The Danger of Illusory IT Excellence

I remember rolling my eyes in August, reading a piece from CIO Insight ("Overstock.com: Feeling the Need for Speed") on how Overstock.com's CIO Shawn Schwegman was running the rootin'est-tootin'est IT shop around. The reason was that it seemed like something out of a time capsule to the late '90's world of Internet time - in all the bad ways - big egos and an apparent dismisiveness of most of what we know about managing business-critical, enterprise-grade development and implementation.

First there was the Schwegman's chest-pounding about implementing a Teradata system faster than Teradata thought possible. How did he do it? Working harder, not necessarily smarter, seems to be Schwegman's way:

How did they do it? "We are working our asses off, to be blunt," he said.  

 

Schwegman's 95-person IT group represents about one out of our every four people at the corporate headquarters building and has almost tripled in size since October.  

 

To make the Teradata deployment happen that quickly, Schwegman's group worked an average of 65 hours/week. "One guy put in 102 hours last week," he said.  

 

Beyond salaried IT staff, Schwegman also "beefed up consulting" and brought in a lot of extra temporary help, he said.

One hundred and two hours in one week - 65 hours/week on average? Even for a 70 day implementation cycle, this will burn people out pretty fast. Of course, it sounds like it did - it looks like Teradata's revised estimate was also 70 days. This required some slightly rephrasing from the uber-CIO. The article continues on:

 

Teradata VP David Scott said that Teradata did, indeed tell Schwegman that the project he was undertaking would probably take nine to 12 months to complete. But he also said it was a rough estimate for a set of services and capabilities much more extensive than the set Overstock initially deployed.  

 

After the project scope was redefined, Teradata's estimate of the time needed for rollout was 70 days, Scott says.  

 

Schwegman, asked about the apparent contradiction between his remarkable time-and-scope description and Teradata's more mundane estimate, said the truth was "a margin between the two stories."  

 

"It would have been a six-month project. If [Teradata] went anywhere else and did it, it would have taken six months," he said. The vendor helped accelerate the project because it "poured in a ton of Teradata resources."
 

Here's where it got completely pre-dot-com-bust for me (in addition to later comments in the article about having "snowboarding hours" after the peak holiday season):

 

"It's very clear when you meet Shawn," Scott said. "He has a cot in his office and a strong pot of coffee. Once we all agreed to the statement of work, their focus was incredible. I'd give them a 9.9—if not a 10—on a scale of being focused."  

 

The 31-year-old Schwegman freely admits to being a workaholic who puts in 70 hours in an average week and is proud of the futon in his office.
 

Now, I'm sure at this point, many executive were contemplating what slackers they had working in IT.  While they might, or they might not, this is not the example to use for the ideal CIO. Although even Gartner was smitten with Schwegman:

 

Workaholics or not, Overstock's people were focused, and they went out of their way to understand and internalize what the business, marketing and operations people needed, according to Bill Gassman, a research analyst at Gartner, who spent a half-day with Overstock's marketing department and was surprised at how well it used technology.  

 

"If we were to draw a bell curve of maturity in IT processes, they're definitely on the high end of the curve," Gassman said. "That takes an IT department that understands that business people are their customer."  

 

"I still use [Overstock] as a benchmark. The types of best practices the company had are really rare," Gassman said. "They ask to look at reports of what people are doing on the site all day. Then, based on that, they may change the site several times a day and keep an eye on the effects of changing that home page. There's a lot of cooperation between the developers and the marketing people. At other companies, making a change like that would take weeks or months. These guys do it multiple times a day."
 

While that approach might work for things like tweaking the home page or tuning the merchandising presentation, this sort of consistent adjustment of fire is usually a great way to sabotage real enterprise systems development.

And this time was no exception. Just a few days after the marvelous puff piece, the problems started to show up. Actually, even the CIO Insight piece points out that Overstock had to delay some earnings numbers because of their ERP implementation. It turns out Schwegman's team was having some problems even beyond that. As it happens, Overstock made true on their name as their CEO announced they are packed to the gills with merchandise and probably won't hit their sales numbers because (drum roll please) the computer upgrades Schwegman's group was performing. (Information Week, among others has a general piece on the subject.) Way to go, champ! Before this, there was an email sent out by Schwegman to real customers (not the so-called internal customers) actually giving a mea culpa for some of the IT problems. Some people have suggested the real problem with this item may have been that Schwegman was too candid. I disagree, but there's no point going into it here (but feel free to read Bob Lewis' comments about it) - suffice it to say that if you're doing your job right, too much transparency shouldn't be an issue.

Of course, possibly Schwegman may have had a sneaking suspicion that although he undoubtedly looked busy he might not be helping the company - explaining his liquidation of nearly his entire holdings of Overstock.com equity in June and September - selling 6,700 shares, leaving him with less than 1,000 shares as of the most recent SEC reports [Yahoo].

So where does this end up? Well, as of last week, it looks like Schwegman is on a leave of absence [Mercury News], barely two months after his CEO called him "The best CIO in America"

What's the take away? I'd encourage my readers to think about this:

How are you, as an executive manager, evaluating the quality of your CIO or its IT function? I have many times had conversations with people who equate hard work (as evidenced, unfortunately, by long hours) as a critical factor ("He's committed," "He works so hard"). My response is invariably the same: the "A for effort" doesn't apply in business - is he doing what he is supposed to be doing and doing it well? These are independent questions.

If you, as an executive, don't have the knowledge to fully assess this, you need someone on your board, or as an outside consultant, who can be honest with you. I recently joined a board for this just reason - neither CEO nor any board member at the time felt fully competent to assess the quality of their CTO and the development function in light of recent comments by a dissident shareholder. The CEO was pretty confident in the hiring process that brought this individual to the organization, but understandably wanted further validation of the individual. (Because he well knew that there was a non-zero chance that the company had been snowed during the hiring process.)

As it happens, the CTO in this case really is quite solid, having both a good vision and strong organizational and execution skills as well as a wealth of relevant experience and managerial maturity. That gave the board and the CEO the confidence to have a more frank conversation about the issue with the shareholders, as well as better information against which to develop related forecasts and plans.