Archive for December, 2009

Merry Snowpocolypse

Wednesday, December 23rd, 2009

As we look forward to a potential snowpocolypse or snowmageddon here in the Twin Cities, or what I like to call as a Clevelander “rush hour flurries,” the wife and I have split last-minute holiday duties. I’ll be going from store to store completing our final gift purchases while she’ll be stocking up on food and, more importantly, beer, in case we are snowed in over the holiday weekend. We will be snuggled up playing Rock Band, enjoying the snow and not sliding down the highway in our Jeep. While we will be comfortably enjoying the weather, one man in a red suit will be out in the storm heading from rooftop to chimney to make sure all those well behaved children have a wonderful December 25th. The wife assures me that any stop at our house will not be on my behalf.

How are we to know where Santa is on his annual journey? I need to know when to put down the beer and fake plastic guitar and go to bed. The solution is NORAD. They have been helping to track old Saint Nick since 1955 and all because of an error in an advertisement. The short story is Sears published an advertisement with a promise to let kids speak to Santa by calling a specific number, but the published phone number was actually a secured military line. Oops! Rather than dash the dreams of the kids that were calling, the airmen pretended they were Santa and gave the callers info on Santa’s progress from the North Pole. The story is especially sweet because we normally only hear about incorrectly published phone numbers when they’d get someone put on the naughty list and not when it develops into something nice. I encourage everyone to read the article from CNET on how this volunteer program has blossomed into this large undertaking. It just helps to prove you never know what great things can come from seemingly small mistakes.

Lorton Data will be closed December 24 and 25 and January 1.

Happy Holidays!


Data Story Time

Thursday, December 10th, 2009

Every Wednesday night for the last fourteen weeks, I’ve had the pleasure of sitting through the academic equivalent of a giant bowl of beets. Before going after a Master’s degree in MIS, I needed to take a prerequisite course in financial accounting. The class has been fantastic for fleshing out my deep transactional understating of business. Financial accounting brings to mind old-time images of guys with adding machines and visors; the reality is this course should have been called “Storytelling from an Economic Perspective.” That’s all numbers really do, provide information to help tell a story. I think we get too caught up in presenting the numbers and forget to talk about what the numbers really mean.

While I am not going to present myself as a business intelligence expert, I did spend two years doing sales operations data analysis and translating between the sales organization, finance and IT departments to get projects completed reasonably on time. This means taking data that has no context and making it actionable. It means geeking out by using the same application I use to make grocery lists. It means being mocked by my family and friends. It means Excel.

I’d like to offer some key points about data analysis and presentation from my experiences that might be useful. As always, your mileage may vary.

1. Numbers tell a story, they aren’t the story.
Nate Silver’s Five Thirty Eight is a fantastic illustration on how to tell a story from the numbers. Nate is a stat-head, but he’s able to translate that information into a story or argument which makes sense for anyone, regardless of their background. The primary purpose of generating data in a number format is to help provide evidence for an opinion. Think back to the last time you had a writing or speech assignment in school and they talked about supporting evidence. Keep that in mind any time you are presenting information that involves numbers.

2. GED – Good Enough Data
This is a concept people have a difficult time getting their arms around, but I found it has been very important for getting things done.  Your information doesn’t have to be 100% accurate all the time.  For example, if you are trying to illustrate to a team of sales reps that gross margin is low and they need to be negotiating better, you don’t have to pull every transaction for the last six years and spend hours slicing and dicing it.  Get close to the pin to illustrate your point.  You aren’t preparing a report for the SEC, so don’t spend more time creating a report than time your audience spends using it.  It doesn’t make sense.

3. Customize for your audience
Seriously, one spreadsheet does not fit all. The chart you make for sales should be different than the one for operations or finance. It’s just like public speaking; extract the data that illustrates your point and makes sense for your audience. Kick out the extraneous stuff that you’ll want to explain, but won’t make sense for your crowd. It is just boring and will confuse your point.

4. Make the information actionable
Summary information should also contain (on another tab) the line item detail that proves your point. Do this not so you can win an argument, even though that is fun, but so that after the presentation people can take your data and do something with it. I know that sounds like crazy talk, but the point of this stuff is to reach a goal. Most meetings aren’t FYI–although one might argue that a lot of them feel that way—they are to get a team together to go somewhere. Make it easy for them.

5. Bite sized chunks
If your data fields stretch out to column AA, it might be just a hair too much. If it stretches to column AZ – delete it. Immediately. It isn’t useful to me, you, or anyone who looks at it. Information overload makes it really easy to be dismissed. If it’s dismissed due to complexity, you’ve killed the whole actionable idea. If it takes hours to figure out what you’ve done, then you are wasting people’s time.

6. Data expires quickly
Any data extract is a point in time. This is especially important to remember for transactional or customer information. Tell your audience what that point in time is and try to keep it reasonably close to the point of presentation. If it takes weeks to massage and tweak your data, it’s probably going to be too old to still be reasonably actionable.