Analytics for Everyone

Analytics is a hot topic with a lot of buzz in the industry.  All too often I fear that business people tune out discussions of analytics because it is perceived as a topic for analysts, technologists, statisticians and “propeller heads”.  The reality is that carefully crafted analytics can be used by everyone making daily business decisions to enhance the speed and efficacy of those decisions.  As an example we might be alerted to a performance issue that requires our attention.  We typically go through a series of questions as we try to characterize that issue including:

• What lead up to this?
• Has it been a gradual decline in performance or something that happened abruptly?
• If I take no action what is likely to happen?
• What was happening the same time last month, last quarter or last year?
• What are the major contributors to the issue?
• Is this an abnormality or a systemic issue?
• Is there a cause and effect relationship that can be identified?

Then we go through some mental assessment of possibly actions to correct the issue, such as:

• What would happen if we could produce more in a particular facility?
• If we changed suppliers and reduced COGS what would be the impact on profit?
• What if we could increase the conversion rate on our online marketing campaigns?
• If we sold a different product mix how would that impact gross margin?

Analytics can help with all of these questions.  For example:

• Historical trend analysis shows what lead up to the issue as well as whether it happened abruptly or over time;
• Forecast projections show what is likely to happen if no action is taken;
• Variance analyis compares current values to prior periods (or benchmarks) to highlight differences;
• Interactive filtering and “drilling into” the data can highlight the root cause and Pareto charts and histograms can show the major contributors;
• Control charts quickly show whether or not a process is “in control” or has wide variances from one data sample to another;
• Scatter charts with trend lines can show whether or not there is a correlation between two metrics;
• What-if scenarion analysis can show the impact of making one or a combination of changes.

Analytics can be very useful and user-friendly if presented within the context of the data being examined and the decisions that are being made based on that data.  We call this “every-person analytics”.