by magical elves riding flying squirrels, they just care about it being fast, affordable and relatively reliable."

Everyone please rate that last message highly so I can use the above quote in next week's newsletter. Seems like that should be crocheted on a pillow or something.
fgoldstein
50%
50%
fgoldstein,
User Rank: Light Sabre
12/5/2012 | 3:39:15 AM
re: Figuring FiOS
These numbers seem both realistic and surprisingly pessimistic. Why would a company like Verizon pay $9000 per home served? It's not rational. The OpEx savings, if real, just aren't worth more than a small fraction of that. But then there may be more going on.

FiOS rollouts began shortly after Mike Powell granted "relief" for unbundling. To hell with the black letter law of 47CFR Sections 251 and 271; Powell decided that unbundling obligations are basically limited to facilities that were in place the day the Telecom Act passed. That's not what the law says but who cares? But Powell (and Martin) needed a quid pro quo. So in exchange for allowing them to cut off competitive LEC access to fiber-served homes, and competitive DSL access to lots more homes, VZ had to pull some FTTH. This is classic "industrial policy". "Fiber" is good, so whatever creates it is good. Forget why fiber was supposed to be good, but get it even if it's done badly.

Remember that back in 1993, the ILECs, especially the Bell Atlantic ones, promised to roll out FTTH over the following decade, on an open access basis, 45 Mbps, in exchange for a removal of rate of return regulation. So they got their relief, and have raised their typical returns above 30% (vs. around 12% under regulation). Yet they didn't pull the fiber. See http://www.teletruth.org/ for more details. So now they're repeating the trick, promising fiber ine exchange for killing their competition. They only have to do it long enough to be credible; they can slow down drastically next year, blaming economic conditions, investors, whatever. Martin will buy absolutely anything they say.

The key to this type of business is take rate, of course. They can theoretically migrate their POTS business to glass, but why bother? They could also gradually replace copper with glass, a sort of routine upgrade, as we sort of envisioned in the 1980s would happen in the 1990s. But DSL saved copper's butt. FiOS also fills in some holes where DSL doesn't reach, but there are cheaper ways to do that (e.g., Pronto, Lightspeed). How much incremental revenue will they get? In some areas they're overbuilding where RCN has already overbuilt cable. As a business, this doesn't make a lot of sense.
davrich
50%
50%
davrich,
User Rank: Light Beer
12/5/2012 | 3:39:17 AM
re: Figuring FiOS
kbg_lem posts
>Anybody knows what it really means to pass a home?

I think you're mostly on target. Chart 38 from VZ's deck http://investor.verizon.com/ne... seems to give the best detail.

It looks like they have the Network & Support (incl. video-head-end) and STB broken out separately from cost-to-pass and cost-to-connect. So I think CPE and install costs inside the home are not included in cost-to-connect.

But then their Gross CAPex calculation on chart 40 doesn't show STB. Which makes me wonder whether STB (and related costs) is indeed buried in the cost-to-connect numbers.

Any opinions on this?
50%
50%
"Ill" Duce,
User Rank: Light Beer
12/5/2012 | 3:39:18 AM
re: Figuring FiOS
I'm not sure about your neighborhood, but here in "the Reddest of Red States" the take rate is probably about 20%. I'm not sure about the logic behind people signing up for broadband without TV. It would cost me more and if the FIOS quality is better than Comcast (which it likely is) I would switch on principle. My ILEC is SBC, the ultimate evildoer, therefore it behoves me to consolidate. That's the point of bundling. If all things are equal and the price is lower, and there is no switching costs, I'll go for it. Your not taking into account actual human behavior. Real humans don't care if their email MySpace access and porn are delivered by magical elves riding flying squirrels, they just care about it being fast, affordable and relatively reliable. Although there seems to be some elasticity in that last point.

As for the provider's CPU being cycled out. That stuff is ungoldy cheap. Right now you're looking at $200 per unit on the high side. Within a year or two that could drop below a c-note.

Is the 9K estimate valid though? You're assuming that the provider installs an OLT and populates it right off. If each blade supports Nx32 homes, they can control the offering by rolling it out in certain neighborhoods and increasing take rates through special offers. The OLT, ONU and splittera are cheap. The labor is what really represents the bulk of the cost. I my neigborhood you'll have to trench. There are no aerial drops. For greenfield apps this doesn't present a problem. You need a ton of prgram management to coordinate installs and builds. Turning up OLT's isn't easy. How do you test it if there are no ONU's hung. I'm not sure where folks get the idea that OPEX is reduced, I'd say the cost is just shifted in product cycle.

Michael Harris
50%
50%
Michael Harris,
User Rank: Light Beer
12/5/2012 | 3:39:19 AM
re: Figuring FiOS
seven, great call on the union angle. The telcos have griped for years about their fundamental cost disadvantage versus non-unionized cable operators. Conventional logic says VZ needs video to compete, which may or may not actually be true. What they do need for sure is an operating cost basis that more closely resembles their competitors in the local loop. With FiOS, seems they're attacking this issue from both angles.
optodoofus
50%
50%
optodoofus,
User Rank: Light Beer
12/5/2012 | 3:39:19 AM
re: Figuring FiOS
I'm not sure how FiOS addresses the union issue. It is still installed, owned and maintained by the Verizon Telecom organization, which means it remains in the union shop. So, maybe the fiber network needs fewer union people around to maintain it, so that helps a little. Except that it doesn't. The union isn't stupid. They see where this is going, and there is going to be one heck of a donnybrook when the Verizon union contract comes up for renegotiation. I think the current contract is up in early 2008, which means the ugliness starts happening in the middle of 2007. If your FiOS isn't installed by then, don't count on it being installed until mid-2008 when the whole mess gets sorted out.

optodoofus
rjmcmahon
50%
50%
rjmcmahon,
User Rank: Light Beer
12/5/2012 | 3:39:20 AM
re: Figuring FiOS
re: You guys have no idea how inefficient the copper network is to run from an Opex standpoint. If you did you would laugh at yourself. You are not even in the right zone to where the problems are. Let me give you a hint - it is spelled u n i o n.

I would guess pensions obligations are also part of the cost problems. From a shareholder perspective it seems like VZ could spin out wireless (include with it the wireline backhaul), push all the debt and union obligations onto wireline access (Fred's loopcos), then bankrupt the loopcos and shift the pensions over to PBGC. Then start all over with new fiber OSPs. (Let muni's raise the bonds before rolling them up.) Isn't that about how capitalism combined wtih deregulation works?
paolo.franzoi
50%
50%
paolo.franzoi,
User Rank: Light Sabre
12/5/2012 | 3:39:21 AM
re: Figuring FiOS

One thing you are all assuming is not true.

If Verizon gets a significant take rate in your neighborhood, you will get your copper line cut and your phone line run over FiOS. None of that changes your -48V phone.

You guys have no idea how inefficient the copper network is to run from an Opex standpoint. If you did you would laugh at yourself. You are not even in the right zone to where the problems are. Let me give you a hint - it is spelled u n i o n.

seven
rjmcmahon
50%
50%
rjmcmahon,
User Rank: Light Beer
12/5/2012 | 3:39:22 AM
re: Figuring FiOS
Here's the issues I took away from a cursory review of the charts and halfway listening to the call on Yahoo.

o Homes passed and connected costs decrease over time but VZ wouldn't separate labor vs materials. It's possible that labor/productivity won't meet those projections and costs will stay flat or increase. Also materials such as fiber were oversupplied during the bubble and that should be correcting.

o Costs didn't include sales/marketting expenses. Anybody following Vonage knows that "me too" services, which FiOS TV is, are very expensive in this area.

o Costs didn't include set top boxes. Customers will demand them to be included. TIVO showed us there is little consumer demand for customer owned STB equipment, even if it's novel.

o As Stephen and others have mentioned, the opex savings assumptions of the fiber plant may be overly optimistic. Personally, I'll probably never give up my line one copper, used for voice, with -48V DC, which is connected to a wall phone. If many people are like me it suggests VZ may have two OSPs instead of one to maintain, hence the savings may not be there.

o Their bundling demand projections were inverted and skewed. They used FiOS TV subscribers as the basis for their "bundling demand" charts saying that TV subscribers also buy other services. Most people interested in FiOS will be buying it for internet access, not FiOS TV. The upsale "or bundle" numbers should start with internet access as the basis and see how many buy other services.

o Their churn rate was exceedingly low, barely accounted for people moving. If there is ever really facilities based video competition with the MSOs one would expect this number to increase significantly. (What's the churn rate for cellular voice? Probably more in line with that)

o VZ would only make 2010 projections and explicity said that wouldn't make any projections for 2007. This makes them less accountable. In general, they seemed to be reticent with respect to revealing the details of their financial models used to prove an ROI.

PS. One way I use to determine a low end cash stream expectation on an investment is to use an annuity estimate from somebody like fidelity. (These tend to be poor investment choices) Give fidelity $90K and they'll guarantee you $1226 per month for 7 years. Hence, VZ needs to *earn* more than $122.6 per month on the low side to justify a $9K investment.
kbg_lem
50%
50%
kbg_lem,
User Rank: Light Beer
12/5/2012 | 3:39:23 AM
re: Figuring FiOS
Anybody knows what it really means to pass a home?

My take is:
1-All the Fiber is installed in a neighborhood.
2-All the Fiber is tested and Connected to an OLT.
3-All the OLT are placed in the Central Office to support a neighborhood.
4-Every OLT port is able to support up to 32 Customers that subscribe.

So: in a neighborhood with 1024 houses, 32 OLT ports are lighted.

When a Customer Subscribes:
Day 1 - Fiber is dropped from the pole to outside of the house.
Day 2 - A verizon Guy comes to the house to do the installation.

Any feedback on this understanding?

Signature: A Happy FiOS customer !!!
optiplayer
50%
50%
optiplayer,
User Rank: Light Beer
12/5/2012 | 3:39:23 AM
re: Figuring FiOS
Stephen,

Adding to your point about opex savings, I suspect that a significant portion of the current improvement they are seeing comes simply from the fact that the plant is new. If they had gone out and replaced the old copper with new copper they likely would have seen tremendous improvement in opex as well. Let the FTTH stuff cycle through some snow, sleet, ice and wind storms for a couple of years then get back to us Verizon.

In fairness to Verizon the 35-40% projection is for broadband. The video share projection is 20-25% which seems reasonable.

All that said, I don't care how much it cost Verizon I demand FTTH even if the government has to step in an subsidize it and it must be a dumb pipe!! (just doing my MaterialGirl impersonation)
DCITDave
50%
50%
DCITDave,
User Rank: Light Beer
12/5/2012 | 3:39:24 AM
re: Figuring FiOS
re: "the financial case for FiOS is driven in large part from opex savings"

Which means layoffs, according to some in the outside plant business.

Good points on ARPU, though. Pretty silly of me to assume that 100 percent was going right back to pay for the network.

ph
stephencooke
50%
50%
stephencooke,
User Rank: Light Beer
12/5/2012 | 3:39:24 AM
re: Figuring FiOS
The other really nasty is that the opex reductions are not what they appear either. Unless Verizon gets lets say >85% (at that level perhaps the rest would be forced onto the fiber architecture) take rate on Fios in your neighbourhood they can't turn off the copper.

In their presentation they also figured that they would achieve 35-40% take rate by 2010. It took cable over 15 years with just rabbit ears as competition to reach 30%. Given the competitive nature of the areas that they are upgrading this should be good to watch. I guess it would be even better if you were in that area.

The most interesting thing though was that their originally installed CPE apparently won't support FMC or MoCA. This means that for Verizon wireless to bring in the 4-play option that the CPE has to be upgraded (or another external WiFi box supplied which would theoretically be taken by the customer when they move. The ONT stays with the house.). To offer the 4-play package it might make sense to replace the ONT which would be more capital cost.

Opex costs would be even lower if the xPON technology used was ubiquitous throughout the rollout. Fios started with BPON and is now GPON and may move to WDMPON before 2010.

I also found it interesting that Doreen dodged the question about 18M households past by 2010 and the thresholds for increasing or decreasing that number. She said that they would re-examine things in 2007 or so.

This is an enormous undertaking with many senior executives' jobs on the line for it. This will be an education for us all.

Steve.
Michael Harris
50%
50%
Michael Harris,
User Rank: Light Beer
12/5/2012 | 3:39:27 AM
re: Figuring FiOS
I think you're assuming the $150/month ARPU is pure profit, which it isn't.

Yep. With that ARPU, profit might be more like $50/month. In that case, the pay back for $9K per sub would take 15 years. Pretty ugly.

In fairness to Verizon, what they've said all along is the financial case for FiOS is driven in large part from opex savings they expect to see from running a passive optical network, as opposed to their legacy copper plant. So really, this is as much about cost savings and existing customer retention as new ARPU.

davrich
50%
50%
davrich,
User Rank: Light Beer
12/5/2012 | 3:39:30 AM
re: Figuring FiOS
I think you're assuming the $150/month ARPU is pure profit, which it isn't.
DCITDave
50%
50%
DCITDave,
User Rank: Light Beer
12/5/2012 | 3:39:30 AM
re: Figuring FiOS
I guess the real key is customer sign up and ARPU, isn't it?

If they get $150/month from the 40 homes I made up above, that pays for itself in less than 5.5 years.


ph
Michael Harris
50%
50%
Michael Harris,
User Rank: Light Beer
12/5/2012 | 3:39:31 AM
re: Figuring FiOS
Nothing like back o' the napkin math. $9K per connected sub? Ouch! What's funny is that Comcast and Time Warner paid under $3,300 per sub for Adelphia. With a market cap around $76B, Comcast itself isn't valued much higher. Were it not for those pesky antitrust rules, it would be smarter for Verizon to buy Comcast rather than overbuild them. ;)
More Blogs from The Philter
Jio's competitive market, fast growth and expanding customer base present some interesting machine learning and analytics challenges for Guavus, its newly announced analytics partner.
It's going to take some televisionary moves for pay-TV providers and big studio owners like AT&T to sort out what consumers want, how to package it and what to call it.
Machine learning is primed to help service providers run more efficient and effective networks, but first the good ideas have to make their way from the lab to the real world – and that's a big challenge, according to the University of Chicago's Nick Feamster.
Light Reading's editors discuss Dish Network, its pioneering past, a few hilarious missteps and why the company seems just as likely as anyone to be the next big player in 5G networks.
Light Reading's Kelsey Ziser discusses the world of SD-WAN, why there are so many vendors and whether enterprise 'white glove' services are handled by butlers.
Featured Video
Upcoming Live Events
September 17-19, 2019, Dallas, Texas
October 1-2, 2019, New Orleans, Louisiana
October 10, 2019, New York, New York
October 22, 2019, Los Angeles, CA
November 5, 2019, London, England
November 7, 2019, London, UK
November 14, 2019, Maritim Hotel, Berlin
December 3-5, 2019, Vienna, Austria
December 3, 2019, New York, New York
March 16-18, 2020, Embassy Suites, Denver, Colorado
May 18-20, 2020, Irving Convention Center, Dallas, TX
All Upcoming Live Events