Les effets de la numérisation
Par Bernard Golden, CEO de Navicasoft

Digitization Causes the New Rustbelt
Quick: what do you think about when you hear the term "Rustbelt"? If you're like me, you think of shuttered steel mills in Pennsylvania: abandoned hulks, exemplars of the long-superseded industrial age. The obituary for America's manufacturing sector is widely known: high-cost labor, inefficient manufacturing processes extended well past their replacement date, entrenched management unwilling to change, and a litany of a thousand more reasons. Frankly, in today's service economy, America's manufacturing past, for many people, sounds like a fable from antediluvian times. In other words, the rustbelt's condition is completely irrelevant to us, with no useful information to tell us about where today's economy is heading. After all, we're the new economy: building businesses that serve one another, making lives more enjoyable or productive - with our basic needs taken care of, companies that focus on meeting needs beyond the basic are those that will prosper. The service economy is rustbelt-immune.

Or is it?

I read an interesting story this morning about Kodak. Its travails over the past few years are well-known, and the story focused on its efforts to find its way in the new world of digital photography. Mentioned as a minor fact near the end of the story: "five mammoth plants where silver halide-based products were made or stored were reduced to rubble." Kodak, which grew fat on millions of people creating "Kodak moments," is a rustbelt business due to the onslaught of digitization.

In terms of understanding the upcoming trends in our economy, Kodak's tale is much more relevant than Bethlehem Steel's. Digitization will wreck current service-based companies and build a new economy on their rubble, and open source will provide the building blocks of the new economy.

Another industry under assault by digitization is widely discussed (perhaps over-discussed): music. With the easy digitization of music, the old business model of sales (and control) based on shipment of physical goods is rapidly shrinking. Obviously, the major music companies haven't responded very well to this trend. Labeling digital music consumers as thieves, suing users on shaky grounds, attempting to control the distribution of music via clumsy DRM mechanisms - the industry has practically developed a textbook in how not to respond to a shift in one's business.

Of course, there's an irony in all this. Despite the reduced sales, there is actually more music being listened to today than ever before in history. The ease of access made possible by digitization has created new and more satisfactory listening options that have allowed people to "consume" more music.

More use, lower sales. What should the industry do?

In "The Long Tail," Chris Anderson stated that digitization is great for musicians, because it helps them do the "really fun" part of being a musician: performing live in front of audiences. By generating more listeners via digitized music, more demand for live performances is created, offering musicians a way to make up for the lost royalties of physical music goods.

Frankly, I'm unconvinced of this. While performing live is undoubtedly one of the joys of being a musician, I'm not sure that most would prefer to replace royalties received in leisure for earnings generated by endless touring. In other words, making a living by performing live is hard work. I believe a solution that leverages digitization is more likely to be a more satisfactory long-term response to digitization. Since music is ubiquitous, finding ways to make special offerings that leverage ubiquity yet do not require in-person service (i.e., standing in front of a bunch of people warbling one's songs) is the way forward.

Just in the past few days the music industry appears to be moving away from customer conflict and toward a more creative solution for prospering in the digital age. The majors all appear to be considering a subscription-based offering, where in return for a set fee per month, music is available on an all-you-can-hear basis. This makes a ton of sense. In return for an easy-to-use service, I pay a single monthly fee, for which I get tremendous freedom to listen to music the way I want to (after all, the subscription model works well for the wireless industry; most people purchase monthly subscriptions despite using far less of the service than they're entitled to - the convenience and peace of mind far outweigh the poor economics of the subscription).

Furthermore, it's bound to be good for the music industry, because easy availability of music will encourage listener experimentation, exposing them to new artists, and further driving up use.

There's one fly in the ointment, however. I predict conflict about how the artists' slice of the revenue stream gets divvied up. Today, musicians are paid like baseball players: under-rewarded during their prime, over-rewarded on the downside of their careers. When subscriptions come in to being, there will be terrific conflict about how to fairly distribute the royalties among musicians, with aging rockstars demanding a share greater than their actual listening percentage.

But imagine what new music businesses can be created. With much more accurate listening statistics available, new offerings will be possible: boy bands formed to address a subsegment of the pre-teen girl audience; special pay-per-listen concerts made available only to subscribers who show a strong listening preference to Yo Yo Ma and Nadja Solerno-Sonnenberg. The music business will shift from selling access to music to selling value-added offerings based on what customers with unlimited access to music choose to expose themselves to.

In other words, digitization makes information much cheaper, so new businesses can (must) be built on the basis of data ubiquity.

As an example of a new type of company that leverages cheap information, take a look at DayJet. This company, which provides air taxi service in low-cost jets, was founded by Ed Iacobucci, formerly of Citrix. What's really interesting about DayJet isn't the planes it uses - new, inexpensive Eclipse jets - but the problem it had to solve to make the cost of air taxis low enough that a new market of time-pressed non-wealthy individuals would be motivated to use them. According to an interview with Iacobucci I saw, much of the time necessary to get the company up and running was spent on writing software to enable the company to provide flexible scheduling - the company dynamically adjusts the route according to who has expressed interest in flying and dynamically adjusts the price according to how willing each individual is to adjust his or her schedule; the more flexible you are in terms of when you'll be picked up and dropped off, the cheaper the flight. DayJet crunches massive amounts of information in real-time to create custom schedules for each passenger. Put another way, DayJet surrounds an inexpensive transport mode with huge amounts of data to offer bespoke travel at a modest premium to normal air travel.

Contrast DayJet's approach to standard airlines, which adjust pricing (mostly) according to how early or soon you book your flight, and provide a (mostly) standardized service based on flying very full large jets between fairly large airports. While airlines undoubtedly have sophisticated yield management systems, the systems are primarily designed to surround an expensive transport mode with off-the-rack travel. Today's airline industry was designed at a time when processing passenger information was incredibly expensive - live humans had to search schedules, write tickets, check status, etc., etc. Because of the cost of managing scheduling, only large jets with inflexible scheduling were economically feasible, because the information processing overhead could be spread across large individual transactions (flights). DayJet has designed a new business based on an underlying reality of cheap information processing, while airlines have remained mired in business operations based on obsolete data realities: they're applying analog methods in a digital world.

What does this have to do with open source? The new data ubiquity must be matched by a software infrastructure with costs appropriate to that scale. However, the assumptions underlying today's software companies are no longer appropriate for a world of data ubiquity. Today's software leaders are mired in the old assumptions of the cost of processing information. They price processing data by the gigabyte when the right scale must be the terabyte.

More piquantly, they are priced at a level appropriate to paving the cow paths of a pre-digital world. The cost is entirely appropriate if you compare it to the cost of processing data with humans, but entirely inappropriate if you recognize that enormous increases in data crunching is necessary to run businesses in the digital age. In short, the cost of software precludes building the businesses of tomorrow. I am certain that DayJet did not design their scheduling software on SAP or Oracle; in fact, I'd bet there's open source at the heart of its software system.

Your reaction is probably something like "software is a trivial part of the overall cost of providing a service, so its price is unimportant." Beyond the fact that this is a rationalization used by software companies to justify why their 95% margin products should be consumed by customers that struggle to make 10% margin, it's no longer relevant. Expensive software made sense for a transitional world moving from physical to digital - providing the means to transform paper to bits was worth (literally) a fortune. But in an all-digital world exploding in the way it uses digitized information, expensive software is not longer a conduit to lower costs, but a bottleneck to increased revenues.

Returning to Kodak, the article notes the business transformation it's attempting. It has moved into an online digital photo storage business and is pushing into high-end commercial printing. Moreover, it's launching an assault on HP with lower-cost computer printers. I can't help but think that it's missing a trick, however. It should be looking to a time when people have had digital cameras for five or ten years and have thousands - perhaps tens of thousands -- of photos. In that world, online storage won't be enough. The biggest problem in that world will be keeping track of one's vast collection - the biggest problem will be finding one photo out of a zillion. Instead of pining for ways to replace high-margin consumables, it should be looking for ways to make it easy to locate a specific "Kodak moment" in the mass of permanent digital photos. That's the new high-margin opportunity.