Bloomberg News reports:
Times Co.’s effort to turn online readers into paying subscribers is being watched by other newspaper companies struggling with a decline in traditional print advertising. The New York-based company is spending $40 million to $50 million on the project and has said it plans to debut it by March.
Picked up and quoted as:
my biggest question right now is how the NY Times spent a reported $40-50 million writing the code (Bloomberg; other sources are consistent). Google was financed with $25 million. The New York Times already had a credit card processing system for selling home delivery. It already had a database management system for keeping track of Web site registrants. What did they spend the $40-50 million on? A monster database server to keep track of which readers downloaded how many articles? They should already have been tracking some of that for ad targeting. In any case, a rack of database servers shouldn’t cost $40 million.
"The project" is unlikely to equal "writing the code", even on a very expansive definition of "writing the code" to include all the project spec and design work. I would very strongly guess that a significant part of that $40m (if it's a hard figure; as far as I can tell the other sources are consistent because they're quoting Bloomberg, which doesn't attribute the number), was market research.
Market research can be very, very expensive, although as with education it's not so expensive when you compare it to the alternative of remaining ignorant. The lion's share of that budget would most likely have been spent on answering the question "Should we have a paywall?". Then "What kind of paywall?". I am not sure I agree with the answer they came up with, but for a largeish company making an unbelievably important decision about its flagship product, large spends on research would seem appropriate.
I am still having a hard time getting to $40m though. A very, very big market research project would be one with a budget of $5m; this would be an account that the entire industry competed for. You could then probably spend quite a bit more on strategy consulting - it's not difficult to run up a bill of $1m a month with McKinsey, and if we say they did twice that for six months, round the whole lot up ... I can see $20m for groundwork. Another c$5-10m for internal cost allocation of senior management time still leaves the actual coding to cost $10m, which is still about an order of magnitude more than it could reasonably cost to do the job. Update Alex thinks could indeed cost that order of magnitude to do the job, making the good point that Greenspun's cheep'n'cheerful paywall for MIT didn't exactly have to handle the volume of the New York Times (and nor did Google back in the days when it was doing things on a budget of $25m, which these days of course it is not). Tying in to the nuclear post above, making things bigger often makes them more complicated too.
Don't forget legal on top of that, adding probably, what? Another $1m tops, and that would be a very high end estimate.
ReplyDeleteI was with you until the last sentence. My wife and I both went into IT after completing our English degrees; I got out, she didn't. So I can assure you that "writing the code" is a tiny element of taking a system from an identifiable set of requirements to, well, a system - and the people who write the code are a tiny (and undervalued) layer of the system development operation.
ReplyDeleteMost software projects fail. I'd bet they built and scrapped two complete implementations before launching this one.
ReplyDeleteThis is basically the same cost problem as government IT procurement projects. If you don't have a very clear idea what you want and what it should cost, the consultants will simply take you for all the money you have.
ReplyDeleteThe more important the project is to senior management, the easier it is to run up the costs.
Well some of the costs could just be because their IT department is really crap. All the evidence points to it. Also licensing costs can prove surprisingly expensive, particularly if you're using off the shelf stuff (Opensource stuff has a higher initial cost, but the curve then slopes to the horizontal as you scale. MS stuff doesn't. Many a company has suddenly found themselves in trouble discovering that fact).
ReplyDeleteI don't agree with Murdoch on a lot of things, but I think he's right about the NYT. Their journalists are some of the most long-winded, pedantic bastards in the business. Whatever else, at least British newspapers know how to get to the point. I really can't imagine anyone under the age of about 45 buying the NYT.
I really can't imagine anyone under the age of about 45 buying the NYT.
ReplyDeleteThe NYT is to this day very popular with students at elite American colleges, who get it for the crossword if nothing else.
...
Also, if they actually spent that much money on McKinsey, wouldn't they get a better strategy than whatever thing they're doing now?
Many a company has suddenly found themselves in trouble discovering that fact
When it comes to MS Office, though, that's just a tax everybody has to pay.
When it comes to MS Office, though, that's just a tax everybody has to pay.
ReplyDeleteI'm talking about server side infrastructure. .NET, Windows Server, MSSQL, etc. The area where Linux is the competitor (or rather the area where Windows is trying to compete, to be rather more accurate). Combination of license costs and the fact that MS stuff just doesn't scale as well.
Having just read about what the NYT is trying to achieve, the reason its expensive is because its incredibly complicated. They have different rules depending upon where you're coming from, monthly limits, social media, etc, etc. Looks like one of the classic bondongles where nobody really wanted to make the hard decision, and so they've got a poorly defined compromise.
ReplyDeleteComplexity costs baby. I'm guessing $40 mil will look cheap at the end of this.
Complexity costs baby. I'm guessing $40 mil will look cheap at the end of this.
ReplyDeleteProbably. I thought they should have designed the thing based on levels of meta-ness. So the really meta stuff, like blogs about other blogs, or regular blog posts, and stuff which basically draws the visitor to NYT website but can't be monetized, would be free, but the non-meta like columns and news and so on would be paywalled.
The problem is that the actual, paper NYT has been for so long been cross-subsidized by its lifestyle section that it's difficult to see it as a realistic business proposition in terms of just news.
I had something to say about this but it grew.
ReplyDelete(Also, Greenspun's project had to handle two different billing rules - are you a subscriber? are you an affiliate? Nothing like "are you a non-logged in member with 15 or more free articles remaining who is following a Twitter link to one of our news analysis pieces?")
ReplyDeleteYork and D-Squared, I posted links to your posts over in Phil's site; I apologize if you are infested with problematic people.
ReplyDeleteI dunno, I've worked for a company that dealt with HLRs for a couple of UK MoTelCos and I just can't imagine NYT needing a couple of million quid to build what's a fairly well-understood set of technologies. Once you decide you're having a paywall, then the matter of the billing rules is a triviality compared to actually setting up all the stuff the ranter mentions in his post. I think DD is more likely right that it's been spent on strategy, research, and more than likely "flowers and chocolate" at the launch party.
ReplyDeleteIt's well understood, but it's very proprietary...I mentioned the HLR as a scale-marker. As you know Bob, there's a lot more complexity in the BSS OSS hooked up to it.
ReplyDelete