Christopher Herot's Weblog

Mossberg's Verdict on Mobile Me: Not Ready For Prime Time

Steve Jobs must yelling in the corridors at Apple today.

Walt Mossberg, whose reviews can make or break a product launch, delivered his verdict on Apple's MobileMe service: Apple’s MobileMe Is Far Too Flawed To Be Reliable.  As I read his description of how MobileMe is intended to synchronize email, calendar, and contacts in real time among PCs, Macs and iPhones I recalled my experience at Lotus when I managed the EasySync business.  My first thought was how ambitious it was for Apple to think they could handle all the subtle variations and undocumented features among the different versions of Outlook, Outlook Express, Windows Mail, iCal, and the iPhone, not to mention suporting web-based version of those programs in various browsers.  My second thought was that a lot of the bugs Mossberg found must have shown up in testing but Apple released it anyway.  I can just see some QA manager telling his boss, "you'd better hope Mossberg doesn't have these problems."

Most of the problems Mossberg found are nits, such as not having the sync'd calendar show up automatically, and may indeed be rare, but he found a lot of them.  Synchronization is a subset of the distributed database problem, and distributed databases are hard.  Not only must they deal with conflicting transactions happening in parallel (e.g. updating a contact's phone number in two places at once) but they must also deal with different interpretations of the various standards, and the tendency of humans to be imprecise, such as storing a "home number" one place and a "home office number" in another.  Add in a couple dozen different builds of Outlook and you have plenty of opportunity for a developer to say "but it works fine on my machine" or a manager to say "that won't happen to most of our users."

Now these bugs will probably get the attention they deserve and the next version of MobileMe will be much improved.  For that we can thank Walt Mossberg, who goes there so we don't have to.

PodCamp Boston - It's About People and Stories

At PodCamp Boston last weekend, one theme ran through all the presentations and discussions, which was that the best way to get and hold people's attention was to tell them compelling stories about other people.  As Jack Valenti said many times about the movie industry, you can spend all you want on famous stars or special effects, but what draws an audience is "a good story, well told."

This point was made many times with regard to web-based businesses, but artist Keith Burtis described how it worked for promoting his art.  Burtis writes about the process of making his woodturned art on his blog, and does live videos on BlogTV.  He said people were much more interested in purchasing a piece of art when they saw or heard the story of how it was made, even to the point on insisting on purchasing the object in the video instead of a nearly identical work made in private.

Burtis offered some additional examples, such as Natasha Wescoat's art tutorial videos and art news blog which have made her a top seller on the Web and led to a lucretive licensing deal with Bed, Bath & Beyond.

Shelley Greenberg made this point and more in her summary of PodCamp's Five Lessons:

1. Social networks are a means, not an end.
2. Content is king.
3. Participate, participate, participate. And then? Participate.
4. Everyone, no matter how much gravitas they command, is human.
5. Help yourself by helping others.

PodCamp Boston

The Boston Twitterati showed up in force for the third PodCamp Boston.  Close to 500 social networking enthusiasts packed the gleaming Joseph Martin Conference Center at the Harvard Medical School this weekend for sessions on social media marketing, podcasting, and other new media topics.

In addition to the sessions, meals and hallway conversations, one of the mist distinctive features was the heavy use of Twitter.  The vast majority of the participants are followers of @pcb3 and used the service to keep up a running commentary on the event.  While most conferences with parallel tracks can induce an anxiety that some other sessions is better than the one you are in, at this event one can follow the tweets from people at other sessions and exercise the Law of Two Feet if another session sounds more relevant.  Following someone (and/or friending them on Facebook) suplanted the usual exchange of business cards.  And of course Twitter proved invaluable organizing the many impromptu lunch and dinner gatherings.

Photos here.

Technology Review on Web 2.0

As MIT's Technology Review completes its transformation from a sleepy alumni magazine to "The Authority on the Future of Technology" it devotes the entire June/July Issue to the topic to "The Next Bubble? - The Future of Web 2.0."  The cover photo (at right) shows Leah Culver, co-founder of Pounce, blowing up a bubble.  You need to turn to page 35 to see the result.

Editor-in-chief and Publisher Jason Pontin recalls his experience as editor of Red Herring which in the first six months 0f 2000 had 500,000 readers and $100 million in revenues and was out of business the following year  He sees a similar enthusiasm today for companies with unproven business models, but more than the potential loss of capital he fears for the future disappointment of the entrepreneurs who put their hearts, souls, and considerable time into the ventures which will not survive the inevitable correction.  Still, he recognizes that speculative manias are an inescapable feature of entrepreneurial capitalism and in fact were responsible for the investments in many businesses, such as transportation and the Internet, that influence us today.

This issue of the magazine takes a close look at social networking which Pontin notes is a "nice return to the Web's foundations."  In fact the last page writes about how Robert Fano (right, with Marvin Minsky), who was my graduate adviser at MIT, organized Project MAC on the theory that ­"general-­purpose, independent, on-line use of computers by a large number of people" appeared feasible as far back as 1963.  That probably explains why he was so tolerant of my doing my thesis work with the group that became the Media Laboratory even though they espoused the idea, thought irresponsible at the time, that ordinary people could have personal computers and that an entire bit of memory could be devoted to lighting up a single pixel on a screen.

Freelancer Bryant Urstadt delves into the details in an article Social Networking Is Not a Business* (*But It Might Be Soon) in which he describes the large user bases and equally large amount of venture capital that have been attracted to KickApps, Ning, and Bebo, not to mention Facebook, MySpace, and LinkedIn.  He also points out how revenues and, especially earnings have lagged behind those numbers.  More tellingly, CPMs vary wildly, from $70.00 at sites that have desirable demographics, such as Technology Review, to $0.13 at Facebook where no one has figured out how to get the social networkers to click through and buy anything.  Jason Calacanis is quoted as saying social networking sites are "a really bad place to advertise."  The reason is that while 2% of Google users click on a given ad, in Facebook the number is less than 0.04%, a disparity explained by the difficulty of targeting users who come to the site to socialize rather than do research on purchases.  As the article explains, attempts to get around the problem by gleaning users intentions from their transactions and demographics embroil the sites in privacy problems, as happened with Facebook's ill-starred Beacon program - "an advertiser's dream - and like many things that are good for advertisers, very annoying to ordinary folks."  The are also issues of "content adjacency" - advertiser's concern at having their ad placed next to user-generated content of uncertain taste.  The consensus of the pundits quoted (Andrew Braccia, Chamath Palihapitiya, Paul Kedrosky, Roger McNamee, Marc Canter, Ron Conway) is that the advertisers will eventually get over their squeamishness and follow the users to where they are spending their time, and that revenue can be generated from a long list of transactions such as commerce.

Other articles cover many of the other hot topics in this space, such as: who owns user's personal information, Facebook's IT infrastructure, Network Neutrality, and even a photo tour of Twitter's world headquarters.

While I agree with the authors in this issue that the high tech industry does have cycles of irrational exuberance followed by painful crashes, I prefer to take the longer view and see those cycles as the tide moving in and out of the beach.  When it comes in it brings lots of interesting stuff ashore, and even though the tide goes out, most of the stuff stays behind.  If the investment of money and human capital was completely rational, we probably wouldn't have railroads or the Internet.  Each expansion brings in waves of new people, but the ones who stay understand the cyclical nature of things and learn to enjoy the bubbles as they expand and remain philosophical when they contract.

In summary, this issue of Technology Review offers a good summary of the issues facing the tech world these days.

Spectacle Island

As you approach Boston Harbor by water, you pass through an extensive network of islands.  For several hundred years, these islands were used and abused as places to keep things the residents did not want nearby, such as exiled Native Americans, newly arrived immigrants, quarantine hospitals and unpleasant industries of varying sorts.  Arguably Spectacle Island suffered more than its share, with a horse-rendering plant, a grease-extraction factory and eventually a city dump.  Leaking toxic waste and smoldering underground fires forced its complete abandonment in 1959.  Fortunately the reconstruction of the Central Artery and Tunnel , the CA/T project  a.k.a. the "Big Dig" offered a means to reclaim the island.  By covering the old dump with 3.8 million cubic yards of dirt excavated from the new tunnels, the island has been transformed into a beautiful 121-acre park, complete with a public beach, visitor center and cafe, and some spectacular views from what is now the highest elevation in the harbor.

The island is 15 minutes from downtown Boston by boat.  During the summer months the Harbor Express high speed ferry leaves every hour from Christopher Columbus Park adjacent to Long Wharf and costs $14 round trip.  In theory one can make reservations but the boats are large and there were plenty of seats on a prime weekend in July.  Once you are on Spectacle Island, you can take a shuttle to the many of the other islands in the harbor, but check the schedule as making connections requires some planning. Also, on our return trip, the high speed ferry was replaced by a much larger and slower boat, which accommodated everyone wanting the leave the island but took 45 minutes to get back.  If you bring your own boat, there is a marina (VHF ch 69).

Spectacle Island has a beach, a visitor center where you can sit on the porch, get a burger from the cafe, and see an exhibit on the islands history and ecology (20,000 plants were re-introduced).  There is a network of trails with picnic tables and views of the harbor and neighboring islands.  Well worth an afternoon - or an entire day if you want to see several islands.

Photos here: herot.typepad.com/photos/spectacle/index.html

For more information:

• Boston Harbor Islands brochure
• Island Alliance web site
• National Park Service site
• Jasper White Summer Shack Clambake every Thursday (6:15 - 9:00)
• Island Regatta, Sept 26-27

Fireworks in Boston

Viewed from the Cambridge side of the Charles River.

The The Boston Globe has some good coverage here of how it was done, including the video tour of the fireworks barge, below, conducted by Eric Tucker of Pyro Spectaculars who has done the past seven Boston shows.

Is YouTube Similar to Videotape?

Wendy Seltzer points out that Wednesday's discovery order in Viacom v. YouTube, which requires Google to release the login names and IP addresses of every person who has ever viewed or embedded a YouTube video may be in contradiction to the Video Privacy Protection Act which prohibits a "video tape service provider" from disclosing "personally identifiable information."  The act defines "video tape service provider" as "any person, engaged in the business, in or affecting interstate or foreign commerce, of rental, sale, or delivery of prerecorded video cassette tapes or similar audio visual materials."

It would seem difficult to argue that video-on-demand is not "similar" to video cassette rental.   Google did cite this law in their argument, but the judge then cited a blog posting by one Google's own engineers that argued IP addresses are not personal because they are usually assigned dynamically.  That may be true in theory, but in practice most DHCP servers hand out the same address each time a given user requests one.  Many people exploit this property to host web sites on their cable modems and the address can stay the same for months, if not longer.

I suspect Google has decided not to appeal this order because if the information in their logs was held to be personally identifiable, the law would prohibit Google from retaining the data any longer than a year, a topic on which their privacy policy in conspicuously silent.

New Data on the Long Tail

Update [21 Jan 2020]: Since I wrote this twelve years ago, the Long Tail theory has fallen into disfavor. Chris Anderson took down the original article I referenced. I have changed the link to a post by Hannah McNamee on what happened.

Chris Anderson's theory of the Long Tail became one of the most influential ideas of the Internet era, firmly establishing the idea that the declining cost of producing and storing digital content would shift the concentration of sales from a few blockbusters to a "long tail" of diverse products.  Although the sales of each product would be smaller, the much higher numbers of such products would result in a greater opportunity.  Now Harvard Business School professor Anita Elberse has published an article in the Harvard Business Review, Should You Invest in the Long Tail?, that argues that sales and profits still rest with the blockbusters and that the Internet's ability to let customers share recommendations may actually increase the concentration of hits.

Elberse cites data from music and video sales and rentals to show that while the tail is growing, most of the sales are concentrated at the short end.  For instance, at Rhapsody, the top 10% of titles account for 78% of all plays, and the top 1% account for 32%.  However, since Rhapsody offers over a million titles, that 1% accounds for 10,000 titles, which is more than most U.S. radio stations play in a year and equivalent to the entire music inventory at a typical Wal-Mart.  However, as readers have pointed out in comments on the HBR site and Chris Anderson's site, whether this data confirms or contradicts the Long Tail theory depends on the definition of long tail. For instance Rhapsody, by carrying 100 times as many titles as Wal-Mart, automatically decreases the percentage of titles represented by the top sellers, but at the same time tripled total sales.  If these sales needed to be fulfilled from a physical warehouse this might not be a profitable move, but in the digital world it's a favorable result.  In fact, the mere presence of such titles might be an incentive for customers to patronize Rhapsody, even if they mostly bought the blockbusters.

Those debates aside, Elberse comes up with some interesting data:

• Customers who purchase a lot of content tend to buy the hits as much as anyone, but they also buy the more obscure titles.
• The light consumers tend to stick to the hits, analogous to the moviegoers who only go to the movies a few times a year in order to see a major blockbuster.
• Consumers of both types tend to give higher ratings to the popular products.
• The long tail products are more likely to be consumed by people who stick to a particular genre, such as classic rock ore romantic comedies.
• Of the 3.9 million digital tracks sold in 2007, 24% sold only one copy and 91% sold fewer than 100 copies.

One lesson here is that the long tail can be quite profitable as long as the cost to build and maintain the inventory is low - or zero in the case of the really obscure titles.  This accords with what I've learned from my recent conversations with people in the music industry, where the back catalog of fully-paid-for works is often the source of all the profit while spending production and promotion budgets on new acts is more risky now that the average payback is lower.   The other lesson is that the maximum return is still at the short end of the curve, but that short end is significantly fatter than it used to be.