Google still winning cloud race

I now have firsthand experience with various cloud based email services, and Google still remains firmly in the lead.

My life is pulled in (at least) three directions when it comes to cloud-based (aka hosted, aka SaaS) mail services:

• This month, my employer (a Microsoft shop) decided to migrate from an Exchange server to the Office 365 hosted services. I guess as an IT-knowledgeable employee, they made me one of the guinea pigs. So this week I'm trying to reconfigure my Mac and cellphone to work with the new servers.
• Meanwhile, on July 1, my wife was forced to migrate from mac.com (aka me.com) to iCloud. As the household IT support desk, the task fell to me, and we still haven’t been able to get it to work with her Mac.
• Finally, my teenager and I are loyal users of gmail and other Google services. My teenager won’t use a client app anymore, while I use my various gmail addresses with my Eudora client. Both of us also use Google Voice.

(I also have an old Yahoo web mail account, but since they don’t support client apps for free, I only give that email address for website registration and other spammers.)

From what I’ve seen so far, Google remains far ahead for web-based services. This is not to minimize the advantages Microsoft and Apple have for their locked in proprietary client customers.

A few years ago, Google and MS were warring over providing hosted mail and office apps to the 23-campus California State University system, America’s largest university system. Google won and at SJSU we were migrated in mid-2010..

However, before that the SJSU business school was an Exchange shop, as was my previous b-school and my current employer. So despite being a Mac user since 1984, I was forced to deal with Mac/Exchange interoperability issues (which as much better than when I was researching my dissertation 15 years ago).

Microsoft Outlook Web Services are not very impressive so far. The web client seems slower than the old MS web app.

It was terrible — absolute pits — for explaining how to configure a 3rd party client (cellphone or whatever). First off, can’t find that help starting from scratch in the online help. I could only find it because my employer provided a link. Secondly, they don't publish their POP/IMAP/SMTP server settings on a web page like normal web services. Apparently the settings are client-specific (which suggests their DNS load balancing technology is inferior to Google’s) Third, if (after logging in) you want to find the mail settings, the steps are so complicated that they want you to watch a video. Since I was on a lousy airport WiFi connection, I figure out how to get through the various windows (also buried in their web page) to find the answer.

On the Apple front, the iCloud migration is going badly. Mac.com and Me.com supported Internet standards, but iCloud deletes POP support and their IMAP implementation is incompatible with my wife’s Eudora client. So we are stuck on webmail until we find a replacement for the client or iCloud.

Meanwhile, Google has a huge lead in features and design. The gmail server supports all the protocols, multiple desktop and mobile clients. And if you find the mail server or client too limiting, you can forward to any other server.

Yes, Apple is going to get me.com customers from their iPad lead and iPhone sales, and Microsoft is going to pick up all the firms that run all-Microsoft shops. But if an IT manager is try to pick the best solution, Google seems to be winning both on its execution and its standards-based approach (allowing third-party integration).

Scalable services

IT enabled scalable services such as Google and other SaaS vendors are quite different from the millennia-old model of labor-intensive services. This was the point I made in an open innovation talk earlier this week — a point I have been mulling over ever since.

Near the end of and near the end I discussed business models and the general shift away from mass production towards services. Below is a slightly expanded version of one of my slides.

Model	Example	Basis
Personal services	Blacksmith, barber	Personal skill, locaiton
Mass production	Cotton gin, Springfield rifle, Model T	Design cost, manufacturing, economies of scale, cost of goods
Information goods	Content: WSJ.com article, iTunes song†	Design cost, economies of scale, cost of license
Automated services	SaaS: Google map, search, mail	Design cost, economies of scale

† excluding information goods delivered in tangible form.

I won’t claim it’s terribly profound, but the act of making the table forced me to think about some implications.

One point was an old one — there are “services” gurus who confound information goods and services to make their area of interest seem more important than it is (even though it would be very important even without exaggeration). Stamping out an identical information good over and over again is not delivering a service — it’s selling an intangible product as is well discussed by the book Information Rules. (Normally we’d think of this as $0 COGS product, but certainly an important class of information goods are sold based on royalties.)

Also, in this intuitive taxonomy I want to hold aside “services” involving the selling and renting of tangible goods, since much (or all) of the value comes from the good and not the customized personal experience. Services involving money also seem very different, even though they’re an important recent area of open innovation research.

Some businesses include a combination of products and services. With my MBA students Wednesday night — talking about disruptive innovation — I noted that for some low-priced commodity products, the cost of providing any personalized service (such as a tech support call) will destroy all the product margins from the product.

In my talk, I also briefly mentioned the “Pharma 2010” view of systems biology by PwC Consulting (now part of IBM). If I’d had time to track down the PDF, I would have put up the opening paragraph of the PwC study (instead of paraphrasing it):

In 2010, the pharmaceutical industry (Pharma) will not only make white powders; it will sell a variety of products and therapeutic healthcare packages that include diagnostic tests, drugs and monitoring devices and mechanisms, as well as a wide range of services to support patients. Companies that learn how to make “targeted treatment solutions”, as we call them, will deliver bigger shareholder returns than they have ever delivered before.

Even this hastily drawn, over-simplified taxonomy communicated the point that I thought was important for the middle managers to understand: don’t think of services the way we used to do — or perhaps the way Accenture or EDS or IBM Global Services does — as labor-intensive, low-margin businesses. Instead, think of them the way Google does — high up front cost, positive returns to scale, that are scaleable indefinitely. These sorts of 21st century business models are completely different than the Bronze Age services model that we normally consider, and are quite feasible for companies that have access to unique and valuable knowledge that can be delivered electronically.

In making this argument, I was dimly recalling (but had no time to look up) what Randy Stross said in the talk he gave last year about his book Planet Google. The oral presentation emphasized that the ideal pursued by Sergei and Larry — in their zealous embrace of algorithmic solutions — that humans shouldn’t touch anything, but instead everything important should be delivered by the computer and not by manual labor.

By using the index and browsing my paper copy of Randy’s book (now in paperback), I was unable to find the relevant passage (which I thought would come in Chapter 3, “The Algorithm.”). However, thanks to Google Books — I did find a discussion of Sergei and Larry’s scalable business model in Chapter 2 (on pp. 48-49).

Michael Vick joins EA's last hurrah

Michael Vick is finally back. No, I don’t mean in the NFL or in the public eye. I mean back in Electronic Arts’ John Madden football game.

In an annual ritual, the latest EA football game (now Madden NFL 10) went on sale at 12:01am Friday morning for Wii, Xbox360, PS3, PS2 and PSP console owners.

However, the announcement of the new Philadelphia Eagles backup quarterback came long after the game had been released for duplication. Instead, the EA developers in Marin County had to scramble to add the backup quarterback to the game for the next online update.

The Madden NFL title has been the perennial EA blockbuster, and the Merc even speculated that it could help the overall industry:

"Madden" arrives just in time for the video game industry, which has suffered from flagging sales due to both a lack of blockbuster titles and the decline in the economy.

According to market researcher NPD Group, sales of video game software, hardware and accessories came in at $848.8 million in July, down 29 percent from a year earlier.

"While year-to-date results are weak, there are some big titles set to be released over the next several months, including 'Madden' this month, which should help spur sales," NPD analyst Anita Frazier said, according to our friends at The Associated Press.

I see this instead as the last hurrah for the best-selling sports game of all time.

I first heard of Madden as the coach of the last decent Raiders team (which brought him the familiar plaque at SFO airport). Most sports fans know him as the quirky sportscaster who refused to fly on an airplane. But now that he’s retired, a new generation of teenage boys will say “who???”

Meanwhile the secular trend for the software products business has been down, down, down. Kids steal software, or they play crappy free stuff online, or they don’t bother. The recession has converted their $500/year videogame habits from a “must” to a “nice to have,” particularly if mom or dad no longer has a job.

My guess is that in two years, NFL football from EA will be available as a fully subscription (SaaS) model, the way that Google already is and big software companies like Microsoft, Adobe and Intuit are heading towards. Hopefully EA will be able to charge more than Google for their subscriptions (or reap billions of dollars in ads for their free offerings).

Week in review

This week I’ve been swamped coaching my two simulation teams in the final rounds of their International Collegiate Business Strategy Competition. A long slog that began on Jan. 7 (during winter vacation) ended Saturday with both teams victorious.

However, I wanted to comment quickly on a few items:

• The WSJ had a great article (available free) on how Vizio used offshore manufacturing to come from nowhere to be one of the top TV makers — in a virtual three-way tie (with Sony and Samsung) for the most LCD sales in North America. Consumer electronics has been a brutal commodity business with entrenched competitors and (except for flat panels) excess capacity, so Vizio’s success (using open innovation) should be an inspiration for upstarts everywhere.
• Red Hat has beat a retreat from the desktop Linux business. Which Windows advantage matter most — the one from network effects or switching costs? (The exact question I tried to answer with my dissertation). I don’t know, but If Red Hat can’t make it, it’s hard to see how Linux is going to be a major factor in consumer or business PCs — at least in countries with high existing PC penetration rates.
• Adobe’s new Photoshop Express website got a very nice writeup in the WSJ — which basically said it’s about as good as the 1.0 of an online photo program can be. This shows that not only is Adobe trying to remake itself into a SAAS company, but its skills are transferrable. Also that by requiring Flash, its can continue to use its position in one software segment to boost another.
• Some (but not all) of the 4G wireless equipment makers agreed to a patent cooperating agreement to speed adoption of the GSM/W-CDMA derived LTE technology. It’s not clear if the parties have agreed to a formal pool or a set of rules — either for valuing patents or (as announced in August) for deciding which patents are essential. Given the past failure of telecom patent cooperation, this seems more like a promise to agree rather than an ironclad agreement.
• The patent “reform” bill S.1145 seems to be dying, with those big IT companies that want to weaken patents (e.g. Apple, Cisco) unable to overcome the opposition of those that like them just as they are. I wonder if anyone in D.C. understands “win-win” — such as recent efforts to make patent examination more rigorous and accurate.

One more cheer for the FSF

In noting the improvements in the latest draft of GPLv3, there is one I missed: dropping the SaaS compulsory sharing. Blogger Fabrizio Capobianco is ranting that the latest draft ruins the GPLv3, which he had counted on closing the “ASP loophole.” Matt Asay is drinking Fabrizio’s Kool-Aid® (or Pellegrino, or [fill in appropriate metaphor here]).

Messrs. Capobianco and Asay think it’s terrible that companies can bring open source in house, use it to offer services over the web, make their own changes and not give back those changes. (Think Google). The GPLv2 defines distribution as, well, distributing your software, so offering services that run on that software doesn’t trigger the compulsory sharing provisions. (If you haven’t been down this path, see Larry Rosen’s book in my OSS bibliography, or the GPLv2 chapter available free online).

At one point, GPLv3 was intended to trigger the GPL’s distribution clause with online services (such as ASPs and Software-as-a-Service), with the intention that any company using GPLv3-licensed code to deliver services would have to make their changes available to the rest of the world. As I found in my consulting practice, such mandatory disclosure would be a great benefit to dual license companies, because it would add one more hassle to slow competitors and force people to pay up.

Today a few diehards are annoyed that Google, Amazon and others run Linux server farms and don’t share their changes. But five years ago, they were desperate for such validation, and later proud that such (oxymoron alert) blue chip dot-coms endorsed a community-developed technology as suitable for the most demanding mission-critical IT operations.

Suppose Linus Torvalds agrees to the GPLv3 (his big issue seems to the anti-DRM clauses of the previous draft). Suppose the GPLv3 does “close the ASP loophole,” requiring Google etc. to share their changes with the world. So what will happen? Perhaps some would share their changes.

But I also think there would be a resurgence of interest in technologies made available under more open licenses (i.e., without compulsory sharing). Things like OpenBSD or FreeBSD might suddenly get popular again. Or the companies using the existing Linux under GPLv2 could just decide not to take any updates under the new license. In a free market, the decentralized choices of private actors are frustrating to social planners: when people have choices, they’ll do what’s in their own self-interest, not what the social planner wants them to do.

The FSF seems to be reacting realistically to market signals, despite what some of its idealist supporters would want it to do. It deserves to be commended for such pragmatism, because dealing with the world as it is will be a far more effective strategy in the long run.

Technorati Tags: Google, GPL, open source licenses, SaaS