February 13, 2010 6:12 AM

Google's "Think Big Gig": What Is And What Should [Will] Never Be

And if you say to me tomorrow
Oh, what fun it all would be
then what's to stop us, pretty baby
but what is and what should never be
-Led Zeppelin, "What Is And What Should Never Be"

With profuse apologies to Led Zeppelin for blaspheming their iconic song title to do a telecom policy blog, this is essentially what Google announced to DC policy makers, via its corporate/policy blog, on Wednesday--except that the policymakers and the press didn't hear the last line.  But, boy, did they eat up the first few . . . you can tell that Valentine's is in the air.

I say the "announcement" was targeted toward policy makers, because absolutely no relevant business information was provided in the announcement--you know . . .  costs, prices, projected revenues, technology to be used, etc.  No vendors, competitors, or even Google's Clearwire partners (a venture from which--according to news reports--Google has been backing away) were interviewed or consulted.  No, but that's OK, because this wasn't a business "announcement."

What the "announcement" really says is how much political clout Google carries in Washington.  On a day when the Gub'ment is closed for a fourth consecutive day, some of the most important Government officials involved in technology policy were intrigued enough to very quickly issue "statements" in reaction to Google's blog post.

For example, the New York Times story actually contains a "statement" from Chairman Genachowski reacting to the Google blog post, and the statement reacts to Google's announcement like it were an "official" announcement--like a firm commitment to enter a market in a specific way, explaining product terms and prices, entry timing, costs, and projected revenues.  The Hill even contains a statement from Senator John Kerry, Chairman of the Senate Commerce Committee's Subcommittee on Communications, Technology, and the Internet.  Moreover, just about every story you'll read really "drank the Kool-Aid."  From the articles I saw on line, only Computerworld got it right.   

But what gives me the right to question Google's ambitiously-admirable, but vaguely-defined, "experiment", the belief of the bulk of the press, and some of the most important officials in Washington?  Well . . . there's this small problem of the facts and the logic.  First, Google's blog never says exactly how they plan to offer this 1 gigabit/sec (1,000 megabit/sec) broadband service at a "competitive price."  Second, the whole theory seems to contain a pretty glaring logical flaw: wouldn't Google deciding to become a broadband ISP allow other Broadand ISPs into Google's monopoly business?

To the first point, when we're talking about high speed Internet access (at any speed), it's worth noting that some "bits" are different from others; there are "Cadillac" bits and "economy car" bits. In other words, there are differences in quality.  Are we talking fiber-to-the-premise, full duplex (same bandwidth in both directions) service for ultra-high speed transmission service at the "carrier-class" standard of "five nines" (99.999% up time) guaranteed reliability?  Or are we talking about a "best efforts", asymmetric, download speed (with a much lower upload speed)?  It makes a difference.  

For the "Cadillac" broadband Internet access, Cogent Communications is a good example of such a company.  You can read about them here.  To save you the trouble of the click-through, Cogent has been in business for over 10 years, has over 1,100 "on net" buildings in North America, and serves most of its customers over that fiber network.  A largely "on net" experience allows Cogent to guaranty a "gold standard" of service.

Cogent's most recent publicly-available 10K (admittedly, almost a year old) for the year ending December 31, 2008, explains that Cogent has over 17,800 retail customer connections.  Moreover, Cogent offers a "1 Gig" Internet access service, as well as a carrier-class "10 Gig" service.  Yet, as they note in their 2008 10K, Cogent's most popular service in North America is its 100 megabit Ethernet Internet access service. (See, 10K, p.7)  Cogent further explains that this service is generally priced at $1,000 per month. Id.

On the other hand, for "best efforts" (lots of restrictions, no guarantees), residential, asymmetric Internet access service, the highest speed residential offer I've seen is Cablevision's Optimum Online Ultra┬« which offers 100 Megs down and 15 Megs up, priced at about $105/month.    

So, do you still wonder why I'm a skeptic?  If we've established a competitive price range of $100 to $1,000 for the range of quality between the highest and lowest quality high-speed Internet access service offered at 1/10th the capacity Google is promising, then one has to wonder what a "competitively-priced" 1 Gigabit fiber service would look like.  This is the problem with the facts.  Internet access service at speeds 10 times the fastest Internet speeds for residential and small business users will, at best, likely range from several hundred to several thousand dollars per month--which is why potential RFI Applicants are asked to provide so much information about per household income and Internet spending.  Just request an application and find out.

The problem with focusing too much on facts, however, is that facts can sometimes be the trees that obscure the forest; so, too, here.  The very thorough Gary Kim of IP Business  provides us with some useful information.  Of the 120 million, or so, U.S. households, about 87 million are online, and 80% of those (about 70 million households) have access to broadband. Similarly, of all residential search engine searches performed by residential customers, Google's share is about 80%. In its most recent quarter, Google obtained about 97% of its record revenues from advertising. Let's step back now, and look at the big picture.

A year and a half ago, it looked as though facilities-based ISPs, like AT&T, Verizon, and Comcast might enter the behavioral-targeted, online advertising business (the same one Google dominates).  The concern of incumbent providers like Google was that the ISPs would participate in the market in the same manner that traditional search engines and social networks gather information from consumers--surreptitiously--and provide this information to advertisers in competition with Google.  Google urged hearings to ensure that Congress would protect consumer's privacy. 

However, the responses of the ISPs surprised their critics.  None of the ISPs had used consumer search information to sell advertising services, but all indicated that they would not use a consumer's information unless the consumer had explicitly "opted in."  See responses to Congress

This approach was considerably more consumer-friendly than the approach used by Google and Facebook, which basically require consumers that use their products to surrender all of their personal information and privacy.  Needless to say, when Google wasn't the "good guy", Congress quickly lost interest.

So, from a matter a pure logic, the main reason that I am suspicious that Google will ever become a broadband operator is that it will essentially wreck Google's existing business, which relies heavily on requiring consumers to accept a lack of privacy in order to use Google's product.  Interestingly, on the same day Google announced its desire to get into the ultra-high-speed consumer broadband business, Google embarrassed itself with a flagrant disregard of consumer privacy in launching its "Buzz" product designed to compete with Facebook and Twitter. 

So, the bottom line is that Google's claims seem specious for two reasons: 1) they have not explained how they will enter a competitive business, on a small scale, and provide better services at lower prices than the firms already in the high-capacity Internet access business; and 2) if Google becomes a broadband ISP, then all the "consumer privacy arguments" and restrictions that Google has previously used to keep broadband ISPs OUT of the Internet advertising market will be shouldered by Google.  Alternatively, if Google does not offer consumers "opt in" policies so that consumers can decide whether they want their search query activity sold to third parties, then Google has no argument to expect other broadband ISPs to do the same.  

If Google's experiment is wildly successful, and they get 500,000 homes, they then allow the broadband ISPs that control the other 70 million households to use the same practices that Google uses to garner customer information to generate advertising revenues.  Google's choice as a broadband ISP is to either 1) offer consumers some privacy protection (not the basis of Google's current business success), or 2) allow other broadband ISPs to use the same tactics Google uses to collect advertising data, and thereby invite multiple effective newcomers into Google's monopoly (by market share).  To Google's management, I'm afraid these facts, and their consequences, make the forecast for Google's broadband ISP experiment a case of What Is And What Should Never Be.

Leave a comment