Link to Profile Semperoper, Dresden Sieg (auf dem Siegesäule), Berlin Brandenburg Tor, Berlin Skyline, Frankfurt am Main

Wednesday, November 01, 2006

A Question of Trust

Yes, gentle reader, it is a little after 9 a.m. and I am here typing away. What gives ... is he abusing his internet access at work? No, one of the fortunate things about working "across the border" in the Bundesland of Baden-Wurrtemburg is that the holiday-schedule is a little different from that of Hessen, so while all of my Hessian friends are toiling away, I got to sleep in until 9. What decadence.

All of which gives me the occasion to ponder the crisis of yesterday ... not much of a crisis for me, rather one for my new German masters. I have commented here in previous posts that I wonder what my colleagues do all day. I found out over the past couple of weeks as I worked my way through the trail of spreadsheets and formulae that they have amassed in the course of doing their jobs, whatever those may be.

The best way to summarize the task at hand would be to describe it as managing roughly 150 assets positioned in 30 countries. To do that, they have developed roughly 180 spreadsheets, one for each asset and then another 30 or so on top to summarize and manipulate the relevant data. Each quarter the 30 "management" spreadsheets grow by 6 as six new worksheets are added to the folders to summarize the data at that point in time and to plan forward.

My boss's boss made a joke to me, after closing the door to my office, that I must surely think they are crazy for not having a real database, which was why I was there. He was right. They are crazy ... nice, but crazy. They are, fortunately, implementing the database that I designed for them during my job interview (they gave me a "test" with 5 case-problems, one of which included how I would get my hands around the problem of managing 150 assets in 30 countries ... they must have liked the outline, because they hired me, although like so many good Germans they started building the solution before the architecht had finished the plans, something I may have to reckon with later). Until then, they must still live with their ad-hoc past.

Self-documenting Code or Spaghetti Code ... a little geek humor. When a small group of people have been working with the same collection of data for some length of time, they become so familiar with it that they all basically understand what any one of them might have done with it. In [computer] programming cirles, they would tell you that their "code" is self-documenting, as in "We hold these truths to be self-evident," anyone can figure it out sort of way. Well, people working with spreadsheets aren't always programmers, and sometimes their skill sets lack the complexity to offer up that lame excuse ... no, the average spreadsheet jockey will timidly reply with something like, "Well, that's the way we have always done it."

Which is what I faced yesterday just before I clicked the "Trace Precedents" button on the Excel Auditing toolbar to look into the convoluted formula adding up results for several assets in the same region. Oops, one of the assets, generating roughly EUR 92 million in EBIT (earnings before income taxes) was counted at two levels, the first being on a regional basis and then again in the global totals on its own. What that meant for this particular "Management Tool" was that the amount upon which the bonus pool would be calculated was overstated by EUR 92 million.

Not a problem for me. Just admit the error, fix it, and move on. My bonus for the year was going to be paltry in any case (it is pro-rated). My boss didn't share my view. His first take was, "Does it really need to be fixed ... I mean, if the number is both in the top and bottom of the fraction, doesn't it cancel itself out?" I'm sure he got a 1 note in his O-level Maths roughly 18 years ago, but it wasn't evident today ... adding the same number to the numerator and divisor of a fraction is not the same as multiplying or dividing the same for the purposes of factoring out a quantity. His "older" boss, who is one year my junior but looks ten years older probably because he is responsible for riding herd and cleaning up after these "kids" immediately saw the flaw with his logic and sent him to his office to workout a quick analysis for himself.

In the meantime, he looked across the table at me and asked, "What would you do?" We knew where my boss stood ... hide. But the boss's boss didn't get to where he is by taking the easy way out. This was an obvious test of my character.

"It's easy for me to come in and say, 'Let's own up to the problem and fix it,' since I have no connection to the history. It's a no-win for us ... if we admit we made a mistake we lose a little or even a lot of credibility. If, on the other hand, we try to cover out tracks and later get caught, we then have no credibility. Besides, anything we do will have to be documented with a memo to file to explain why we deviated from our own formulae. Better to clean it up now and be done with it."

By this time my boss has come back and says, "Yes, it does make a big difference, but we already communicated this amount as a part of the [performance] targets for this year to several dozen managers ... if we fix the error, we will have to re-communicate the targets to them, and they will wonder why. Wouldn't it be easier to take this through to the end of the year and fix it then?"

Yeah, and then explain why next year's targets suddenly got so much easier while explaining to the files and future auditors why we paid X-million more in bonuses than the accounting systems said we should. Hmmm.

The Boss's boss went with what we discussed. Nevertheless, I still have to look to the guy to whom I report ... another warning light goes off in my head.


Blogger christina said...

Reading that gave me a giant stomach ache. I don't know how you do it.

5:43 PM, November 01, 2006  

Post a Comment

Links to this post:

Create a Link

<< Home