It’s time for a few I-told-you-so’s (ITYS), some industry commentary, public policy ramblings, and just enough connection to IT leadership to justify the rest.
ITYS #1: In his Fatal Exception blog, InfoWorld’s Neil McAllister asks, “Is the SaaS experiment finally over?” (7/1/2010). Citing Gartner, which elsewhere continues to cheerlead Software as a Service, he points out that SaaS adoption continues to be both limited and disappointing.
No surprise here. See “Trend, fad, or blah blah blah?” (Keep the Joint Running, 2/13/2006) and a number of other articles on the subject in the KJR Archives (search on “SaaS”).
ITYS #2: If you think you understand how and why our economy fell apart and haven’t read The Big Short: Inside the Doomsday Machine (Michael Lewis (no relation), W. W. Norton Y& Company, Inc. 2010), you’re wrong.
The book is that important. Even more amazing, Lewis is such a good and entertaining writer that he’s able to hold your attention while explaining what Credit Default Swaps (CDSs), Collateralized Debt Obligations (CDOs), and the mysteriously un-acronymized “tranches” are and why they matter.
Best of all, it turns out I got two points right (“Lehman and mean,” KJR 9/22/2008). It was entirely by accident, and in the greater scheme of things one of them wasn’t all that important, but I’ll take what I can get.
The unimportant one: Circular references really were involved. Astonishingly, this isn’t satire: The geniuses running the financial sector of our economy actually created securities whose underlying portfolios included each other. So if security A increased in value, security B, which included A, increased; security C, which included B, increased, and … this is the wonderful part … security A, which included security C, increased again, because of no good reason other than the circular reference.
The one that mattered: It turns out the same geniuses really did assume that if you aggregate a bunch of loans, each with a high risk of default (triple-Bs), that together they formed a AAA-rated security. Except they probably were assuming no such thing. They knew better, and sold the mis-rated securities anyway.
Does The Big Short have any relevance to you in running IT? I don’t care — you need to read it regardless — although there’s some value to be had in recognizing the role played by Wall Street’s rating agencies and how they came to play it.
S&P and Moody’s both considered the companies they rated to be their customers. So instead of being watchdogs, they trusted the big investment firms’ AAA risk assessments for fear of losing their rating business to competitors.
It also turns out that S&P and Moody’s pay their analysts a tiny fraction of what the Wall Street investment firms pay their analysts. One consequence: The people constructing what turned out to be fraudulently rated securities were much more skilled at hiding the fraud than the ratings agencies’ analysts would have been at detecting it, even if they’d had management support for trying to do so.
Lesson for you: IT has some watchdog roles to play in the enterprise. If you have “internal customers,” and if you underpay your staff (or overstretch them — another version of the same thing) you’ll fail in them.
ITYS #3 (sort of): Andy Grove, Intel’s former CEO and chairman of the board, just published the most important article of the decade (thanks to Dan Jacobs for bringing it to my attention). Titled “How America Can Create Jobs,” (Bloomberg/Business Week, 7/1/2010), it describes with compelling, detailed logic how America’s business leaders have systematically separated economic growth from domestic job creation (the sort of ITYS: “The two economies,” KJR, 4/19/2004).
Grove points out what’s changed: The “scaling up” phase — turning an innovation into a manufacturing success — has been sent offshore. The economic consequences: (1) For every job Silicon Valley creates onshore it creates ten offshore; and (2) sending manufacturing offshore has led to sending a lot of innovation offshore with it.
Our economic recovery isn’t, as it turns out, jobless. It’s creating jobs. But they aren’t here.
As an IT leader, you do have a role to play in this … a minor one, but a role nonetheless, and it might surprise you:
Switch to Agile methodologies. More specifically, switch to methodologies that require a high degree of direct interaction between programmers and end-users.
Waterfall methodologies are easily supported by offshore sourcing systems. With high levels of programmer/end-user interaction, Agile works much better when everyone is local.
That keeps jobs here. This being the July 4th weekend, consider it your patriotic duty.
ITSY#3 — Selling agile methodologies to business unit managers is very difficult. They would prefer that you “define the requirements,” with a minimum of disruption, then go away and develop the solution, preferably while the requirements are still reasonably current. In one of your books you described how to actually succeed at this (embed the developers with the users), but that’s heretical, at best, and runs the risk that your developers will, gasp, align with your users.
Involving the users in defining the requirements is costly, time-consuming, and fraught with danger. Success requires that they actually understand the business, rather than just a segment of the process. No one will have bothered to teach them (or allow them to learn) the business, so you get trivial requirements, e.g., “can you change the color of this tag?”
Agile development is terribly difficult to “schedule.” (I’m convinced that the worst thing that ever happened to software development is Microsoft Project, which casts a fantasy schedule, filled with uncertainty, into concrete.) Agile development takes … well, as long as it takes, plus some more time to do it better, now that the developers actually understand what’s really needed. Pretty difficult to explain to “the boss” why you can’t definitely say when the effort will be complete, which means you also can’t say with any certainty when that development team will be available for one of the hundreds (or thousands) of other backlogged projects.
[Worse, if IT doesn’t respond, the business unit may develop a solution in Sharepoint, and then expect you to support it!]
Compared to the “fixed cost” of off-shoring the development, and lacking any understanding of the subsequent cost of actually making the solution work, it’s pretty difficult to make the business case. And any attempt to disparage the offshore developers immediately leads to the conclusion that your staff are incompetent to define the requirements, creating an immediate opportunity for a consulting company to engage (at a cost that far exceeds what the agile development would have cost).
Bottomline: we need a heckuva lot more senior executives who have actually written a couple of lines of code, maybe even had to debug it, and then had to deal with the user astonishment at just how badly the code didn’t even address the need.
So … how do you make this clear to senior executives (presuming that they didn’t already understand and hire you because of that)?
Rog
Rog
Second note for IT under ITYS#2: Pay to get good people!!!!
Thought I’d share this one – an article on Apple, and the ten things that make it great, published in FastCompany
http://www.fastcompany.com/magazine/147/apple-nation.html?page=0%2C0&partner=homepage_newsletter
Steve Jobs is quoted, quoting Henry Ford:
“If I’d have asked customers what they wanted, they would have told me, ‘A faster horse!’ ”
To put it another way, if you only respond to what your customers ask for, you’ll end up being behind their needs curve. Jobs says customer needs provide inspiration for Apple, but not guidance.
At some point, if you place yourself in the mode only fulfilling what people ask for, you’re in the commodities business, and that means, you’ll be in a race to the lowest margin.
On the financial crisis, there is an even better one. In the book Too Big to Fail, Andrew Sorkin describes how Lehman was booking a profit by planning on buying their own debt back. The theory was they could issue debt (bonds) at one price and as their business failed, they could buy them back at a discount. Instant profit! So the lower their bonds went, the better they did. Too bad that theory didn’t end up actually working.
On the Agile comment above. As someone who is leading a very large Agile project, there are ways to sell it to Senior Management. Actually, Mr. Lewis provided some consulting services that helped me learn this. Estimating/creating specs for waterfall has a large price tag that can be saved through Agile. With Agile, you can estimate a project based on story points and the team velocity. We spent maybe 5% of the time it would have taken with waterfall to estimate the cost of the project and so far the numbers have held up. I can’t go into all the detail in a blog comment, but it can be done and it’s easier than it seems.
Another lesson from The Big Short, is the danger of incrementalism or applying models you don’t understand (which reminds me of an old joke about Statisticians and Economists). The models used by the rating agencies were based on earlier models for consumer credit debt that; step-by-step assumptions were relaxed (i.e., substituting CDOs for the prime mortgage portion of the mix) until there was nothing but crap in the mix.
Patriotic Duty
We in IT have a duty to our company first. If our company succeeds it will help the country. Making a change to IT policy to keep jobs here, is not a good justification as it may hurt the company instead of helping it. There are many justifications for moving from waterfall to agile (most of which I learned from you), creating jobs is a side benefit. If I put the creation of jobs as a justification for an IT policy, I would expect to be fired- or at least laughed out of the management meeting.
I love your guidance on using agile methods. After using them for a while, you begin to realize that anything different is just nonsense.
With respect to the piece from Andy Grove. That article is completely spot on. I realized during the last downturn (when I worked in IT supporting manufacturing) that in present day America, American companies have absolutely ZERO loyalty to the country they are from. We were discussing job cuts in an economic downturn and how those cuts would just hurt the economy more. The fact was it didn’t matter in the current quarter, so jobs were cut (and yes the company in question really didn’t need to cut the jobs). The great irony of Grove’s piece is that in this time where it is very in vogue to be a strict constitutionalist, the solution he recommends is to utilize the tariff to restore American manufacturing. The tariff being the primary method of government funding laid out by the constitution. Perhaps the founding fathers were right on that one.