New York Times Online: An Inside View

September 24, 2014

Check out the presentation “The Surprising Path to a Faster NYTimes.com.”

I was surprised at some of the information in the slide deck. First, I thought the New York Times was first online in the 1970s via LexisNexis.

image

This is not money. See http://bit.ly/1rus9y8

I thought that was an exclusive deal and reasonably profitable for both LexisNexis and the New York Times. When the newspaper broke off that exclusive to do its own thing, the revenue hit on the New York Times was immediate. In addition, the decision had significant cost implications for the newspaper.

The New York Times needed to hire people who allegedly create an online system. The newspaper had to license software, write code, hire consultants, maintain computers not designed to set type and organize circulation. The New York Times had to learn on the fly about converting content for online content processing. Learning that one does not know anything after thinking one knew everything is a very, very inefficient way to get into the online business. In short, the blow off of the LexisNexis deal added significant initial and then ever increasing on-going costs to the New York Times Co. I don’t think anyone at the New York Times has ever sat down to figure out the cost of that decision to become the Natty Bumpo of the newspaper publishing world.

I had heard that the newspaper raked in the 1970s seven figures a year while LexisNexis did the heavy lifting. Yep, that included figuring out how to put the newspaper content on tape into a suitable form for LexisNexis’ mainframe system. Figuring this out inside the New York Times in the early 1990s made this sound: Crackle, crackle, whoosh. That is the sound of a big company burning money not for a few months but for DECADES, folks. DECADES.

image

Photo from US Fish and Wildlife.

When the newspaper decided that it could do an online service itself and presumably make more money, the newspaper embarked on the technical path discussed in the slide deck. Few recall that the fellow who set up the journal Online worked on the online version of the newspaper. I recall speaking to that person shortly after he and the newspaper parted ways. He did not seem happy with budgets, technology, or vision. But, hey, that was decades ago.

image

How some information companies solve common problems with new tools. Image thanks to Enlgishrussia.com at http://bit.ly/1ps0MPF.

In the slide deck, we get an insider’s view of trying to deal with the problem of technical decisions made decades ago. What’s interesting is that the cost of the little adventure by the newspaper does not reflect the lost revenue from the LexisNexis exclusive. The presentation does illustrate quite effectively how effort cannot redress technical decisions made in the past.

This is an infrastructure investment problem. Unlike a physical manufacturing facility, an information centric business is difficult to re-engineer. There is the money problem. It costs a lot to rip and replace or put up a new information facility and then cut it over when it is revved and ready. But information centric businesses have another problem. Most succeed by virtue of luck. The foundation technology is woven into the success of the business, but in ways that are often non replicable.

The New York Times killed off the LexisNexis money flow. Then it had to figure out how to replicate that LexisNexis money flow and generate a bigger profit. What happened? The New York Times spent more money creating the various iterations of the Times Online, lost the LexisNexis money, and became snared in the black hole of trying to figure out how to make online information generate lots of dough. I am suggesting that the New York Times may be kidding itself with the new iteration of the Times Online service.

Does this apply to Google? Sure does. Google is built on innovations from CLEVER, voting, and the ineptitude of companies with arguably better technology. I am thinking about AltaVista.com, which contributed a core of wizards who built the original Google search system. With the infusion of inspiration from Yahoo/Overture/GoTo.com, Google became a one trick pony—ad revenue. The Google plumbing is getting old. What is presented as innovations strike me as software wrappers. Software wrappers like interfaces are cheaper and easier than tackling the core information architecture problem. The reason is that the cost and risk of going back to Square One are too much for even well funded management teams. Why does John Dvorak grouse about the flaws in Google Maps? He perceives a quality degradation. Will these be fixed? Yes, sort of. But there are quite a few problems Google has with its core search function. Forget deep changes to the core of Google search.

Many initial technology decisions are driven in response to situations that may not be fully understood at the time. The solution to each challenge is good enough. If a solution generates big revenue or saves the day, luck may have had more of a role than some want to consider. The fact that systems work is mostly a matter of timing, good fortune, and expediency. Once someone takes a look at the whole shebang as the Times Online person did, the problem becomes one that is too tough to solve even after two decades of trying.

There is a gotcha in this infrastructure investment problem. First, new technology may not solve the core issues. Software wrappers add a fix and increase the complexity of the system. Many systems are too complex to redo or get working as the owner desires. Why are some mainframe or wild and crazy client server systems still chugging away? Who has the money, time, or expertise to redo these legacy systems? Examples include air traffic control, the IRS computer system, and Google search, among others.

Second, today’s whiz kids do not want to accept the enormity of the task. Someone says, “Fix HealthCare.gov.” Okay, how is that working out? What about the HealthCare.gov fast track registration system? It sort of works, but it is an interface, not a fix for the deeper problems of the system. For example, what about the back end? Well, that’s something few want to talk about or figure out how to address. This means that some commercial and government systems will never work as users hope or expect. What about those fancy predictive analytics systems? Did these systems nail the Ebola outbreak, ISIS, or the behavior of the Australian terrorist a day or so ago? Nah. Why? These systems don’t work because those who rely on them are in the same position as the New York Times when it figured that it could do online its own way.

Third, the manager, consultants, or inside employees who point out that a problem may be non fixable are ignored. This is a fairly predictable  management reaction. Yesterday I wrote about the hatchet job done on Lucid Works in one blog and then the bandage hastily slapped on by another blog. In the tempest in a teapot, a very poignant factoid emerged. The president of Lucid Works said here:

“This sh#t is hard.”

A pithy, if not elegant or sophisticated observation.

Well, no kidding.

The New York Times development professionals have a much better understanding of the challenges the Times Online poses than senior management.

It would be an interesting exercise to figure out the total, fully loaded cost of dumping LexisNexis and doing the Frank Sinatra “My Way” thing over the last 20 years, factoring in appropriate interest rates, estimates for opportunity costs, direct and indirect costs, etc.

I think these might be quite ugly numbers. My hunch is that the numbers will be uglier going forward. Online complexity is worse than maintenance for one’s Ferrari.

Stephen E Arnold, September 24, 2014

Comments

One Response to “New York Times Online: An Inside View”

  1. New York Times: All the News that Fewer Staff Can Produce : Stephen E. Arnold @ Beyond Search on October 1st, 2014 10:42 am

    […] Cutbacks in Newsroom Staff.” I summarized the Times’s management decisions about online in this post. The Times has been floundering with new media for […]

  • Archives

  • Recent Posts

  • Meta