Why Mouse Trap sucked and so does your data.
The board game Mouse Trap was first created in 1963. The goal of the game was to catch a mouse by building a mouse trap designed in a way that would make Rube Goldberg proud (well aside from the fact that the company refused to pay him royalties, but we can leave that bit to the lawyers).
Despite the name of the toy company, Mouse Trap was anything but ideal as games go. Quite frankly it sucked! Understanding why is an exercise in model building (model — not mouse trap). Along the way, you may also learn why you often feel the same way about your data infrastructure.
1) Good models simplify things and games should do the same. Think about Monopoly, Risk, The Game of Life — all of these were simple models that allowed you to focus on the strategies of how to win. Mouse Trap was a convoluted (Rube Goldberg) model. There was nothing simple about it, and frankly, I don’t even recall what the strategy of the game was… it was more than just catching the mouse, I promise you.
Now think about your data infrastructure. Does the phrase ‘simplicity at its best’ come to mind or does it take 6 months to teach new employees how to use it? An hour to discuss the most recent data issue? Another three to determine the data lineage of the poorly named metrics on that table you found the other day?
2) Good models work faster than the things they are modeling. They speed up the cycles. Imagine if it took a life time to play The Game of Life. Monopoly was an incredibly fast way to build a real estate empire and lots of people still complain the game takes too long. Real mouse traps are wickedly fast. Mouse Trap the game, was not.
Back to your data infrastructure — my bet is it takes longer for the data to arrive in your warehouse than it did for your customers to create that data in the first place. We live in a day and age when real time should be a reality — and yet 24 hour lags are more the rule (and some reading this only wish they had their data that fast).
3) The Mouse Trap was made from dozens of disparate pieces. And while you can occasionally revel in the complexity of models like this — they only lead to disappointment. Good models appear to be clean and have continuity. Nothing like getting to the end only to realize that the dog ate one of the parts!
Disparate data has become the rule. Too much is piecemeal in the world of data collection. And while I have often reveled in the complexity of corporate data infrastructures — I have also been forced to deal with lots of disappointment.
This is not to say that data infrastructures can’t bring together disparate data, they just shouldn’t look like it. So hide the boot. Hide the hand crank. And for Pete sake, don’t use marbles!
4) Great games (and models) have clear and simple rules so you can focus on the strategy. This is why they are fun! Can anyone name one of the rules of Mouse Trap? Did anyone bother to read them? My bet is that if you ever ventured to play this game, you quickly stopped following the rules and just tried to get the trap to work.
Have you ever referred to your data infrastructure as fun? Does that question seem silly? Why? Shouldn’t data analysis be fun? It is full of Sherlock Holmes style investigation, story telling, and loads of surprises… well unless you spend all your time just trying to figure out how to get the data. Or comprehend what it all supposed to mean.
5) And finally, the damn thing never worked! Models should work. And there is no excuse for a game. Mouse Trap was doomed to fail. The complexity, the speed, all those pieces, and I am not even considering the rules here. I built that mouse trap dozens of times and every time something got stuck, was out of position, or just plain missing.
How often is your warehouse down? How often are reports ignored because the data isn’t right? How often often do you use the word ‘broken’?
Mouse Trap wasn’t a good game. It wasn’t a good model. If your data infrastructure is too complicated, too slow, built from too many pieces, has convoluted rules, or just never seems to work — it is going to suck too.
Originally published at www.linkedin.com.
