Idiots who can't make up their minds
I want you to try to imagine having your heart ripped out through your mouth. Not a pleasant thought, and definitely not a pleasant experience. Trust me, I know. Over the last few days, the idiots in the system have convinced some people that the sky is falling in, and that this would be an appropriate time to panic. A Japanese version of Chicken Little - Tanuki Little?
Now, the data doesn't support the notion that the firmament is collapsing - in fact, it's probably telling us that it looks like rain. Maybe even snow. But it's equally as evident that there will be a rainbow in the very near future, which the peasants will tell everyone is a divine sign of the god's favor. We all know that there are simpler explanations than supernatural intervention, but that's not as emotionally powerful for the milling masses.
Which all brings me to remind you all of the 4 S Principle for Data:
Store - Start by making sure you store your data to an accessible database that's structured, and architected to ensure multiple views into it. That means no more Excel spreadsheets, no more Word documents, no more dumb ETL operations (unless they produce XML that's stored in your db), no more personal Access databases.
Share - Next, demand and enforce complete transparency into the data. That means forcing your team into regular reporting to a wide audience, and forcing your audience to pay attention via visually attractive interfaces like Dashboards and Table Mats. While most people think that this should be the last step, my advice is to do it early and do it often. Everyone begins to expect certain behaviors and almost always wants to offer advice on how you could do it better.
See - Use the data to generate insights that take the form of actionable recommendations for your own team and your clients. This is the equivalent of answering the "so what" question. Remember that not everyone has the same degree of familarity with the data sets, and often don't see the the relevance of what you're trying to present. The answer is to make it clear that the purpose of the exercise is to power the kaizen process, and that there are obvious returns on the time, money, and energy you're spending on it.
Simplify - the problem with most data geeks is that they expect that everyone is as interested as they are in sucking on the data firehose. Danger, Will Robinson! Almost never true, and always dangerous. So make sure that you're presenting the data in ways that make visual sense and that are painfully obvious to even the most "gut-feel" observer. That means making the metrics construct really simple - to the point where you're asking your boss to look at just 3 or 4 things, rather than multi-dimensional graphs in 8 point font with thousands of data points.
My preferred position is to design 4 predictive constructs that can be expressed as vectors - that means that they have a value (position) and a direction (velocity). I call these constructs "diagnostics", like you get from your doctor. And I feel that these should be a Sales Health diagnostic, a Sales Force Health diagnostic, a Product diagnostic, and an Operations diagnostic. And when you roll these up, you should get a quick health check on how you're performing against strategy. Each diagnostic probably has up to 10 key variables that roll up into a number that has relevance for senior management.
Because I think that executives don't need to know what we screwed up; they need to know what's going wrong!
Or am I an idiot?
Now, the data doesn't support the notion that the firmament is collapsing - in fact, it's probably telling us that it looks like rain. Maybe even snow. But it's equally as evident that there will be a rainbow in the very near future, which the peasants will tell everyone is a divine sign of the god's favor. We all know that there are simpler explanations than supernatural intervention, but that's not as emotionally powerful for the milling masses.
Which all brings me to remind you all of the 4 S Principle for Data:
Store - Start by making sure you store your data to an accessible database that's structured, and architected to ensure multiple views into it. That means no more Excel spreadsheets, no more Word documents, no more dumb ETL operations (unless they produce XML that's stored in your db), no more personal Access databases.
Share - Next, demand and enforce complete transparency into the data. That means forcing your team into regular reporting to a wide audience, and forcing your audience to pay attention via visually attractive interfaces like Dashboards and Table Mats. While most people think that this should be the last step, my advice is to do it early and do it often. Everyone begins to expect certain behaviors and almost always wants to offer advice on how you could do it better.
See - Use the data to generate insights that take the form of actionable recommendations for your own team and your clients. This is the equivalent of answering the "so what" question. Remember that not everyone has the same degree of familarity with the data sets, and often don't see the the relevance of what you're trying to present. The answer is to make it clear that the purpose of the exercise is to power the kaizen process, and that there are obvious returns on the time, money, and energy you're spending on it.
Simplify - the problem with most data geeks is that they expect that everyone is as interested as they are in sucking on the data firehose. Danger, Will Robinson! Almost never true, and always dangerous. So make sure that you're presenting the data in ways that make visual sense and that are painfully obvious to even the most "gut-feel" observer. That means making the metrics construct really simple - to the point where you're asking your boss to look at just 3 or 4 things, rather than multi-dimensional graphs in 8 point font with thousands of data points.
My preferred position is to design 4 predictive constructs that can be expressed as vectors - that means that they have a value (position) and a direction (velocity). I call these constructs "diagnostics", like you get from your doctor. And I feel that these should be a Sales Health diagnostic, a Sales Force Health diagnostic, a Product diagnostic, and an Operations diagnostic. And when you roll these up, you should get a quick health check on how you're performing against strategy. Each diagnostic probably has up to 10 key variables that roll up into a number that has relevance for senior management.
Because I think that executives don't need to know what we screwed up; they need to know what's going wrong!
Or am I an idiot?
posted by Terry at
6:31 am
0 Comments:
Post a Comment
Subscribe to Post Comments [Atom]
<< Home