Comment by callmeed | Hacker News Reader

Comment by callmeed | Hacker News Reader

Completely inaccurate back-of-the-napkin calculations:

- There are about 1 million households in LA (3M+ population but 3 people per household in LA County)[1]

- Assuming the vendor got 75% sign-up for the free tier, that's 750K households

- I'll give them a generous freemium-to-paid conversion rate of 10% (sorry but 5Mbs is fine for many people). That's 75K households.

- I'll be generous and say that most paid households will get some kind of bundle and pay $150/month

MONTHLY HOUSEHOLD REVENUE: $11,250,000

- There are 250K non-farm businesses in LA County. I'm not sure how many are in the city limits, but let's say 25K buy a paid. Since we know ISPs love to gouge offices with higher prices for the same service, let's say they get $250/mo from each business.

MONTHLY BIZ REVENUE: $6,250,000

- Quick search puts Comcast's ad revenue is about 4% of total (Charter is 5%). So I'll throw in an extra $1M/MO in ad revenue from this deal, which I think is generous.

TOTAL: $18.5M/MO

TIME TO RECOUP $5B: 270 Months (22.5 years)

[1]http://quickfacts.census.gov/qfd/states/06/06037.html

EDIT: Updated with ad revenue numbers

Great analysis! Some suggested tweaks below.

SALT:

1. You estimated revenues - let's multiply that by AT&T's 5.7% 2012 net income margin [1]. Y1 earnings: $12.7 million.

2. That seems low - let's double your business revenue estimate. Adjusted Y1 earnings: $17 million.

3. Los Angeles-Long Beach-Anaheim real GDP grew at 2.4% a year between 2009 and 2012. Assume that continues into perpetuity. AT&T pays 5.3% and 3.7% on its 2041 and 2029 debt, respectively - let's say you can borrow between those.

--> Present value of cash flows: $0.6 to $1.3 billion. Quadruple your revenue estimates and borrow at 3.75% to hit $5 billion.

CAYENNE:

1. $5 billion to roll out fibre across Los Angeles seems optimistic. The issue is compounded by the city council having no skin in the game, even with regards to bureaucratic inefficiencies (e.g. some land use or environmental commission jamming the entire project for 5 years).

SUGAR:

1. Your revenue estimates look pessimistic - businesses could pay differing rates based on usage and more than 1/4 of the population could sign up. You also don't include government demand. But it's thorough enough to look closer than 4x off.

∴ The RFP in its present form looks more like a political stunt than a serious proposal. Not sure what the LA electoral calendar looks like over the next 3-6 months.

[1] http://finance.yahoo.com/q/is?s=T+Income+Statement&annual

This assumes 100% gross margin. If they are buying a bundle, you probably mean TV, so they are going to have to pony up to the networks. Then they'll have to pay for upstream (I don't think AT&T etc is going to freely peer with a 750k residential/commercial network), then they'll have to pay for all the support costs associated with running a fiber network.

I'm thinking they'd be lucky to be able to use half of $18.5M/MO to pay off their $5B investment. We're now at 50 years! And that assumes they can keep their price high despite technological advancements over the next half-century (imagine paying $150/mo for an internet connection using technology from 50 years ago). Oh, and they have to wholesale it to their competitors. There is 0 chance this is happening with these numbers.

Yes, good point.

My thought with a bundle was TV, land-line phone/fax (businesses especially still need this), and possibly even mobile service because I'm assuming AT&T and Verizon are on the short list of vendors anyway.

My short list of vendors who are capable-and-possibly-dumb-enough to consider this are: AT&T, Comcast, and Verizon.

On the other hand, some tech companies with streaming deals/services could maybe do this without any bundle. Could even force some networks to get with the times. So my shot list of vendors who are awesome-enough-but-probably-way-too-smart to consider this are: Amazon, Netflix, Google & Apple.

In my experience, Cable/Telco companies don't build fiber networks with greater than 15 year payback, with 30 being the hard cap. I think your math is fairly accurate.

Source: See my other comment, I did fiber rollouts with telcos including Comcast in the Bay Area.

Right on. And it will 100% go over budget given the amount of work needed, ancient infrastructure, non-compliant landlords and the "I just don't give a shit to change" LA attitude.

I remember two years ago when I was the CTO of a 100 employee company in Beverly Hills just how difficult and annoying it was to get a single 10/10Mbps Fiber link to our offices.

It took 6 months from the date the order was placed to the provisioning. And it took another 2-months to go from 10-20Mbps. It's totally insane that these things take so long and it was 50% the fault of AT&T and I kid you not - 15 employees at AT&T dealing with this - and 50% the fault of the building who didn't even know what Fiber was and thought it was a back link from their telco room to our servers.

Insanity

*- There are about 1 million households in LA (3M+ population but 3 people per household in LA County)[1]

What am I not seeing? The census link states population: 9,962,789, Housing units: 3,449,273.

10M is the county population. Choose LA from the city dropdown on that census link and you get 3,857,799, and 1,413,995 housing units.

That's still a pretty significant jump over the 1M household estimate the OP had made, though.

Sorry, I didn't see the city dropdown. I was just using the 2.99 persons per household to estimate based on 3M people. I can't edit anymore unfortunately.

Reply on news.ycombinator.com