Understand the Shift From the Old Probability to the New Entropy.
John Barnes 12/19/2016 23 comments
Managers need to understand the conceptual shifts away from traditional statistics, such as the change of emphasis from probability to entropy.
Nonlinear Problems: When Brute Force Is Best
John Barnes 3/14/2013 6 comments
Analysts would do well to get out of the rut of using linear regressions by default.
Analyze for Accuracy or Precision, or Vice Versa
John Barnes 2/18/2013 24 comments
Sometimes your results require accuracy and sometimes precision. Knowing the difference matters.
Rock, Paper, Scissors – Trend!
John Barnes 2/4/2013 11 comments
Rule-based behavior models offer a good alternative to guesswork and folk wisdom.
Know Your Perimeters, Parameters & Paradigms
John Barnes 1/14/2013 17 comments
Keeping these three words, often jumbled in business discourse, separate and precise can help you be a better decision maker.
Don't You Regress With Your Regressions!
John Barnes 12/3/2012 11 comments
Always keep in mind that regression to the means happens — and watch those outliers.
The Landslide That Wasn't: A Lesson in Distributions
John Barnes 11/16/2012 17 comments
Every day, thousands of managers confuse numbers with distributions, as did many 2012 presidential pundits.
Take Away the Mystery of Regression vs. ANOVA
John Barnes 10/23/2012 5 comments
Knowing when to use a regression and when to use analysis of variance needn't be a difficult to determine.
When You Think of Apple, Think of Trees & Pies
John Barnes 10/8/2012 4 comments
Tree-and-pie diagrams are about seeing a choice and its implications clearly.
Decision to Make in an Uncertain Future? Try the Payout Matrix
John Barnes 9/24/2012 5 comments
A fundamental aspect of game theory, the payout matrix is a powerful tool for insight into whatever you're up against.
Keeping Your Numbers Real
John Barnes 9/10/2012 12 comments
Four cautionary tales and seven rules to keep you out of trouble.
It Takes a Dummy to Get Context Right
John Barnes 8/29/2012 9 comments
Significant codes may be the more common method, but dummy variables might really help deliver the context you need.
Plotting a Presidential Victory
John Barnes 8/13/2012 22 comments
Not who won, or who will win, but how he can win -- the third use for metric graphics.
When Being a Tedious, Numeric Nerd Is a Good Thing
John Barnes 7/30/2012 37 comments
Why analysts must always make sure managers understand probabilities, especially if doing so is difficult.
The Correlation Matrix: Simple Tool, Powerful Insights & Clear Priorities
John Barnes 7/16/2012 18 comments
The correlation matrix deserves more than just a quick glance for the insight it can provide.
Turning Numbers Into Narratives as Easy as 1, 2, 3
John Barnes 7/5/2012 31 comments
But watch that the automated reporting doesn't mug you.
If Correlation Doesn't Equal Causality, What Else Might it Equal?
John Barnes 6/18/2012 16 comments
If it's not causality, consider four other things it might be.
Why 100 Can Make for a Messy Metric
John Barnes 6/4/2012 37 comments
Using 100 as a metric can sometimes confuse more than clarify.
When a Little Confidence Is All You Need
John Barnes 5/8/2012 17 comments
What you've learned about the term 'statistical significant' in statistics classes doesn't necessarily apply in the world of marketing analytics.
How a 'Calibrating Urn' Can Help Set Better Targets
John Barnes 3/23/2012 14 comments
Analyzing expected value using this technique can help you make informed estimates.
Sometimes You Can Learn From a Distraction
John Barnes 3/8/2012 16 comments
Analyzing distractors in your training questionnaires can help point out weaknesses for improvement.
Understanding Comes With Normalized Medians
John Barnes 2/23/2012 19 comments
How a normalized median graph can help get your point across.
What Car Dashboards Teach About Choosing Metrics
John Barnes 2/6/2012 15 comments
Here's a primer on domain, range, and responsiveness.
Making Heads or Tails of a Highly Skewed Distribution
John Barnes 1/25/2012 15 comments
Consider these nine factors when working with disruptions.
Behind the Powerful 80-20 Rule, an Even More Powerful Statistical Trick
John Barnes 1/16/2012 22 comments
Don't stop at the 80:20 rule. Use these analytics techniques to get even more out of your stats.
Getting Your Feet Wet With Hierarchical Clustering
John Barnes 1/3/2012 16 comments
Here's a guide to what to look for as you take on clustering in your next exploratory data analysis.
Coding: The Bottleneck on the Road to True Data-Driven Management
John Barnes 12/13/2011 15 comments
If we can solve the coding problem, we can move the whole analytics field forward.
Viral Vampires: How to Measure a Marketing Hashtag's Potency
John Barnes 11/18/2011 13 comments
The epidemiological equation can help marketers figure out which hashtags offer their best chance for spreading the word.
How to Measure Kindliness Like a Lazy Professor
John Barnes 11/7/2011 30 comments
Here's a measurement technique from academia you can apply to defining hiring criteria.
Mensurability: Concept for the Toolbox & Analytics-Readiness Test
John Barnes 10/25/2011 15 comments
Why the best metrics derive from quantities that are measurable and mensurable.
Why Factor Analysis Doesn't Have to Be a Disgrace
John Barnes 10/10/2011 12 comments
Though often maligned, factor analysis actually happens to be one of the most useful exploratory data analysis tools out there.
For Best Analytics Work, See What's Not There
John Barnes 9/26/2011 27 comments
Don't fall into the trap of analyzing the wrong population.
Reporting Epidemiological Equation in a Useful Way
John Barnes 9/12/2011 9 comments
The epidemiological equation can provide for powerful social media analytics, but sometimes it requires translation for those who aren't mathematically inclined.
Everything I Need to Know About Bad Metrics I Learned in School
John Barnes 8/30/2011 26 comments
Taking a lesson from the American public school, here are seven things never to do with your metrics.
Survey Smarts: JCPenney Asks a Great Question
John Barnes 8/15/2011 26 comments
How adding this one simple question to your survey will give you a way to distinguish real customers providing honest feedback for high-quality analysis from the freebie seekers.
What a Difference a K (& a J) Can Make in Social Media Analytics
John Barnes 8/5/2011 3 comments
Here's a look at how an equation typically used in epidemiology can prove invaluable as a social media metric, too.
Finessing Social Hiring With Creative Metrics
John Barnes 7/28/2011 12 comments
Here are some quick and relatively easy analytics to apply when assessing those hot social media experts upper management is so gung-ho to hire.
Let's Make Friends With the Friendly Elephant, Zipf
John Barnes 7/18/2011 10 comments
Just like trying to ignore an elephant in the room, you might get squished while pretending Zipf isn't there.
How Desktop Monte Carlo Detects a Dangerously Bad Metric
John Barnes 7/6/2011 10 comments
This practical example shows how the Monte Carlo testing method can come into play in analyzing the effectiveness of Web delivery mechanisms.
The Monte Carlo Where Everyone Wins
John Barnes 6/27/2011 14 comments
This method of testing metrics can be a sure bet for anyone looking to assure the accuracy and value of analytics.